Chemical Nature of Catalytic Active Sites for the Oxygen Reduction

Article pubs.acs.org/JPCC

Chemical Nature of Catalytic Active Sites for the Oxygen Reduction Reaction on Nitrogen-Doped Carbon-Supported Non-Noble Metal Catalysts Yingdan Qian,† Pan Du,† Ping Wu,*,† Chenxin Cai,*,† and Dominic F. Gervasio*,‡ †

Jiangsu Key Laboratory of New Power Batteries, Jiangsu Collaborative Innovation Center of Biomedical Functional Materials, National and Local Joint Engineering Research Center of Biomedical Functional Materials, College of Chemistry and Materials Science, Nanjing Normal University, Nanjing 210097, P.R. China ‡ Department of Chemical & Environmental Engineering, University of Arizona, 1133 East James E. Rogers Way, Tucson, Arizona 85721, United States S Supporting Information *

ABSTRACT: Much work has been devoted to synthesizing the non-noble metal catalyst such as nitrogen-doped carbonsupported transition metal catalysts (denoted as metal−N−C catalyst) for the oxygen reduction reaction (ORR). However, the catalytic mechanisms and precise chemical nature of the active sites in this kind of catalyst are still controversial, which hinders the development and commercialization of this novel ORR catalyst. The objective of this work is to study the nature of active sites for ORR in the Fe−N−C catalysts. We synthesized a new family of nitrogen-doped carbon with iron catalysts (denoted as Fe−N−C catalysts) by pyrolyzing the mixtures with various ratios of a nitrogen-atom rich heterocycle compound, 1-ethyl-3-methylimidazolium dicyanamide (EMIMdca), and iron chloride (FeCl3). The ORR activity (JK at 0.8 V vs RHE, in 0.1 M KOH solution) of a typical catalyst, Fe15−N− C1000, in this family is 6.65 mA/mg, which is much higher than the values of the Fe−C (0.48 mA/mg) and N−C catalysts (0.25 mA/mg). The relationship between the ORR activity and the structures (the possible active sites in particular) of the catalysts was studied under different conditions. The active site in the catalyst is found to be the Fe−N species (most likely in the form of Fe3N). Metallic iron (Fe) particles, Fe3C species, and N−C species are not catalytically active sites, nor do these moieties interact with the Fe−N active sites during the catalysis of the ORR. High pyrolysis temperatures and increasing the Fe content during the synthesis favor the formation of the Fe−N active sites in the final catalyst. Our study opens up new synthetic control of parameters affecting the final structure and catalyst performance and allows modifying the unexplored avenues toward new multiply heteroatom doped nonprecious ORR catalysts.

1. INTRODUCTION The oxygen reduction reaction (ORR) is an essential reaction in electrochemistry because of its importance in cathodic reaction in polymer electrolyte membrane fuel cells (PEMFCs), a clean and sustainable high energy density power source. Pt and Pt-based alloy catalysts have long been regarded as the best catalysts for the four-electron oxygen reduction reaction (O2 + 4H+ + 4e− → 2H2O).1−25 However, the kinetics of the ORR on Pt-based catalysts are sluggish; the Pt catalysts have poor durability, and the costs of Pt catalysts are too high.1−6 To address these issues with Pt, one promising strategy is to design and synthesize a non-noble metal catalyst such as a nitrogendoped carbon-supported transition metal (metal−N−C catalysts) catalyst.26 The open literature describing non-noble metal catalysts dates back to 1964, when Jasinski27 reported the use of metal phthalocyanine complexes as substitutes for expensive Pt-based catalysts. Following Jasinski’s discovery, it was demonstrated © XXXX American Chemical Society

that a significant increase of catalyst stability and activity (improved by several orders of magnitude) could be achieved by thermal treatment of these phthalocyanine- and porphyrinbased materials.28,29 Next, Gupta et al.30 demonstrated that ORR activities achieved by thermal treatment of metal macrocycles could also be obtained with nitrogen-doped carbon-supported cobalt catalysts, which were synthesized by heat-treating a mixture of Co(II) salt and polyacrylonitrile. Gupta’s results showed that the precursor to pyrolysis need not be a metal in a macrocycle structure (like phthalocyanine and porphyrin) in order to get an active non-noble metal catalyst. Following the above works, numerous nitrogen-doped carbon-supported metal catalysts (especially Fe and Co in metal−nitrogen−carbon networks) have been now synthesized Received: March 15, 2016 Revised: April 27, 2016

A

DOI: 10.1021/acs.jpcc.6b02670 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C and studied in attempts to find non-noble metal catalysts with improved catalytic activity for the ORR.31−44 These non-noble metal catalysts are often synthesized by pyrolysis of transition metal salts with nitrogen-containing molecules or polymeric organic precursors. For example, Wang et al.33 synthesized a non-noble metal catalyst (Fe−N−C catalyst) for ORR by pyrolyzing a composite of carbon-supported Fe-doped graphitic carbon nitride (Fe−g-C3N4@C) at temperature above 700 °C. Kramm et al.34 synthesized a Fe−N−C catalyst by pyrolyzing porphyrins (CoTMPP and FeTMPPCl) together with metal oxalates (iron oxalate dihydrate or anhydrous tin oxalate) and showed that the resulting catalyst exhibited good catalytic activity for the ORR. Besides synthesizing metal−N−C catalysts and characterizing their activity for the ORR, there have been a number of attempts made to obtain a detailed understanding of the structure and mechanism of the catalysts for the ORR.32,34−36,45−47 However, the exact nature of the active sites in the metal−N−C catalysts is still controversial and remains unsolved until today. Some works consider that the nitrogen functional groups in the carbon support are directly responsible for the ORR activities.48,49 They contend that the metal (such as Fe or Co) only facilitates the incorporation of various nitrogen sites (e.g., pyridinic-N, pyrrolic-N, and graphitic-N) into the carbon matrix with a strong Lewis basicity.50 In this N-doped carbon, the N atom is quite negative because N atoms have a more electronegative (the electronegativity of the N and C is ∼3.0 and 2.5, respectively) and strong electron-accepting ability than carbon atoms. Most of the compensating positive charge is distributed on the adjacent carbon atoms, which create the net positive charge centers (via intramolecular charge-transfer) on the adjacent carbon atoms to readily facilitate parallel diatomic adsorption of O2 (because O2 tends to be adsorbed at the sites with positive charges) and consequently increase the catalytic activity toward the ORR by effectively weakening the O−O bond. Theoretical calculations based on density-functional theory (DFT) suggest an energetically favorable O2 adsorption with increasing the number of nitrogen atoms around a CC bond, and only one of the carbons directly bonded to nitrogen (N−C catalytic sites) acts as a catalytic site.51 Other works consider that the interaction of the nitrogen atoms with the metal dopants is responsible for the catalytic activity, e.g., the nitrogen groups serve as a coordinating environment for metal ions. For instance, CoNx (CoN2 and CoN4) moieties have been proposed to act as ORR active sites in the nitrogen-doped carbon-supported cobalt catalysts, which were synthesized by heat treatment of cobalt polypyrrole (Co-PPy) or carbon nanotube-supported Co-PPy at 800 °C.5 In Fe-based catalysts, various kinds of catalytic sites, such as FeN2+2, FeN2, FeN4 moieties, etc., have been proposed.5,34,42,45 Furthermore, both Anderson’s52 theoretical calculations and Kramm’s34 Mössbauer spectra studies have revealed that only ferrous FeN4 sites catalyze the O2 reduction and that ferric FeN4 centers are excluded because of their strong bonding to water. In a recent publication, Tylus et al.35 have concluded from their in situ XANES study that an interaction of FeN4 sites with metal particles is required in order to obtain high onset-potentials for the ORR in an acidic electrolyte. This conclusion was confirmed by Hu et al.53 who proposed that the metal particles are promoters for the ORR on FeN4 sites. However, Zitolo et al.54 and Kramm et al.34 excluded the requirement of the additional metal particles to get good ORR activity, based on the results obtained on the

catalysts free of metallic Fe species. This metal-free condition was achieved by keeping the metal content and the pyrolysis time rather low, or by purifying the as-synthesized catalysts. Furthermore, many authors very recently also proposed that M3C (M = Fe, Co, or Ni) moieties were formed in the nitrogen-doped carbon-supported metal, and the M3C (M = Fe, Co, or Ni) sites actually are the catalytic sites for ORR.55−58 Much work still needs to be done before we can be sure that the exact active sites in the catalysts are known and that the catalytic mechanisms are fully understood for ORR on the nitrogen-doped carbon-supported metal catalyst (metal−N− C). There is a great need to have a detailed understanding of the nature of the active sites in metal−N−C catalysts, and their roles in the ORR, because this knowledge will have a profound effect on successful implementation and commercialization of this kind catalysts and associated fuel cells and also metal air batteries. The objective of this work has been to study the chemical nature of catalytically active sites for ORR on a nitrogen-doped carbon-supported iron catalyst (denoted as Fe−N−C catalyst). For this aim, we first synthesized the Fe−N−C catalyst by pyrolyzing a mixture of iron salt (FeCl3) and 1-ethyl-3methylimidazolium dicyanamide (EMIM-dca) ionic liquid (IL). EMIM-dca was selected as the nitrogen-doped carbon precursor because the nitrogen atom is the only heteroatom in this IL, so we could exclude any confusion due to the possible effects of the other heteroatoms such as S, P, B atoms, etc. In order to better understand the nature of the composition of the catalysts and catalytic sites of the synthesized catalysts, additives, like carbon, were excluded and contaminants, like metallic iron, were removed. The only carbon in the catalyst was from the pyrolyzed IL, that is, there was no other carbon added, such as carbon black, which is commonly added during the synthesis of an ORR catalyst in order to increase the dispersion of precious metals and to reduce the cost of the obtained catalyst. To exclude confusion from the effects of metallic iron and iron oxides particles, which might be formed during the N-doping of the carbon surface, we carefully purified the as-synthesized catalysts by overnight refluxing the synthesized catalyst at 110 °C in 2 M H2SO4 in air, then heat-treating the purified catalyst in forming gas (Ar atmosphere containing 10% H2), and finally repeating a second acid-reflux of the material, similar to the first step of refluxing. The morphologies and their ORR activities of the catalytic materials were carefully characterized and compared at each step of the synthesis. The effects of the initial Fe salt content in the synthesis mixture and the pyrolysis temperature were studied as parameters affecting the composition of the catalyst as well as the catalytic sites for the ORR. We further compared the catalytic activity of the Fe−N−C catalysts with that of the Fe−C catalysts in order to study the role of the doped nitrogen in carbon for enhancing the activity of the Fe−N−C catalysts.

