Atomic Mechanism of Electrocatalytically Active CoâN Complexes in

Research Article www.acsami.org

Atomic Mechanism of Electrocatalytically Active Co−N Complexes in Graphene Basal Plane for Oxygen Reduction Reaction Feng Li,† Haibo Shu,*,†,‡ Chenli Hu,† Zhaoyi Shi,† Xintong Liu,† Pei Liang,† and Xiaoshuang Chen‡ †

College of Optical and Electronic Technology, China Jiliang University, 310018 Hangzhou, China National Laboratory for Infrared Physics, Shanghai Institute of Technical Physics, Chinese Academy of Science, 200083 Shanghai, China

‡

S Supporting Information *

ABSTRACT: Superior catalytic activity and high chemical stability of inexpensive electrocatalysts for the oxygen reduction reaction (ORR) are crucial to the large-scale practical application of fuel cells. The nonprecious metal/N modified graphene electrocatalysts are regarded as one of potential candidates, and the further enhancement of their catalytic activity depends on improving active reaction sites at not only graphene edges but also its basal plane. Herein, the ORR mechanism and reaction pathways of Co−N codoping onto the graphene basal plane have been studied by using firstprinciples calculations and ab initio molecular dynamics simulations. Compared to singly N-doped and Co-doped graphenes, the Co−N codoped graphene surface exhibits superior ORR activity and the selectivity toward a four-electron reduction pathway. The result originates from catalytic sites of the graphene surface being modified by the hybridization between Co 3d states and N 2p states, resulting in the catalyst with a moderate binding ability to oxygenated intermediates. Hence, introducing the Co−N4 complex onto the graphene basal plane facilitates the activation of O2 dissociation and the desorption of H2O during the ORR, which is responsible for the electrocatalyst with a smaller ORR overpotential (∼1.0 eV) that is lower than that of Co-doped graphene by 0.93 eV. Our results suggest that the Co−N co-doped graphene is able to compete against platinum-based electrocatalysts, and the greater efficient electrocatalysts can be realized by carefully optimizing the coupling between transition metal and nonmetallic dopants in the graphene basal plane. KEYWORDS: graphene, doping, electrocatalyst, oxygen reduction reaction, active site, density functional theory

1. INTRODUCTION Fuel cells, as a highly efficient and clean power source, have received great attention in the past decade. The realization of large-scale application of fuel cells depends on the development of highly efficient and low-cost electrocatalysts for use in electrochemical energy conversion process, in particular for the cathodic oxygen reduction reaction (ORR).1−4 Platinum (Pt)based materials are the most efficient catalysts for ORR toward the four-electron transfer pathway (O2 + 4H+ + 4e− → 2H2O) discovered so far.5,6 However, high cost and scarcity of Pt-based catalysts are the main obstacles to the widespread commercialization of fuel cells.7,8 Therefore, it is strongly desired to develop cost-effective electrocatalytic materials for the replacement of Pt-based fuel cell catalysts. Heteroatom-doped carbon-based nanostructures, such as carbon nanoparticles,9 carbon nanotubes,2 and graphene,10,11 were regarded as one of the possible alternatives to Pt-based catalysts due to their lower price, greater abundance, and high stability. Among these candidates, heteroatom-doped graphenerelated materials have attracted considerable interest because of their outstanding properties, including large surface area, high carrier mobility, excellent mechanical flexibility, and superior chemical stability. Therefore, there are numerous reports on © XXXX American Chemical Society

graphene doped with various heteroatoms such as nitrogen (N),10,12 phosphorus (P),11,13 boron (B),14 sulfur (S),15 selenium (Se),16 antimony (Sb),17 and their mixtures.18,19 These studies have demonstrated that the doping of nonmetal atoms in graphene is an effective way to tune their electronic structures and consequently to enhance their ORR activity. Especially, the dual-doped graphene nanosheets, such as B−N and S−N co-doping in graphene,18,19 exhibit superior ORR activity and stability as compared to the doping of single nonmetal elements onto graphene, which is ascribed to a synergistic coupling effect between heteroatoms. Meanwhile, they have also found that the ORR activity of graphene-based catalysts is strongly correlated with the number of exposed doping edge sites.19−21 Hence, a prerequisite for an active graphene-based ORR catalyst is to maximize the number of exposed edge doping sites relative to surface sites in the graphene basal plane. But thermodynamics favors the formation of the graphene basal plane, thus limiting the number of active sites in graphene-based catalysts. Received: September 29, 2015 Accepted: November 13, 2015