2. EXPERIMENTAL SECTION 2.1. Synthesis of the Fe−N−C Catalysts. In a typical synthesis of the Fe−N−C catalysts, 570 mg of 1-ethyl-3methylimidazolium dicyanamide (EMIM-dca, >98.5%, SigmaAldrich) was mixed with 0.1 g of FeCl3 (the concentration of FeCl3 in the mixture is 15 wt %) by vigorously stirring overnight until a homogeneous mixture evidenced by a clear color was achieved. The mixture was then transferred to a corundum crucible, which was placed in center of a tube B


Article

The Journal of Physical Chemistry C

mapping of the catalysts were recorded by a JEOL JEM-2100F transmission electron microscope, operating at an accelerating voltage of 200 kV. The TEM samples were prepared by drying a droplet of the catalyst solution on a carbon film-support grid. 2.3. Electrochemical Measurements. The electrochemical measurements were carried out with an Autolab PGSTAT302N electrochemical station (Metrohm) in a conventional three-electrode cell with the sample volume of ∼10 mL at ambient temperature. A glassy carbon rotating disk electrode (RDE, 5 mm in diameter, Metrohm) was used a working electrode. A Pt foil (1 cm × 1 cm, Metrohm) and an Ag/AgCl electrode (saturated with KCl solution, Metrohm) were used as counter electrode and reference electrode, respectively. Aqueous KOH solution (0.1 M, AR grade) was used as supporting electrolyte. Prior to use, the RDE was sequentially polished with metallographic abrasive paper (no. 6) and slurries of 0.3 and 0.05 μm alumina to create a mirror finish. The electrode was then sonicated for about 1 min in absolute ethanol and then double-distilled water to remove traces of alumina grit from the electrode surface. After being rinsed thoroughly with double distilled water, it was dried in air at ambient temperature. For immobilization of the catalyst on the electrode surface, 12 mg of the catalyst power were ultrasonically dispersed in a mixed solvent of 450 μL of distilled water, 500 μL of ethanol, and 50 μL of 5 wt % Nafion solution, then ten microliters of asprepared catalyst ink were cast onto the RDE surface with a microsyringe, and the solvent was allowed to be evaporated in air at ambient temperature, resulting in a catalyst loading of 0.61 mg/cm2. The catalyst on the electrode surface is fairly stable since the electrochemical response of the electrode after 200 cycles of continuous scanning between 1.1 to 0.3 V (vs RHE) at 50 mV/s in a N2-saturated 0.1 M KOH solution remained almost unchanged except for a slight decrease during the initial five cycles (total decrease was by ∼3%). Moreover, the electrochemical response of the electrode after storage in a desiccator for 4 weeks virtually did not change compared to the response with a freshly prepared electrode. For comparison, a commercial 20 wt % Pt/C catalyst (JM) was modified on RDE surface using the above procedures. The loading of the Pt/C catalyst on RDE surface is estimated to be ∼20 μgPt/cm2. All electrode potentials in this work are referred to the reversible hydrogen electrode (RHE) by assuming junction potentials are negligible and adding 0.989 V to the potentials measured versus the Ag/AgCl electrode in 0.1 M KOH solution at pH = 13, i.e., ERHE (in V) = EAg/AgCl (in V) + 0.989 V. The ORR activity was measured in O2-saturated 0.1 M KOH solution, which was purged with high purity O2 for at least 30 min, and an O2 flow was streamed over the solution surface. Before the data were recorded, the RDE was cycled in the potential range of 1.1 to 0.3 V (vs RHE) at a scan rate of 50 mV/s for at least 20 cycles until a reproducible cyclic voltammogram was obtained. The ORR polarization curves were recorded by cathodically scanning the RDE potential from 1.1 to 0.3 V with a scan rate of the 5 mV/s with varying rotating speed from 400 to 3600 rpm. Current densities were normalized in terms of geometric electrode area (mA/cm2). Koutecky−Levich plots (J−1 against ω−1/2) were analyzed at various electrode potentials, and the slopes of their best linear fit lines were used to calculate the number of electrons transferred (n) according to eqs 1− 3:

furnace, for pyrolysis under an atmosphere of inter gas (high purity of Ar gas). The mixture was first heated by ramping temperature 5 °C/min to 1000 °C and then holding at 1000 °C for 2 h to complete the pyrolysis. The furnace was turned off, and the temperature was allowed to cool down naturally to an ambient temperature. The pyrolysis products were purified by overnight refluxing in 100 mL of H2SO4 solution (2 M) in air at 110 °C to remove any metallic Fe and iron oxides particles produced during pyrolysis, then heating in forming gas (Ar/10% H2) at a temperature of 1000 °C for 2 h, followed by refluxing in 2 M H2SO4 solution, similar to the first refluxing procedure. After each refluxing step, the suspensions were filtered and washed with distilled water until the filtrate exhibited a pH value of ∼7. The catalyst was finally obtained after drying the filter cakes in air at 80 °C and was denoted as Fe15−N−C1000 catalyst. For study of the effects of the Fe salt contents and the pyrolysis temperature on the catalytic activities and active sites of the Fe−N−C catalysts, the catalysts were synthesized by changing the FeCl3 contents from 5 to 15 wt % in the synthesis mixture while keeping the pyrolysis temperature at 1000 °C, and the pyrolysis temperature was varied from 700 to 1000 °C while keeping the FeCl3 content at 15 wt %. These catalysts are denoted as Fex−N−Cy, where the value of x stands for the FeCl3 content in the synthesis mixture and the value of y for the pyrolysis temperature. For example, Fe10−N−C900 means the catalyst was synthesized by pyrolyzing the mixture of EMIMdca and FeCl3 at 900 °C with the mass ratio of FeCl3 of 10 wt % in the synthesis mixture. For evaluating the role of the doped nitrogen atom in the carbon support for the enhancement of the catalytic activity of the catalyst, Fe−C catalyst (without doped N in carbon) was synthesized by pyrolysis of a mixture of FeCl3 and glucose (15 wt % of FeCl3 in the mixture) at 1000 °C using the same procedures as those for the Fe−N−C catalyst synthesis except that EMIM-dca was replaced by glucose. The N−C catalyst was synthesized by pyrolyzing the EMIMdca at 1000 °C in the same manner as that of the synthesis of the Fe−N−C catalysts and treated with acid in the same procedures as those of Fe−N−C catalysts. 2.2. Catalyst Characterizations. The specific surface areas of the catalysts were calculated based on the nitrogen sorption isotherms, which were measured using a Quadrasorb MP porosimetry apparatus (Quantachrome). Samples were degassed at 200 °C for 10 h at reduced pressure before measurements. The content of Fe in the catalysts was quantitatively determined by optical emission spectroscopy using inductive coupled plasma as ionization source (ICPOES), which was performed using an Optima 7300 DV (PerkinElmer). The surface characteristics of the catalysts were examined by X-ray photoelectron spectroscopy (XPS), which was measured with ESCALAB 250 XPS spectrometer (VG Scientifics) using a monochromatic Al Kα line at 1486.6 eV. Binding energies were calibrated with respect to the C 1s peak at 284.6 eV. Peak fit analysis was performed using the XPS PEAK program (V4.0). The structures of the catalysts were studied using their X-ray diffraction (XRD) patterns, which were recorded on a Rigaku/Max-3A X-ray diffractometer with Cu Kα radiation (λ = 0.15418 nm). Raman spectra were recorded on a Labram HR 800 microspectrometer (Jobin Yvon) with an excitation source of 514 nm. The morphologies, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images, and element analysis C


Article


Figure 1. TEM images of the synthesized Fe15−N−C1000 catalyst before (a−d) and after purification (e−h). HAADF−STEM image of the catalyst (i), and the corresponding elemental mapping for C (j), N (k), and Fe (l). The catalyst was synthesized by pyrolyzing the mixture of FeCl3 (15 wt %) and EMIM-dca at 1000 °C.

1 1 1 1 1 = + = + J Jk JL Jk Bω1/2

(1)

B = 0.62nFC0D02/3v−1/6

(2)

Jk = nFkC0

(3)

Fe at 44.7, 65.0, and 82.4 degrees (Figure 2a, JCPDS number 65-4899), respectively.

where J is the measured current density, Jk and JL are the kinetic- and diffusion-limiting current densities, ω is the angular velocity of the disk (ω = 2πN, where N is the linear rotation speed), n is transferred electron number, F is the Faraday constant (96485 C/mol), C0 is the bulk concentration of O2 (1.2 × 10−6 mol/mL), ν is the kinematic viscosity of the electrolyte (0.01 cm2/s), D0 is the O2 diffusion coefficient (1.9 × 10−5 cm2/s), and k is the electron-transfer rate constant. The number of electrons transferred (n) and Jk can be obtained from the slope and intercept of the Koutecky−Levich plots, respectively.

Figure 2. XRD patterns of the Fe15−N−C1000 catalyst before (a) and after purification (b). The curve (c) represents the XRD patterns of the Fe−C catalyst, which was synthesized by pyrolyzing the mixture of FeCl3 (15 wt %) and glucose at 1000 °C.

3. RESULTS AND DISCUSSION 3.1. Morphologic and Structural Characterization. The observed morphology indicates that the synthesized catalyst is loosely and uniformly dispersed (images (a) and (e), Figure 1). Before purification, there are many small dark particles with sizes of about 4−5 nm in diameter wrapped in the carbon layers as shown in image (b) (marked with red circles) and the magnified image in image (c). High-resolution TEM (HRTEM) images indicate that these small particles have a high crystallinity with a lattice spacing of 0.213 nm (image (d), Figure 1), which is slightly larger than the standard value of the (110) plane of metallic Fe (0.204 nm). The larger lattice space might be ascribed to intercalation of nitrogen (and/or carbon) atom into the spaces of iron,59 which accordingly enlarges the lattice constant. Thus, we can conclude that the observed small dark particles are metallic Fe particles, which are formed in the synthesis of the catalyst. This conclusion can be further verified by the XRD pattern measurements, which show strong diffraction peaks of (002), (102), and (103) planes of metallic

The metallic Fe particles were effectively removed from the surface of the catalyst by our purification procedures, as verified by TEM images depicted in images (f)−(h) (Figure 1). Even in the HRTEM images, we do not find the existence of any metallic Fe particles. We only observed the typical graphenelike structure (images (g) and (h), Figure 1). Because of the heterogeneous nature of the catalyst under study, and to confirm that the results depicted in images (g) and (h) in Figure 1 represent the typical HRTEM images of the synthesized catalyst, we randomly selected eight different sites to record the HRTEM images, which also show that only the graphene-like structure can be observed (Figure S1), implying that the metallic Fe particles have been completely removed after the purification or the purification can reduce the size of metallic Fe particles to that below the resolution of HRTEM. However, the study of the element distribution of the purified D


Article


Figure 3. (a) XPS spectrum of the synthesized Fe15−N−C1000 catalyst. The inset shows the amplified XPS of N 1s and Fe 2p. (b−d) Highresolution XPS spectra of the C 1s (b), N 1s (c), and Fe 2p (d), respectively, and their related curve-fitted components.

acid-refluxing, any iron oxides were reduced to metallic iron and removed in the second acid-refluxing. XPS was used to study the detailed chemical environments of the catalyst surface. The survey XPS spectrum shows the presence of carbon, nitrogen, oxygen, and iron elements (Figure 3a). The presence of the oxygen is a typical phenomenon as the surface of carbon materials tends to be oxidized slightly when handled under ambient conditions. The atomic ratios of the C, N, O, and Fe are about 92.07, 1.26, 6.53, and 0.14 at %, respectively. Although ICP-OES measurements showed the weight ratios of the iron species in the catalyst are ∼1.72 wt % after the purification, only ∼0.63 wt % of Fe is detected by XPS. This is probably a result of the XPS sampling being restricted to the surface, so only the iron species on the catalyst surface (the XPS signals of Fe 2p come from only 1.5− 2 nm depths of the catalyst) were detected.57 Therefore, we can conclude from the TEM image, HAADF-STEM elemental imaging, and XRD results that iron species are well dispersed and incorporated into the carbon layer and form catalytically active centers. The shapes of the C 1s core level spectrum is asymmetrical and shows a broad satellite at higher values of binding energy between ∼287.6 and 291.2 eV (Figure 3b). A satellite feature in this region is typical for graphitic carbon binding situations and is caused by the delocalized π-electron system. This observation is in agreement with the XRD patterns, corroborating a higher degree of graphitization of the N-doped carbon (see TEM image depicted in Figure 1g,h). Furthermore, in this higher binding energy region, oxidized carbon species appear, which could be the reason for the observation of these satellite peaks. The presence of these peaks is typical for carbon materials handled in an air atmosphere. The two dominant contributions to the spectrum at ∼284.6 and 285.8 eV are of higher significance. At ∼284.6 eV carbon atoms neighboring other carbon atoms in a sp2 binding environment are represented, while at the higher binding energies (∼285.6 eV), carbon atoms are bound to more electronegative binding partners.60 Here, this can be ascribed to the presence of nitrogen-bound carbon

catalyst using HAADF−STEM indicates the existence of C, N, and Fe elements (images (i)−(l), Figure 1); moreover, these elements are uniformly distributed on the catalyst surface, suggesting that these Fe species are chemically coordinated to the nitrogen-doped carbon, most likely in the form of Fe−N or/and Fe−C bonds. This can be confirmed by XRD measurements, which can offer a further insight into the structural long-range order of the samples. As depicted in Figure 2b, the XRD patterns of the purified catalyst show a peak at ∼26.0 degrees, which corresponds to the (002) carbon plane.11 The strong intensity of this peak indicates the high degree of graphitization, which agrees with that observed from HRTEM images. The XRD peak of the (002) plane of metallic Fe (at ∼44.8 degrees) is too weak to be quantitatively detected, suggesting that the metallic Fe particles are removed after our purification, or a very small portion of the residual metallic Fe particles is buried inside the catalysts too deep to be detected by XRD. This very small amount of the metallic Fe should not affect the analysis of the active sites of the catalyst. It is noteworthy that the XRD of the catalyst shows an apparent peak at ∼43.6 degrees, which is significantly different than the XRD peak of the sample before purification (Figure 2b). This peak is attributed to the characteristic peak of (111) plane of Fe3N (JCPDS number 49-1664), implying Fe3N has been formed during pyrolysis. Furthermore, we observed that there is a small amount of the Fe3C in the purified catalyst because of the appearance of a weak peak at ∼42.8 degrees, which is attributed to the diffraction peak of the (211) plane of Fe3C (JCPDS number 65-2412). The ascription of the peak at ∼42.8 degrees to Fe3C can be further confirmed from the XRD patterns of the Fe−C catalyst depicted in Figure 2c, which shows a strong peak at ∼42.7 degrees corresponding to the (211) plane of Fe3C. No patterns corresponding to the iron oxides were found, implying that there are no iron oxides present in the catalyst, probably due to the heat treatment of the catalyst in the reducing atmosphere (Ar/10%H2) after first E


Article


Figure 4. (A) RDE measurements in O2-saturated 0.1 M KOH solution for the carbon (curve a), Fe−C (curve b), N−C (curve c), unpurified Fe15− N−C1000 (curve d), commercial 20 wt % Pt/C (curve e), and purified Fe15−N−C1000 catalyst (curve f). The measurements were performed at ambient temperature with a rotating speed of 1600 rpm and a scanning rate of 5 mV/s. The loading of the catalyst was 0.61 mg/cm2, and the loading of Pt/C catalyst is 20 μgPt/cm2. (B) The values of the ORR onset potential (Eonset, defined as the potential at the cathodic current reaches 0.1 mA/ cm2) and half-wave potential (E1/2) obtained at different catalysts under study. (C) The electron transfer number (n) at 0.4 V and the mass activity (JK at 0.8 V vs RHE) for the studied catalysts. (D) Tafel ORR plots obtained at the purified Fe15−N−C1000 (a), 20 wt % Pt/C (b), unpurified Fe15−N−C1000 (c), N−C (d), and Fe−C (e) catalysts.