A

DOI: 10.1021/acsami.5b09169 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

(AIMD) simulations. In order to understand the role of Co−N complexes in enhancing the ORR activity of graphene electrocatalysts, the oxygen reduction on N-doped and Codoped graphene surfaces have also been taken into account in our simulations. We find that a Co−N co-doped graphene basal plane exhibits high stability and ORR activity as compared to N-doped and Co-doped graphene surfaces. The AIMD simulations demonstrate that the introduction of a Co−N4 complex into the graphene surface facilitates the ORR toward a four-electron transfer pathway. Based on the analysis of electronic structures, the mechanism of an electrocatalytically active Co−N4 complex in graphene has been revealed.

To improve the activity of graphene-based catalysts, one feasible way is to make full use of surface sites in the graphene basal plane. However, surface sites of graphene incorporated with nonmetal dopants are not enough active for the ORR relative to the edge sites due to the weak electronegativity of nonmetal dopants.22 For instance, N dopants in the graphene basal plane exhibit weaker ability for the adsorption and activation of O2 molecule than those at the graphene edges during the ORR.23 Moreover, lower active nonmetal dopants on the graphene surface will lead to the ORR toward the twoelectron pathway with the formation of H2O2 intermediate (O2 + 2H+ + 2e− → H2O2). This is because the four-electron reduction pathway requires relatively active catalytic sites to activate O2 molecules, resulting in the O−O bond breaking and the formation of the O−H bond. On the other hand, most of these metal-free graphene-based catalysts are highly efficient only in alkaline medium, but they show poor ORR activity in acid electrolyte.19,24 It is a well-known fact that a fuel cell working in an acidic medium is of practical significance due to the high CO2 tolerance during the ORR at the cathode.25,26 Recent studies have found that introducing active nonprecious transition metal (e.g., Fe, Co, and Ni) atoms or nanoparticles into N-doped graphene and carbon nanostructures could induce the electrocatalysts with high ORR activity and durability in acid aqueous environment.27−36 For example, Wu et al.27 reported high-performance carbon−nitrogen catalysts incorporated with Fe and Co. The catalysts exhibit remarkable performance stability for 700 h at a fuel cell voltage of 0.4 V and excellent four-electron selectivity. Zhu et al.28 synthesized Fe/N-modified carbon electocatalysts with high ORR activity and good durability in acid electrolytes. Kong et al.29 prepared ordered hierarchically micro- and mesoporous Fe−Nx embedded graphite architectures. They found that the catalysts presented good ORR selectivity and high durability in both alkaline and acidic media. Jiang et al.30 reported the ORR performance of Co/N co-modified graphene materials that was comparable to commercial Pt/C electrocatalysts. Very recently, Zitolo et al.36 reported the synthesis of Fe−N co-doped graphene materials, and the structures of active sites in Fe−N− C catalysts have been identified. These studies uniformly demonstrated the possibility for the nonprecious metal/N modified graphene and carbon nanomaterials as alternatives to Pt-based fuel cell electrocatalysts. However, current knowledge about the ORR theory on transition metal (TM)/N doped graphene surfaces is very limited,37−39 and the fundamental questions about electrocatalytic activity and ORR mechanism in the catalysts have not been clarified, including the stability mechanism of dual-doped graphene catalysts, the relationship between the binding ability of the catalysts to oxygenated species and the ORR activity, and the oxygen−reduction difference between dual-doped and single-doped graphene catalysts. More importantly, the experimental measurement techniques, such as X-ray diffraction (XRD), scan electron microscopy (SEM), transmission electron microscopy (TEM), and ORR polarization curve, cannot provide a visualizing process to insight into the atomic-scale ORR mechanisms in graphene-based electrocatalysts. This adds the difficulty of developing effective methods to improve the ORR performance of the catalysts. In this work, we present a comprehensive theoretical study on electrocatalytic activity and dynamics of a Co−N co-doped graphene surface for the ORR by using density functional theory (DFT) calculations and ab initio molecular dynamics