nitrogen atoms truly replacing carbon atoms in graphene-sheet binding environments with the excess electron delocalized in the π-system (sp2 hybridized). The peak at ∼401.4 eV corresponds to quaternary N, and the peak at a rather high binding energy (∼402.2 eV) is ascribed to graphitic N. These descriptions were supported by accurate ab initio model calculations of the core level ionization potentials and the core level shifts.61 When we decompose the N 1s peak, another peak, shifted up by ∼1.4 eV from the binding energy of pyridinic N, has to be included in the fit. This corresponds to the binding energy of ∼399.8 eV. This is the energy of nitrogen bonds to metal,47 the Fe−N bond (most likely in the form of Fe3N) in this work, which is in good agreement with that observed from XRD measurements. The Fe atom in this Fe−N bond should be bound with pyridinic N because Fe bound to pyrrolic N, such as in porphyrins, appears at a lower binding energy of 398.5 eV.62 Moreover, Artyushkova et al. also confirmed the presence of Fe−N species in the range 399.8−400 eV by density functional theory (DFT) calculations of binding energy shifts for metal−N−C graphene-based materials.63 The percentage contents of different types of N in total N amount were calculated to be about (in atomic ratio) 42.03, 22.87, 8.5, 14.09, and 12.51 at % for pyridinic-, pyrrolic-, quaternary-, graphitic-N, and Fe−N, respectively. To further confirm the presence of Fe−N bond in N 1s spectrum, we have done the analysis of spectrum in Fe 2p region for the same sample (Figure 3d). In addition to a small peak due to metallic Fe (∼707.4 eV), there is a distinct peak due to the Fe−N bond at ∼710.4 eV. Moreover, it can be concluded that the Fe species are in the form of Fe(II) as

atoms. Thus, the carbon core spectrum reveals N-functionalized carbon in which the nitrogen atoms are anchored within a carbon backbone with a graphitic microstructure. However, we did not detect the presence of a peak corresponding to a carbon atom bound to Fe atom, implying there are very low amounts of iron carbide species in the catalysts. The core level spectrum of the N 1s is provided in Figure 3c, which shows that the N 1s core level spectrum has an asymmetrical shape and thus a broadened contribution at high binding energy values of more than 404 eV. This feature can, on one hand, be attributed to oxidized nitrogen species on the catalyst’s surface, or can, on the other hand, be interpreted as a satellite feature caused by π-electrons of nitrogen atoms bound in an aromatic chemical environment.32 This satellite peak reveals the chemical state of the nitrogen atoms in differently functionalized carbon materials as being firmly incorporated into the carbon backbones. This can be confirmed from the presence of the two dominant peak contributions, which appear at binding energies of ∼398.4 and 400.7 eV, respectively, and two relatively small peak contributions, which appear at binding energies of ∼401.4 and 402.2 eV, respectively, in N 1s spectrum. The former spectral feature (two dominant peaks) corresponds to pyridinic or pyrrolic nitrogen, representing nitrogen atoms bound at the edges of graphite layers.47,60 The peak at ∼398.4 eV is commonly ascribed to be pyridinic N contributing to the π-system with one π-electron as confirmed from reference spectra acquired from polypyridine.32 The peak at ∼400.7 eV is ascribed to pyrrolic N with two π-electrons as established by reference spectra acquired from the polypyrrole.32 The latter spectral feature (two relatively small peaks) corresponds to the graphitic nitrogen atoms, referring to F


Article

The Journal of Physical Chemistry C evidenced by the peaks at ∼712.5 and 723.3 eV. The content of Fe(II) is estimated to be 79.4 at%. Although Fe(III) species are also detected because of the appearance of the peaks at ∼712.5 and 725.8 eV, the amount of the Fe(III) species is only 9.8 at %, which is very low in comparison with that of Fe(II) species. This observation is significant because it is known that only ferrous iron sites catalyze the O2 reduction, and ferric iron centers are not active sites for the ORR.34,42 3.2. Electrocatalytic Characteristics and Active Sites. In this section, the relationship is studied between the catalytic activity, as determined by cyclic voltammetry (CV) and RDE measurements (performed in O2-saturated 0.1 M KOH solution), and the structures (in particular, the active site structure), as determined by XPS and XRD for the catalysts synthesized under different conditions. The catalytic activity of a commercial 20 wt % Pt/C catalyst serves as a reference. CV results indicate a well-defined cathodic peak for O2 reduction is observed at ∼0.71 V (vs RHE) for the unpurified Fe15−N−C1000 catalyst (Figure S2b). After purification, this cathodic peak appears at ∼0.80 V (Figure S2c), which is about a 90 mV positive shift in comparison with that of unpurified catalyst. This cathodic peak for O2 reduction is even slightly more positive than that obtained at the commercial 20 wt % Pt/ C catalyst, which shows the O2-reduction peak at ∼0.79 V (Figure S2a). The cathodic peak current (in mA/cm2) obtained at the purified Fe15−N−C1000 catalyst is ∼1.22 mA/cm2, which is higher than that obtained at the commercial Pt/C (∼0.85 mA/cm2) and unpurified catalyst (∼0.56 mA/cm2). These results suggest that the purified Fe15−N−C1000 catalyst has a high ORR activity in alkaline media, and metallic Fe particles observed in the unpurified catalyst do not contribute to the O2 reduction, at least those metallic particles do not play a key role in the catalytic O2 reduction; otherwise, the ORR activity of the unpurified Fe15−N−C1000 catalyst should be higher than that of the purified one. To study the origin of the high catalytic activity of the purified Fe15−N−C1000 catalyst, we further perform the O2 reduction at the Fe−C (N free), N−C (metal free), and carbon (both N and metal free) catalysts. The catalytic activity of the carbon catalyst is too low to be clearly observed (Figure S2d). The peak potentials of the O2 reduction at Fe−C and N−C catalysts are ∼0.64 (Figure S2e) and 0.66 V (Figure S2f), respectively. The peak currents are ∼0.12 and 0.80 mA/cm2 obtained at the Fe−C and N−C catalysts, respectively. These results indicate that the high catalytic activity observed at the purified Fe15−N−C1000 catalyst does not come from the Fe3C; otherwise, the Fe−C catalytic should have the similar catalytic activity to that observed at the purified Fe15−N− C1000 catalyst (the presence of the Fe3C species in the synthesized Fe−C catalyst can be confirmed by its XRD patterns depicted in Figure 2c, which shows the (211) plane of Fe3C at ∼42.8 degrees). N-doped carbon catalyst (N−C) does not also contribute much to the high catalytic activity of the purified catalyst (Of note, the amount of the N element in the N−C and purified Fe15−N−C1000 catalysts is similar; it is ∼1.86 and 1.26 at %, respectively, estimated from XPS measurements). Therefore, the high catalytic activity of the Fe15−N−C1000 catalyst may come from the Fe−N species (most likely in the form of Fe3N) in the catalyst. To further determine the active sites in the Fe15−N−C1000 catalyst, RDE measurements were performed on the synthesized catalyst, and the results are depicted in Figure 4. The data were analyzed based on the value of the ORR onset potential

(Eonset, defined as the potential at the cathodic current reaches 0.1 mA/cm2),64,65 half-wave potential (E1/2), mass activities (JK at 0.8 V vs RHE), and the electron transfer numbers (n). The RDE gives similar results to those obtained from CV; the catalytic activity of the carbon catalyst is almost negligible because the catalytic current is almost undetectable even at potential of 0.4 V (curve a, Figure 4A). The catalytic activity of the Fe−C catalyst is low (curve b, Figure 4A) reflected by the low magnitude of the catalytic currents. The catalytic activity of N−C catalyst (curve c, Figure 4A) is higher than that of carbon catalyst, implying the contributions of the doped N to the catalytic activity. However, it is still lower than that of the purified Fe15−N−C1000 catalyst (curve f, Figure 4A), which is much closer to the hydrodynamic limit and the corresponding oxygen transport conditions, and is almost the same as that obtained at the Pt/C catalyst (curve e, Figure 4A). We thus conclude that the presence of metal in the catalyst appears important for high levels of ORR activity, rivaling high surface area Pt. The value of the catalytic current at the unpurified Fe15−N− C1000 catalyst (curve d, Figure 4A) is considerably lower than that of purified Fe15−N−C1000 catalyst, probably due to that the inert metallic Fe particles can block the reactant (O2 molecules) to access to the active sites or even due to contamination leading to a change of mechanism. Improved ORR activity for the purified Fe15−N−C1000 catalyst (see the discussion presented below) provides evidence supporting the notion of the active site blockages by the inert metallic Fe particles. Thus, from the analysis of the catalytic current, we can conclude that the catalytic activity of the purified Fe15−N− C1000 catalyst is much higher than those of the N−C and unpurified Fe15−N−C1000 catalyst. The value of the Eonset obtained at the purified Fe15−N− C1000 catalyst is ∼0.99 V (vs RHE) (Figure 4B). This value is 40, 100, 170, and 190 mV more positive than that obtained at the Pt/C (∼0.95 V), unpurified Fe15−N−C1000 (∼0.89 V), N−C (∼0.82 V), and Fe−C (∼0.80 V) catalysts, respectively. The value of E1/2 obtained at the purified Fe15−N−C1000 catalyst (∼0.85 V) is also more positive than that obtained at other studied catalysts (Figure 4B). Those results indicate that the catalytic activity of the purified Fe15−N−C1000 catalyst is higher than that of Pt/C and unpurified Fe15−N−C1000 catalyst. The diffusion-limited current at the purified Fe15−N− C1000 catalyst is close to that at the Pt/C catalyst, indicating both promote a 4e− O2 reduction mechanism. This conclusion can be further verified by the analysis of mass activities and the calculations of electron transfer numbers. Figure 4C displays the electron transfer number (n), which is obtained from the analysis of the slope of the Koutecky−Levich (a typical Koutecky−Levich for purified Fe15−N−C1000 catalyst is depicted in Figure S3), and the ORR mass activity (JK) at 0.8 V of the various catalysts under studied. The results of carbon catalyst are not included here due to its negligible catalytic activity. The results of the Pt/C catalyst are also not compared here because the mass loading of the Pt/C (∼20 μgPt/cm2) and the synthesized catalysts (0.61 mg catalyst/cm2, corresponding to ∼10 μgPt/cm2 for the purified Fe15−N− C1000 catalyst) on the electrode surface are different (see Experimental Section for details); thus, their mass activities do not have comparability. It is clear that purified Fe15−N−C1000 catalyst displays the optimal activity with the highest electron transfer of 3.98 and the highest mass activity of 6.65 mA/mg at 0.8 V, which is much higher than the values of unpurified Fe15− G


Article


mass activities (JK, at 0.8 V vs RHE), and n as a function of Fe content are also depicted in Table 1. With increasing of Fe content, the catalytic activity of the catalyst was seen to improve significantly as revealed by the positive shifts of the values of the Eonset (from 0.95 to 0.99 V) and E1/2 (from 0.80 to 0.85 V), and the increase of the JK value (at 0.8 V) from 3.16 at the Fe content of 5 wt % (in synthesis mixture) to 6.65 mA/mg at the Fe content of 15 wt %. Moreover, increasing of the Fe content also enhances the ORR selectivity of the catalyst as demonstrated by the increase of n values from 3.88 to 3.98. Because the catalytic activity of the synthesized Fe−N−C catalyst is related to the concentration of the active sites that exist as Fe−N species (most likely in the form of Fe3N), we can conclude that, by increasing of the Fe content in the synthesis mixture, there was a significantly enhancement of the amount of the Fe−N species in the resulting catalyst until all pyridinic N is coordinated with Fe ions. The increase of the Fe3N species can be verified by the XRD measurements as depicted in Figure S5 in which the XRD peak corresponding to Fe3N (at ∼43.6 degrees) becomes more prominent as the initial Fe content in the synthesis mixture was increased from 5 to 15%. XPS spectral analysis also indicates that the percentages of Fe−N species making up the total amount N species increase as Fe content in the synthesis mixture increases (Figure S6), that is there are ∼7.57 at % of Fe−N species in the catalyst when 5 wt % FeCl3 were added to the synthesis mixture, and ∼12.51 at % of Fe−N species with 15 wt % FeCl3 were added to the synthesis mixture, implying that the amount of the N−C species decreases since the total nitrogen content remains essentially constant. These results again suggest that the Fe−N species (most likely the Fe3N in this work) in the synthesized Fe−N−C catalyst act as the active sites in the ORR, and the N−C species do not play a key role or serve as active sites in the ORR. Moreover, the contributions from interaction of the Fe−N species with N−C species can also be excluded because the amount of N−C species in the catalyst decreases with the increasing of the initial Fe content. 3.4. Effects of the Pyrolysis Temperature. The pyrolysis temperature may affect the formation of the active sites, we studied the ORR catalytic activity of the catalyst as a function of the pyrolysis temperature by varying the pyrolysis temperature from 700 to 1000 °C while keeping the Fe content in the synthesis mixture constant at 15 wt %. The catalytic activities of the purified catalysts for O2 reduction are shown in Figure 5. We see that the diffusion-limited current observed at the Fe15− N−C700 catalyst (∼3.2 mA/cm2, curve a, Figure 5A) is much lower than those observed at Fe15−N−C800 (∼5.5 mA/cm2), Fe15−N−C900 (∼5.6 mA/cm2), and Fe15−N−C1000 (∼5.6 mA/cm2). The low diffusion-limited current of the Fe15−N− C700 catalyst may be caused by the low value of its specific surface areas as calculated from the nitrogen sorption isotherms (Figure S7), which show the BET surface areas of the Fe15−N− C700 (184.5 m2/g) are almost half of those for Fe15−N−C800 (356.2 m2/g), Fe15−N−C900 (361.4 m2/g), and Fe15−N− C1000 (365.3 m2/g). Low surface area does not favor the access of O2 molecules to the catalyst surface and, thus, results in a low diffusion-limited current for this catalyst as observed in the Figure 5A. Alternatively, the low pyrolysis temperature may not allow the active site to form and therefore leads to a change of mechanism for Fe15−N−C700 compared to the catalysts pyrolyzed at higher temperatures. It is clear that the Fe15−N−C700 catalyst has a low catalytic activity by noting the relatively low values of Eonset (∼0.78 V),

N−C1000 (1.88 mA/mg), Fe−C (0.48 mA/mg), and N−C catalyst (0.25 mA/mg) catalysts. The above results imply that the purified Fe15−N−C1000 catalyst has high efficiency in catalyzing the ORR. The improvement in catalytic activity observed for purified Fe15−N−C1000 catalyst implies that the metallic Fe particles can be excluded as an active site in the Fe15−N−C1000. We can also exclude that the interaction of the metallic Fe particles with the Fe−N species serves as active sites for ORR, in agreement with previous reports.35,53 The low catalytic activity observed for the Fe−C and C−N catalysts suggest that the N− C species and the Fe3C species do not serve as active sites in catalyzing the ORR. Although we cannot exclude that the interaction of the Fe−N species with N−C species or/and the Fe3C species contributes to the catalytic activity for the ORR, we can conclude from the RDE data that the Fe−N species (Fe3N in this work) in the synthesized Fe−N−C catalyst plays an important role in catalyzing ORR and may serve as active site in the catalyst, which is the same as the conclusion obtained from the CV analysis. The Tafel slope was also determined to evaluate the ORR kinetic features of the synthesized catalysts (Figure 4D). The Tafel plots of the catalyst show two distinct slope regions. In comparison to the well-established “dual” Tafel slope for the ORR on Pt (60 and 120 mV/dec at potential higher and lower than 0.8 V, respectively), the Tafel slope values of ∼62 and 63 mV/dec observed with the unpurified and purified Fe15−N− C1000 catalysts, respectively, at high potential region, which are very close to the those obtained at Pt-based catalysts. The Tafel slope value of ∼82 and 80 mV/dec is obtained at the Fe−C and N−C catalysts, respectively. The difference in Tafel slope values for the Fe−N−C catalysts and Fe−C and N−C catalysts implies a different nature of the ORR active sites in these catalysts. 3.3. Effects of Fe Loading. We studied the correlation between the initial Fe content in the synthesis mixture and the corresponding residual content in the final catalysts of Fe and N, and the correlation with their ORR activity. The compositions of the Fex−N−C1000 catalyst (x = 5, 10, 15) are tracked by x, where x denotes the content of FeCl3 wt % in the synthesis mixture, which are summarized in Table 1. The data show that an increase in initial Fe content leaves the nitrogen content essentially constant (1.24−1.26 at %), while the amount of Fe follows the initial Fe content. The ORR polarization curves of the purified Fex−N−C1000 catalyst in O2-saturated alkaline solution with different Fe contents are depicted in Figure S4. The values of Eonset, E1/2, Table 1. Effects of the Fe Content in the Synthesis Mixture on the Catalytic Characteristics of the Obtained Fex−N− C1000 Catalyst (x = 5, 10, and 15) initial Fe content (wt %) 5 10 15

Fe (at %)

Na (at %)

Eonsetc (V vs RHE)

E1/2 (V vs RHE)

JKd (mA/mg)

n

0.08 (1.41)b 0.11 (1.62)b 0.14 (1.72)b

1.25 1.24 1.26

0.95 0.97 0.99

0.80 0.84 0.85

3.16 5.54 6.65

3.88 3.91 3.98

a

a

These values refer to the atomic ratios and are obtained from the XPS analysis. bThe values in the parentheses are Fe wt % content in the synthesized catalysts obtained from ICP measurements. cThe values are defined as the potential at the cathodic current reaches 0.1 mA/ cm2. dThe mass activities are obtained at potential of 0.8 V (vs RHE). H


Article


Figure 5. (A) RDE measurements in O2-saturated 0.1 M KOH solution for the purified Fe15−N−C700 (curve a), Fe15−N−C800 (curve b), Fe15− N−C900 (curve c), and Fe15−N−C1000 catalysts (curve d). The measurements were performed at ambient temperature with a rotating speed of 1600 rpm and a scanning rate of 5 mV/s. The loading of the catalyst was 0.61 mg/cm2. (B) Dependence of the values of Eonset, E1/2, and JK (at 0.8 V) on the pyrolysis temperature of the catalyst synthesis.

Figure 6. XRD patterns (A) and Raman spectra (B) of the purified Fe15−N−C700 (curve a), Fe15−N−C800 (curve b), Fe15−N−C900 (curve c), and Fe15−N−C1000 catalysts (curve d).