2. COMPUTATIONAL DETAILS The doping models of graphene are created on the basis of an in-plane periodic graphene supercell structure consisting of 96 C atoms. Ndoped, Co-doped, and Co−N co-doped models are established by introducing N atom, Co atom, and Co−Nx complexes into the supercell, respectively. The possible configurations of three graphene doping models considered in the present study are shown in section S1 of the Supporting Information (SI). There are three N-doped graphene (N-G) configurations (SI Figure S1), including graphitic N, pyridinic N, and pyrrolic N, and three Co-doped graphene (Co−G) configurations (SI Figure S2), including Co embedded into a carbon single vacancy (Co−G-I), Co embedded into a carbon divacancy (Co−G-II), and Co adatom (Co−G-III), respectively. For the Co−Nx co-doping in graphene (Co−Nx−G, x = 1−4), 10 possible configurations have been considered (SI Figure S3), including three Co−N1−G, four Co−N2−G, two Co−N3−G, and one Co−N4−G structures. The stability of various doping models is evaluated by calculating their formation energies (Ef) as follows,

Ef = E T − E P + ΔnCμC − ΔnNμ N − ΔnCoμCo

(1)

where ET and EP are the energies of doped and pristine graphene sheets, respectively, μi is the chemical potential of atomic species i (i = C, N, and Co), and Δni is the difference of atomic species i in doped and pristine graphene. Here μC, μN, and μCo are referred to the energy of the C atom in graphene, half of the energy of an N2 molecule, and the energy of the Co atom in bulk, respectively. The reaction scheme of O2 reduction to H2O in acid aqueous environment follows two possible pathways as demonstrated in Figure 1. One is the four-electron pathway (path I; see eq 2) in which O2 is reduced directly to H2O without the formation of H2O2 intermediate and another is a series two-electron pathway (path II; see eqs 3 and 4) in which O2 is reduced to H2O via H2O2.40 O2 + 4(H+ + e−) → 2H 2O

(2)

O2 + 2(H+ + e−) → 2H 2O2

(3)

H 2O2 + 2(H+ + e−) → 2H 2O

(4)

For the four-electron pathway, there exist two different reaction mechanisms (path I1 and path I2; see Figure 1). Path I1 requires first the breakage of an O−O bond of an adsorbed O2 molecule and then the transfer of proton/electron (H+ + e−), while path I2 has the reverse process. The free-energy diagrams in the reaction pathways were calculated on the basis of a computational hydrogen electrode (CHE) model that was introduced by Nørskov et al.3 In this model, the adsorption energies of O2 and intermediates are calculated first by the relation ΔE = ET − Eslab − Eadsorbates, where ET, Eslab, and Eadsorbates are the total energies of an adsorbed system, an isolated catalytic substrate, and a gas-phase adsorbate, respectively. Then these adsorption energies ΔE are converted into Gibbs free energies with the relationship3

ΔG = ΔE + ΔZPE − T ΔS + ΔGU

(5)

where ΔE is obtained directly from DFT calculations, ΔZPE is the change of zero-point energies (ZPE), T is temperature which refers to B


Research Article


three electrocatalysts as their initial structures of ORR, respectively. The introduction of surface H atoms is used to add the probability of oxygen reduction during the ORR. AIMD simulations were performed in the canonical (NVT) using 5000 time steps with a 1.0 fs time step at 300 K. As it is hugely time-consuming to run an AIMD simulation for a large system (there are about 100 atoms), a 2 × 2 × 1 k-point mesh is used in the calculations for the Brillouin-zone integration.

3. RESULTS AND DISCUSSION 3.1. Geometries and Stability. We have first investigated geometries and stability of N-doped, Co-doped, and Co−N codoped graphene sheets with the purpose of elucidating groundstate structures of three catalytic surfaces. The potential configurations of N−G, Co−G, and Co−Nx−G have been shown in SI Figures S1−S3, respectively. The stability of these potential catalytic surfaces is evaluated by calculating their formation energies, as listed in Table S1 of the SI. For N-doped graphene, the graphitic N (Figure 2a) is the most stable Figure 1. Schematic ORR pathways on a Co−N co-doped graphene surface in acid aqueous electrolytes. Paths I (I1 and I2) and II follow four-electron reduction and two-electron reduction pathways, respectively. The dashed frame indicates the initial step for the ORR, and atomic models in solid frames represent released molecules during the ORR. The Co, O, N, C, are H atoms are colored by purple, red, blue, brown, and cyan, respectively. Here asterisks (*) denote the adsorbed species on catalytic surfaces. room temperature (T = 298.15 K), and ΔS is the change in entropy. ΔGU is the contribution of electrode potential to ΔG, and it is computed on the basis of the assumption that the chemical potential of a proton/electron (H+ + e−) in solution is equal to half of the chemical potential of a gaseous H2 and shifting ΔG by − eU in each proton/ electron transfer step, where e is the number of electrons transferred and U is the applied electrode potential. The ZPE and entropies of ORR intermediates are calculated from the vibrational frequencies, and those of gas-phase molecules are obtained from the standard thermodynamic database.41 Details of free-energy calculations and DFT energies, ZPE, entropies of ORR intermediates, and molecules are shown in section S2 of the Supporting Information. All calculations in the present work are performed by using the DFT as implemented in the Vienna ab initio simulation package (VASP).42,43 The electronic exchange-correlation energy is treated by the spin-polarized generalized-gradient approximation (GGA) of Perdew−Burke−Ernzerhof.44 The electron−ion interactions are described by projector augmented wave (PAW) potentials.45 The kinetic energy cutoff for the plane-wave expansion is set to 400 eV, and the Monkhorst−Pack k-point mesh of 4 × 4 × 1 is found to provide sufficient accuracy in the integration of the Brillouin zone. For the geometry optimization, the convergence criteria of electronic and ionic iterations are 10−3 eV and 10−2 eV/Å, respectively. Because the GGA functional fails to describe partially occupied d and f electronic states of transition metal elements,46 thus the GGA with the correction of a Hubbard U term is used to compute electronic structures of graphene with the Co dopant. For the Hubbard parameter Ueff = U − J, we use a typical value of 5 eV that is proved a reasonable value to describe the correlation energy of localized Co 3d states on the basis of the careful tests of electronic structures and magnetic interactions, as shown in section S3 of the SI. The Bader charge analysis47,48 is used to investigate charge transfers between adsorbates and catalytic surfaces during the ORR. The climbing-image nudged elastic band (cNEB) method49 is employed to search the transition states of O2 and HOO dissociation on catalytic surfaces. In order to understand the ORR kinetics on N-doped, Co-doped, and Co−N co-doped graphene surfaces, ab initio molecular dynamics simulations have been carried out. In the calculations, we used the model with the size of 17.04 Å × 14.76 Å × 15 Å in which some adsorbed H atoms and a free O2 molecule were put on the surface of

Figure 2. Optimized geometries (upper panel) and simulated STM images (lower panel) of N-, Co-, and Co−N4-doped graphene surfaces. (a−c) Top views of the most stable atomic structures of (a) N−G, (b) Co−G, and (c) Co−N4−G. (d−f) Simulated STM images of the most stable (d) N−G, (e) Co−G, and (f) Co−N4−G at the voltages of −1.0 V using Tersoff−Hamann approximation.

configuration and its formation energy is at least 1.24 eV lower than other N-doped configurations. The result is in good agreement with extensive experimental observations50,51 in which N dopants prefer to occupy at the substitution sites of the graphene lattice. For Co-doped graphene, it is found that the Co adatom has the smallest formation energy. But, the Co adatom is unstable on the graphene surface due to the weak cobalt−graphene interaction, which easily induces the diffusion of Co adatoms and the formation of Co clusters on the graphene basal plane. Moreover, our calculation shows that the adsorption of the O2 molecule will induce the migration of a Co adatom out of the graphene surface. In other words, the Co adatom (Co−G-III) cannot become a stable catalytic site for the ORR. Instead, the Co atom in Co−G-I (Figure 2b) is strongly interacted with its neighboring C atoms, and it is difficult to be broken during the ORR. Moreover, its formation energy is lower than that of Co−G-II, and thus it is considered as the configuration of Co−G for the ORR. For Co−N codoping in graphene (Co−Nx−G), we find that Co−N4−G in which the Co atom is located at the center of four N atoms of graphene (Figure 2c) is the most stable configuration, and it has a smaller formation energy (1.32 eV) that is far lower than Co-doped graphene (∼4.14 eV). According to the doping concentration C ∝ exp(−Ef/(kBT)), where Ef, kB, and T are the formation energy, Boltzmann constant, and temperature, respectively, the lowest formation energy implies that the C