E1/2 (∼0.59 V) (Figure 5B), and magnitude of the mass activity (JK at 0.8 V, ∼ 0.16 mA/mg). These values shift to ∼0.94 V (Eonset), 0.71 V (E1/2), and 3.52 mA/mg for a pyrolysis temperature of 800 °C and reach the highest values of the parameters when the pyrolysis temperature is 1000 °C (Figure 5B), indicating that the catalytic activity of the catalyst increases as the pyrolysis temperature increase. The enhancement in the catalytic activity with the pyrolysis temperature implies that the high pyrolysis temperature favors the formation of the active sites in the catalyst. To confirm this conclusion, we recorded and compared the XRD patterns of the catalysts synthesized at different pyrolysis temperature. As seen in Figure 6A, the intensity of the characteristic peak of the (111) plane of Fe3N species is too weak to be seen when the pyrolysis temperature was 700 °C (curve a, Figure 6A), implying the amount of the Fe3N species in the catalyst is very low. The intensity of this characteristic peak increases with increasing pyrolysis temperature (curves b−d, Figure 6A), suggesting that the amount of the Fe3N species in the catalysts increases with the pyrolysis temperature. The above XRD results are consistent with XPS results, which indicate that the percentages of Fe−N in the total amount of N increase as the pyrolysis temperature increases. The percentage of Fe−N goes from 1.8 at % at a pyrolysis temperature of 700 °C to 9.30, 11.28, and 12.51 at % at pyrolysis temperature of 800, 900, and 1000 °C (Figure S8), respectively, implying that a higher pyrolysis temperature favors the formation of the Fe−N active sites. It is interesting to note that the intensities of the peak corresponding to the (211) plane of Fe3C (at ∼42.8 degrees) remain essentially constant with increasing the pyrolysis temperature, which suggests that the amount of the Fe3C species in the catalysts does not increase much as the pyrolysis temperature increases from 700 to 1000 °C. These results further confirm that the Fe−N (in

the form of Fe3N) species is the active site rather than the Fe3C species for ORR in the synthesized catalysts. The increase of the conductivity of the synthesized catalysts may also partially contribute to the enhancement in the catalytic activity with the pyrolysis temperature because a higher pyrolysis temperature results in a higher degree of graphitization of the carbon materials. This can be confirmed from the full width at half-maximum (fwhm) of the XRD characteristic peak of the (002) plane of carbon materials (at ∼26.0 degrees). The values of fwhm of this peak decrease from 1.32 degrees for catalyst synthesized at 700 °C to 1.26, 1.07, and 1.02 degrees for the catalysts synthesized at pyrolysis temperature of 800, 900, and 1000 °C, respectively, indicating the enhancement in the graphitization and, in turn, the enhancement in the conductivity and the catalytic activities. The increase in the graphitization with the pyrolysis temperature can further be verified by Raman spectroscopic measurements. Raman spectroscopy is usually used to characterize the degree of graphitization of carbon materials. Raman spectra of the synthesized catalysts exhibit two peaks at ∼1347 and 1592 cm−1 corresponding to the well-defined D and G bands (Figure 6B), respectively. The ratio of ID/IG is usually used to evaluate the disorder in the graphene structures. The G band is related to the E2g vibration mode of sp2 carbon domains and is associated with the degree of graphitization, while the D band is associated with structural defects and partially disordered structures of the sp2 domains. The D and G bands for the Fe15−N−C700 catalyst cannot be well separated, implying the low graphitization at this pyrolysis temperature. When the pyrolysis temperature is higher than 700 °C, two well-defined peaks are observed and the value of ID/IG is estimated to be ∼1.29, 1.55, and 1.59 for the catalyst synthesized by pyrolyzing at 800, 900, and 1000 °C, respectively, suggesting the increase in the degree of I


Article


Figure 7. Cyclic voltammetric responses of the purified Fe15−N−C1000 (a) and Pt/C catalyst (b) in N2-saturated 0.1 M KOH solution at 50 mV/s before and after ADTs. RDE measurements in O2-saturated 0.1 M KOH solution for the purified Fe15−N−C1000 (c) and Pt/C catalyst (d). The measurements were performed at ambient temperature with a rotating speed of 1600 rpm and a scanning rate of 5 mV/s. The loading of the synthesized catalyst was 0.61 mg/cm2, and the Pt/C catalyst was 20 μgPt/cm2.

durability of Fe15−N−C1000 catalyst may be due to the active sites in the catalyst being not easy to be oxidized and reduced. However, Pt on carbon is known to undergo area loss due to a dissolution and precipitation via a corrosive process. The Fe being highly dispersed and coordinated to N, which is doped into carbon, appears to stabilize the Fe atoms in the catalyst against corrosive dissolution. Consequently, the synthesized Fe15−N−C1000 catalyst is more stable than Pt/C catalyst.

graphitization with the pyrolysis temperature, which help catalyst activity by minimizing polarization (due to resistive voltage losses). 3.5. Stability of the Active Sites. The stability of the active sites in the ORR catalysts is an important issue in evaluating the performance of the catalyst. To probe the longterm stability of the synthesized catalysts, the ORR activity of the Fe15−N−C1000 catalyst before and after an accelerating durability tests (ADTs), which were performed by potential cycling of the catalyst between 0.6 to 1.0 V (vs RHE) at 50 mV/s in a N2-saturated 0.1 M KOH solution for 5000 cycles, was recorded and compared with that for a commercial Pt/C catalyst. The CV curves of the Fe15−N−C1000 catalyst before and after the ADTs are depicted in Figure 7a. From which, we can see there is almost no change in the CV responses, implying that the active sites (Fe3N species) in the catalyst are stable. The active sites do not dissolve from the catalyst. However, Pt/ C suffers from greater degradation during the ADTs, as reflected by the decreasing of the redox peaks associated with the adsorption/desorption of hydrogen and the formation/ reduction of Pt oxides (Figure 7b). This is because Pt particles are likely to dissolve in electrolyte, aggregate into larger particles, and detach from the support, resulting in poor durability. The ORR activities of the Fe15−N−C1000 catalyst before and after ADTs are shown in Figure 7c, which shows almost no changes being detected for the values of Eonset, and only slightly negative shift of E1/2 (∼12 mV) and decrease in JK (decrease by ∼6%), indicating the catalytically active sites remain intact during the ADTs in alkaline media. In contrast, the negative shifts of 40 and 20 mV are found for the E1/2 and Eonset, respectively, for commercial Pt/C catalyst after the ADTs (Figure 7d). Moreover, the mass activity also decreases by ∼15%. The results suggest that the Fe15−N−C1000 catalyst exhibits better durability than commercial Pt/C under these same test conditions in alkaline electrolyte. The enhanced

4. CONCLUSIONS We have successfully synthesized new highly active family of Fe−N−C catalysts by using EMIM-dca as the nitrogen-doped carbon precursor and have investigated the chemical nature of active sites for the ORR in the catalysts. These new catalysts have been systematically tested for their ORR activities by RDE measurements and have been structurally characterized based on the spectroscopic measurements. We have established that the active sites for the ORR are Fe−N species (most likely Fe3N). We have concluded that metallic Fe particles, Fe3C, or N−C species does not serve as active sites or play a key during the catalysis of the ORR. Furthermore, we have also excluded the Fe−N active site having any synergistic interactions with metallic Fe particles, Fe3C species, or N−C species during the catalysis of the ORR. Our results demonstrated that a high pyrolysis temperature favors the formation of the active sites, and the increase of the Fe content in the synthesis mixture can significantly enhance the amount of the Fe−N active species in the catalyst. Our study has confirmed that the introduction of Fe during the synthesis is critical for high ORR activity. This shows that the formation of the catalytically active site is strongly influenced by the doping heteroatoms and in turn that the ORR activity of the catalysts is controlled by the number and chemical nature active sites, and while iron is present in the catalytically active site and may also serve as a mediator to generate the active site during the synthetic process. J