Research Article


Figure 3. (a) Reaction pathway of the reduction of O2 to H2O2 and (b) free-energy diagram of ORR on the N−G surface with two-electron (2e−) and four-electron (4e−) reduction processes. The images of 1−4 indicate optimized atomic structures of the reaction steps. TS is the transition state for dissociating HOO* to HO* + O*on the N−G surface. The numbers of horizontal axis n(m) in panel b denote the number of transferred electrons and image number of reaction steps, respectively.

Figure 4. (a) Reaction pathway and (b) free-energy diagram of ORR on the Co−G surface with the 4e− reduction process. Images of 1−7 indicate optimized atomic structures of reaction steps. Red and blue lines in panel b represent reactions at zero electrode potential (U = 0 V) and the equilibrium potential (U = 1.23 V), respectively. TS is the transition state of O2 dissociation on the Co−G surface.

N4−G surfaces, the oxygen reduction on the N−G surface is investigated first based on the CHE model.3 As demonstrated in Figure 1, the ORR includes a two-electron reduction pathway that induces the production of H2O2 intermediate and four-electron reduction pathway that leads to the production of H2O in an acid environment. Both of the two reduction pathways have been considered to obtain the optimum one. Figure 3a and SI Figure S6 show the reaction pathways of ORR on the N−G surface with two-electron and four-electron reduction processes, respectively. In the first three steps, the process is the same for both of the two pathways. In the initial step (1 → 2 in Figure 3a), an O2 molecule is adsorbed on the N−G surface but the O2 molecule is not strongly interacted with the surface due to the weak surface activity of N−G. The calculated adsorption energy of O2 on the N−G surface is only −0.13 eV. Undergoing the first (H+ + e−) transfer step (2 → 3) leads to the formation of a HOO* intermediate; here an asterisk (*) denotes that the oxygenated species is adsorbed on the N−G surface. It can be found that HOO* binds with an active C atom adjacent to the N dopant, but its adsorption energy is not large enough (−0.39 eV). Accordingly, the HOO* species is easily released from the N−G surface. In the second (H+ + e−) transfer step (3 → 4), HOO* has the chance to form a H2O2 molecule or an O* by releasing a H2O molecule on N−G (O* + H2O; see SI Figure S6). The latter reaction needs to break the O−O bond of HOO*; thus it requires a large extra energy in the reduction process. The dissociation barrier of O2 on the N−G surface is 1.20 eV from

Co−N co-doping can induce a high dopant density in the graphene basal plane as compared to the singly Co-doped case. Although the Co−N2 complex in graphene has also been reported as the catalyst for the ORR,39 its formation energy (∼3.14 eV) is far larger than that of the Co−N4 complex. Thus, the Co−N4−G is considered as the ground-state structure of Co−N co-doped graphene for the ORR. In order to further identify atomic structures of three catalytic surfaces, the STM images of N−G, Co−G, and Co− N4−G are simulated. As reported previously,50 the graphite N dopant in the graphene surface is indicated by dark areas in STM images and the nearest-neighboring C atoms appear bright (Figure 2d) due to an increase of electronic density of states. The STM image of Co−G (Figure 2e) exhibits a triangular bright spot that originates from unpaired electrons of Co delocalized to its adjacent C atoms. A similar phenomenon was also observed in B-doped graphene.52 The comparison of STM images between the N−G and the Co−G indicates that the dopant image of Co−G is more obvious and bright, implying that the Co−G may have higher surface activity than the N−G due to high charge distribution around the Co dopant. For the Co−N4−G, it can be found that the bright spots not only appear at the center Co dopant but also its neighboring two pentagons and two hexagons (Figure 2f), suggesting the charges transfer between the Co atom and its adjacent N and C atoms. 3.2. ORR Pathways and Free-Energy Diagrams. To provide a reference for studying the ORR on Co−G and Co− D