Article


(10) Wu, P.; Zhang, H.; Qian, Y.; Hu, Y.; Zhang, H.; Cai, C. Composition- and Aspect-Ratio-Dependent Electrocatalytic Performances of One-Dimensional Aligned Pt−Ni Nanostructures. J. Phys. Chem. C 2013, 117, 19091−19100. (11) Hu, Y.; Zhang, H.; Wu, P.; Zhang, H.; Zhou, B.; Cai, C. Bimetallic Pt-Au Nanocatalysts Electrochemically Deposited on Graphene and Their Electrocatalytic Characteristics towards Oxygen Reduction and Methanol Oxidation. Phys. Chem. Chem. Phys. 2011, 13, 4083−4094. (12) Tan, X.; Prabhudev, S.; Kohandehghan, A.; Karpuzov, D.; Botton, G. A.; Mitlin, D. Pt−Au−Co Alloy Electrocatalysts Demonstrating Enhanced Activity and Durability toward the Oxygen Reduction Reaction. ACS Catal. 2015, 5, 1513−1524. (13) Zhu, H.; Zhang, S.; Guo, S.; Su, D.; Sun, S. Synthetic Control of FePtM Nanorods (M = Cu, Ni) To Enhance the Oxygen Reduction Reaction. J. Am. Chem. Soc. 2013, 135, 7130−7133. (14) Liu, L.; Pippel, E. Low-Platinum-Content Quaternary PtCuCoNi Nanotubes with Markedly Enhanced Oxygen Reduction Activity. Angew. Chem., Int. Ed. 2011, 50, 2729−2733. (15) Lin, S.-P.; Wang, K.-W.; Liu, C.-W.; Chen, H.-S.; Wang, J.-H. Trends of Oxygen Reduction Reaction on Platinum Alloys: A Computational and Experimental Study. J. Phys. Chem. C 2015, 119, 15224−15231. (16) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Improved Oxygen Reduction Activity on Pt3Ni(111) Via Increased Surface Site Availability. Science 2007, 315, 493−497. (17) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.; Lucas, C. A.; Wang, G.; Ross, P. N.; Markovic, N. M. Trends in Electrocatalysis on Extended and Nanoscale Pt-Bimetallic Alloy Surfaces. Nat. Mater. 2007, 6, 241−247. (18) Heggen, M.; Oezaslan, M.; Houben, L.; Strasser, P. Formation and Analysis of Core-Shell Fine Structures in Pt Bimetallic Nanoparticle Fuel Cell Electrocatalysts. J. Phys. Chem. C 2012, 116, 19073−19083. (19) Snyder, J.; McCue, I.; Livi, K.; Erlebacher, J. Structure/ Processing/Properties Relationships in Nanoporous Nanoparticles as Applied to Catalysis of the Cathodic Oxygen Reduction Reaction. J. Am. Chem. Soc. 2012, 134, 8633−8645. (20) Koenigsmann, C.; Scofield, M. E.; Liu, H.; Wong, S. S. Designing Enhanced One-Dimensional Electrocatalysts for the Oxygen Reduction Reaction: Probing Size- and Composition-Dependent Electrocatalytic Behavior in Noble Metal Nanowires. J. Phys. Chem. Lett. 2012, 3, 3385−3398. (21) Li, H.-H.; Ma, S.-Y.; Fu, Q.-Q.; Liu, X.-J.; Wu, L.; Yu, S.-H. Scalable Bromide-Triggered Synthesis of Pd@Pt Core-Shell Ultrathin Nanowires with Enhanced Electrocatalytic Performance toward Oxygen Reduction Reaction. J. Am. Chem. Soc. 2015, 137, 7862−7868. (22) Zhang, J. L.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adzic, R. R. Controlling the Catalytic Activity of Platinum-Monolayer Electrocatalysts for Oxygen Reduction with Different Substrates. Angew. Chem., Int. Ed. 2005, 44, 2132−2135. (23) Zhang, J.; Vukmirovic, M. B.; Sasaki, K.; Nilekar, A. U.; Mavrikakis, M.; Adzic, R. R. Mix-Metal Pt Monolayer Electrocatalysts for Enhanced Oxygen Reduction Kinetics. J. Am. Chem. Soc. 2005, 127, 12480−12481. (24) Strasser, P.; Koh, S.; Anniyev, T.; Greeley, J.; More, K.; Yu, C.; Liu, Z.; Kaya, S.; Nordlund, D.; Ogasawara, H.; et al. Lattice-Strain Control of the Activity in Dealloyed Core-Shell Fuel Cell Catalysts. Nat. Chem. 2010, 2, 454−460. (25) Gan, L.; Heggen, M.; Cui, C.; Strasser, P. Thermal Facet Healing of Concave Octahedral Pt−Ni Nanoparticles Imaged in Situ at the Atomic Scale: Implications for the Rational Synthesis of Durable High-Performance ORR Electrocatalysts. ACS Catal. 2016, 6, 692− 695. (26) Jaouen, F.; Proietti, E.; Lefèvre, M.; Chenitz, R.; Dodelet, J.-P.; Wu, G.; Chung, H. T.; Johnston, C. M.; Zelenay, P. Recent Advances in Non-Precious Metal Catalysis for Oxygen-Reduction Reaction in Polymer Electrolyte Fuel Cells. Energy Environ. Sci. 2011, 4, 114−130.

In conclusion, a new class of ORR catalyst has been made with heteroatom doped carbon with non-noble metal iron. New synthetic control of parameters have been developed, which affect the final structure and catalyst performance and allow the rational design and integration of two or more heteroatoms that may lead to even more new catalyst materials with tunable catalytic properties.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.6b02670. HRTEM images, CV responses in 0.1 KOH solution, ORR polarization curves at different rotating speeds and their related Koutecky−Levich plot, nitrogen sorption isotherms, XRD, and XPS spectra of the synthesized catalysts (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Tel: 86 25 85891780. *E-mail: [email protected]. Tel: 86 25 85891780. *E-mail: [email protected]. Tel: 1 520 621 6044. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is supported by NSFC (21175067, 21273117, 21375063, 21335004, and 21405083), NSF of the Jiangsu Higher Education Institutions (14KJB150012), and Priority Academic Program Development of Jiangsu Higher Education Institutions.

■

REFERENCES

(1) Wagner, F. T.; Lakshmanan, B.; Mathias, M. F. Electrochemistry and the Future of the Automobile. J. Phys. Chem. Lett. 2010, 1, 2204− 2219. (2) Debe, M. K. Electrocatalyst Approaches and Challenges for Automotive Fuel Cells. Nature 2012, 486, 43−51. (3) Guo, S.; Zhang, S.; Sun, S. Tuning Nanoparticle Catalysis for the Oxygen Reduction Reaction. Angew. Chem., Int. Ed. 2013, 52, 8526− 8544. (4) Nie, Y.; Li, L.; Wei, Z. Recent Advancements in Pt and Pt-Free Catalysts for Oxygen Reduction Reaction. Chem. Soc. Rev. 2015, 44, 2168−2201. (5) Morozan, A.; Jousselme, B.; Palacin, S. Low-Platinum and Platinum-Free Catalysts for the Oxygen Reduction Reaction at Fuel Cell Cathodes. Energy Environ. Sci. 2011, 4, 1238−1254. (6) Xia, W.; Mahmood, A.; Liang, Z.; Zou, R.; Guo, S. EarthAbundant Nanomaterials for Oxygen Reduction. Angew. Chem., Int. Ed. 2016, 55, 2650−2676. (7) Reddington, E.; Sapienza, A.; Gurau, B.; Viswanathan, R.; Sarangapani, S.; Smotkin, E. S.; Mallouk, T. E. Combinatorial Electrochemistry: A Highly Parallel, Optical Screening Method for Discovery of Better Electrocatalysts. Science 1998, 280, 1735−1737. (8) Li, Y.; Li, Y.; Zhu, E.; McLouth, T.; Chiu, C.-Y.; Huang, X. Stabilization of High-Performance Oxygen Reduction Reaction Pt Electrocatalyst Supporting on Reduced Graphene Oxide/Carbon Black Composite. J. Am. Chem. Soc. 2012, 134, 12326−12329. (9) Wang, X.; Vara, M.; Luo, M.; Huang, H.; Ruditskiy, A.; Park, J.; Bao, S.; Liu, J.; Howe, J.; Chi, M.; et al. Pd@Pt Core-Shell Concave Decahedra: A Class of Catalysts for the Oxygen Reduction Reaction with Enhanced Activity and Durability. J. Am. Chem. Soc. 2015, 137, 15036−15042. K


Article

The Journal of Physical Chemistry C (27) Jasinski, R. A New Fuel Cell Cathode Catalyst. Nature 1964, 201, 1212−1213. (28) Jahnke, H.; Schönborn, M.; Zimmermann, G. Organic Dyestuffs as Catalysts for Fuel Cells. Top. Curr. Chem. 1976, 61, 133−181. (29) Bagotzky, V. S.; Tarasevich, M. R.; Radyushkina, K. A.; Levina, O. A.; Andrusyova, S. I. Electrocatalysis of the Oxygen Reduction Process on Metal Chelates in Acid Electrolyte. J. Power Sources 1978, 2, 233−240. (30) Gupta, S.; Tryk, D.; Bae, I.; Aldred, W.; Yeager, E. Heat-Treated Polyacrylonitrile-Based Catalysts for Oxygen Electroreduction. J. Appl. Electrochem. 1989, 19, 19−27. (31) Lefèvre, M.; Proietti, E.; Jaouen, F.; Dodelet, J.-P. Iron-Based Catalysts with Improved Oxygen Reduction Activity in Polymer Electrolyte Fuel Cells. Science 2009, 324, 71−74. (32) Sahraie, N. R.; Paraknowitsch, J. P.; Göbel, G.; Thomas, A.; Strasser, P. Noble-Metal-Free Electrocatalysts with Enhanced ORR Performance by Task-Specific Functionalization of Carbon Using Ionic Liquid Precursor Systems. J. Am. Chem. Soc. 2014, 136, 14486−14497. (33) Wang, M.-Q.; Yang, W.-H.; Wang, H.-H.; Chen, C.; Zhou, Z.-Y.; Sun, S.-G. Pyrolyzed Fe−N−C Composite as an Efficient Nonprecious Metal Catalyst for Oxygen Reduction Reaction in Acidic Medium. ACS Catal. 2014, 4, 3928−3936. (34) Kramm, U. I.; Herrmann-Geppert, I.; Behrends, J.; Lips, K.; Fiechter, S.; Bogdanoff, P. On An Easy Way to Prepare MetalNitrogen Doped Carbon with Exclusive Presence of MeN4-Type Sites Active for the ORR. J. Am. Chem. Soc. 2016, 138, 635−640. (35) Tylus, U.; Jia, Q.; Strickland, K.; Ramaswamy, N.; Serov, A.; Atanassov, P.; Mukerjee, S. Elucidating Oxygen Reduction Active Sites in Pyrolyzed Metal-Nitrogen Coordinated Non-Precious-Metal Electrocatalyst Systems. J. Phys. Chem. C 2014, 118, 8999−9008. (36) Singh, D.; Tian, J.; Mamtani, K.; King, J.; Miller, J. T.; Ozkan, U. S. A Comparison of N-Containing Carbon Nanostructures (CNx) and N-Coordinated Iron-Carbon Catalysts (FeNC) for the Oxygen Reduction Reaction in Acidic Media. J. Catal. 2014, 317, 30−43. (37) Byon, H. R.; Suntivich, J.; Shao-Horn, Y. Graphene-Based NonNoble-Metal Catalysts for Oxygen Reduction Reaction in Acid. Chem. Mater. 2011, 23, 3421−3428. (38) Zheng, Y.; Jiao, Y.; Chen, J.; Liu, J.; Liang, J.; Du, A.; Zhang, W.; Zhu, Z.; Smith, S. C.; Jaroniec, M.; et al. Nanoporous GraphiticC3N4@Carbon Metal-Free Electrocatalysts for Highly Efficient Oxygen Reduction. J. Am. Chem. Soc. 2011, 133, 20116−20119. (39) Lin, L.; Zhu, Q.; Xu, A.-W. Noble-Metal-Free Fe−N/C Catalyst for Highly Efficient Oxygen Reduction Reaction Under Both Alkaline and Acidic Conditions. J. Am. Chem. Soc. 2014, 136, 11027−11033. (40) Wang, Y.; Kong, A.; Chen, X.; Lin, Q.; Feng, P. Efficient Oxygen Electroreduction: Hierarchical Porous Fe−N-doped Hollow Carbon Nanoshells. ACS Catal. 2015, 5, 3887−3893. (41) Yang, L.; Su, Y.; Li, W.; Kan, X. Fe/N/C Electrocatalysts for Oxygen Reduction Reaction in PEM Fuel Cells Using Nitrogen-Rich Ligand as Precursor. J. Phys. Chem. C 2015, 119, 11311−11319. (42) Charreteur, F.; Jaouen, F.; Ruggeri, S.; Dodelet, J.-P. Fe/N/C Non-Precious Catalysts for PEM Fuel Cells: Influence of the Structural Parameters of Pristine Commercial Carbon Blacks on Their Activity for Oxygen Reduction. Electrochim. Acta 2008, 53, 2925−2938. (43) Sasan, K.; Kong, A.; Wang, Y.; Chengyu, M.; Zhai, Q.; Feng, P. From Hemoglobin to Porous N−S−Fe-Doped Carbon for Efficient Oxygen Electroreduction. J. Phys. Chem. C 2015, 119, 13545−13550. (44) Fan, X.; Peng, Z.; Ye, R.; Zhou, H.; Guo, X. M3C (M: Fe, Co, Ni) Nanocrystals Encased in Graphene Nanoribbons: An Active and Stable Bifunctional Electrocatalyst for Oxygen Reduction and Hydrogen Evolution Reactions. ACS Nano 2015, 9, 7407−7418. (45) Lefèvre, M.; Dodelet, J.-P.; Bertrand, P. O2 Reduction in PEM Fuel Cells: Activity and Active Site Structural Information for Catalysts Obtained by the Pyrolysis at High Temperature of Fe Precursors. J. Phys. Chem. B 2000, 104, 11238−11247. (46) Lefèvre, M.; Dodelet, J.-P.; Bertrand, P. Molecular Oxygen Reduction in PEM Fuel Cells: Evidence for the Simultaneous Presence of Two Active Sites in Fe-Based Catalysts. J. Phys. Chem. B 2002, 106, 8705−8713.

(47) Artyushkova, K.; Serov, A.; Rojas-Carbonell, S.; Atanassov, P. Chemistry of Multitudinous Active Sites for Oxygen Reduction Reaction in Transition Metal-Nitrogen-Carbon Electrocatalysts. J. Phys. Chem. C 2015, 119, 25917−25928. (48) Gong, K. P.; Du, F.; Xia, Z. H.; Durstock, M.; Dai, L. M. Nitrogen-Doped Carbon Nanotube Arrays with High Electrocatalytic Activity for Oxygen Reduction. Science 2009, 323, 760−764. (49) Yang, L.; Jiang, S.; Zhao, Y.; Zhu, L.; Chen, S.; Wang, X.; Wu, Q.; Ma, J.; Ma, Y.; Hu, Z. Boron-Doped Carbon Nanotubes as MetalFree Electrocatalysts for the Oxygen Reduction Reaction. Angew. Chem., Int. Ed. 2011, 50, 7132−7135. (50) Niwa, H.; Horiba, K.; Harada, Y.; Oshima, M.; Ikeda, T.; Terakura, K.; Ozaki, J.; Miyata, S. X-Ray Absorption Analysis of Nitrogen Contribution to Oxygen Reduction Reaction in Carbon Alloy Cathode Catalysts for Polymer Electrolyte Fuel Cells. J. Power Sources 2009, 187, 93−97. (51) Sidik, R. A.; Anderson, A. B.; Subramanian, N. P.; Kumaraguru, S. P.; Popov, B. N. O2 Reduction on Graphite and Nitrogen-Doped Graphite: Experiment and Theory. J. Phys. Chem. B 2006, 110, 1787− 1793. (52) Anderson, A. B.; Sidik, R. A. Oxygen Electroreduction on FeII and FeIII Coordinated to N4 Chelates. Reversible Potentials for the Intermediate Steps from Quantum Theory. J. Phys. Chem. B 2004, 108, 5031−5035. (53) Hu, Y.; Jensen, J. O.; Zhang, W.; Cleemann, L. N.; Xing, W.; Bjerrum, N. J.; Li, Q. Hollow Spheres of Iron Carbide Nanoparticles Encased in Graphitic Layers as Oxygen Reduction Catalysts. Angew. Chem., Int. Ed. 2014, 53, 3675−3679. (54) Zitolo, A.; Goellner, V.; Armel, V.; Sougrati, M.-T.; Mineva, T.; Stievano, L.; Fonda, E.; Jaouen, F. Identification of Catalytic Sites for Oxygen Reduction in Iron- and Nitrogen-Doped Graphene Materials. Nat. Mater. 2015, 14, 937−942. (55) Jiang, H.; Yao, Y.; Zhu, Y.; Liu, Y.; Su, Y.; Yang, X.; Li, C. Iron Carbide Nanoparticles Encapsulated in Mesoporous Fe−N-Doped Graphene-Like Carbon Hybrids as Efficient Bifunctional Oxygen Electrocatalysts. ACS Appl. Mater. Interfaces 2015, 7, 21511−21520. (56) Hu, Y.; Jensen, J. O.; Zhang, W.; Martin, S.; Chenitz, R.; Pan, C.; Xing, W.; Bjerrum, N. J.; Li, Q. Fe3N-based Oxygen Reduction Catalysts: Synthesis, Hollow Spherical Structures and Applications in Fuel Cells. J. Mater. Chem. A 2015, 3, 1752−1760. (57) Liao, Y.; Pan, K.; Wang, L.; Pan, Q.; Zhou, W.; Miao, X.; Jiang, B.; Tian, C.; Tian, G.; Wang, G.; et al. Facile Synthesis of HighCrystallinity Graphitic Carbon/Fe3C Nanocomposites as Counter Electrodes for High-Efficiency Dye-Sensitized Solar Cells. ACS Appl. Mater. Interfaces 2013, 5, 3663−3670. (58) Zhang, Y.; Jiang, W.-J.; Guo, L.; Zhang, X.; Hu, J.-S.; Wei, Z.; Wan, L.-J. Confining Iron Carbide Nanocrystals inside CNx@CNT toward an Efficient Electrocatalyst for Oxygen Reduction Reaction. ACS Appl. Mater. Interfaces 2015, 7, 11508−11515. (59) Parvez, K.; Yang, S.; Hernandez, Y.; Winter, A.; Turchanin, A.; Feng, X.; Müllen, K. M. Nitrogen-Doped Graphene and Its Iron-Based Composite as Efficient Electrocatalysts for Reduction Reaction. ACS Nano 2012, 6, 9541−9550. (60) Wu, P.; Qian, Y.; Du, P.; Zhang, H.; Cai, C. Facile Synthesis of Nitrogen-Doped Graphene for Measuring the Releasing Process of Hydrogen Peroxide from Living Cells. J. Mater. Chem. 2012, 22, 6402−6412. (61) Bao, X. G.; Nie, X. W.; von Deak, D.; Biddinger, E.; Luo, W. J.; Asthagiri, A.; Ozkan, U.; Hadad, C. A First-Principles Study of the Role of Quaternary-N Doping on the Oxygen Reduction Reaction Activity and Selectivity of Graphene Edge Sites. Top. Catal. 2013, 56, 1623−1633. (62) Kothandaraman, R.; Nallathambi, V.; Artyushkova, K.; Barton, S. C. Non-Precious Oxygen Reduction Catalysts Prepared by HighPressure Pyrolysis for Low-Temperature Fuel Cells. Appl. Catal., B 2009, 92, 209−216. (63) Artyushkova, K.; Kiefer, B.; Halevi, B.; Knop-Gericke, A.; Schlogl, R.; Atanassov, P. Density Functional Theory Calculations of L


Article

The Journal of Physical Chemistry C Xps Binding Energy Shift for Nitrogen-Containing Graphene-Like Structures. Chem. Commun. 2013, 49, 2539−2541. (64) Higgins, D. C.; Hoque, M. A.; Hassan, F.; Choi, J.-Y.; Kim, B.; Chen, Z. Oxygen Reduction on Graphene-Carbon Nanotube Composites Doped Sequentially with Nitrogen and Sulfur. ACS Catal. 2014, 4, 2734−2740. (65) Wu, G.; More, K. L.; Johnston, C. M.; Zelenay, P. HighPerformance Electrocatalysts for Oxygen Reduction Derived from Polyaniline, Iron, and Cobalt. Science 2011, 332, 443−447.

M


Chemical Nature of Catalytic Active Sites for the Oxygen Reduction

Recommend Documents