Research Article


Figure 5. (a) Reaction pathway of oxygen reduction on the Co−N4−G surface with the 4e− reduction process. Images 1−6 indicate optimized atomic structures in each reaction step. (b) Possible initial and final structures in the second (H++e−) transfer step. The inset refers to an intermediate structure during the optimization. (c) Free-energy diagram of ORR on the Co−N4−G surface. Red and blue lines represent reactions at zero electrode potential (U = 0 V) and the equilibrium potential (U = 1.23 V), respectively.

previous reports.53 In contrast, the production of H2O2 more easily happens due to its weak interaction with the surface. Hence, a series two-electron reduction mechanism should be energetically favorable for the oxygen reduction on the N−G surface. The free-energy diagram in Figure 3b shows that the ORR on the N-doped graphene surface is limited by the reduction of O2* to form HOO* that induces an energy increase of 0.79 eV. So the step from O2* to HOO* is the rate-limiting step during the ORR on the N−G surface. Since the HOO* is formed on the N−G surface, the rest of the reaction steps are energetically downhill for both the 2e− and 4e− reduction pathways. But the direct four-electron reduction pathway is kinetically unfavorable. Based on transition-state (TS) calculations, we find that it needs to overcome a barrier energy of 1.02 eV for the dissociation of HOO* to induce the step from HOO* to O*, while there is nearly no barrier for activating HOO* to produce H2O2 following the 2e− reduction pathway. The result suggests that the oxygen reduction on the N-doped graphene surface is totally different from that of N-doping in graphene edges in which the direct four-electron reduction pathway is more favorable.54 Once the H2O2 species is formed on the N−G surface, it is easily released from the catalytic surface due to the weak interaction between H2O2 and the N−G surface. Although there is the possibility to induce further reduction of H2O2 to H2O followed by pathway II shown in Figure 1, it needs to overcome a large energy barrier for breaking the O−O bond of the H2O2 molecule. Therefore, the reduction of H2O2 to H2O is relatively difficult and the formed H2O2 species can be considered as the ORR product on the N−G surface. Figure 4 shows the reaction pathway and free-energy diagram of ORR on the Co−G surface. Unlike the N−G surface, an adsorbed O2 molecule prefers to directly bind at the top of the Co dopant in the initial reaction step. The adsorption energy of O2 on the Co−G surface is −1.74 eV and is far larger than that of O2 on the N−G surface, which indicates a stronger bonding interaction between the adsorbed O2 (O2*) and the Co dopant. The Co−O bonding interaction weakens the O−O bond of O2*. We find that the O−O bond length of 1.38 Å on the Co−

G surface is larger than that of free O2 molecule (1.23 Å). The elongated O−O bond length implies the possibility to dissociate O2* into two O* (2 → 3; see Figure 4a). The transition-state calculation shows that the energy barrier of O2 dissociation is only 0.17 eV. The geometry optimization also indicates that the 2O* structure is more stable than the O2*. Once the 2O* is formed on Co−G, the first and second (H+ + e−) transfer steps (3 → 5) will not produce H2O2. In other words, O2 is easily activated by its strong interactions with the Co−G surface, leading to the ORR toward a direct fourelectron reduction pathway. To explore the oxygen−reduction activity of Co−G, the freeenergy diagram of ORR is calculated in Figure 4b. At the zero electrode potential (U = 0 V), free energies of all reaction steps are downhill except for the last step (7 → 8 in Figure 4a) associated with the H2O desorption. This endothermic step arises from the exceptional stability of HO* on the Co−G surface. The adsorption energy of HO* is −4.07 eV, indicating a strong interaction between HO* intermediate and the surface. Hence, the step for the reduction of HO* to H2O becomes an endothermic process. At the equilibrium potential (U = 1.23 V) that corresponds to the maximum thermodynamic potential of the fuel cell, free energies in all (H+ + e−) transfer steps (4 → 7) are uphill except for the first transfer step. The rate-limiting step lies in the last (H+ + e−) transfer with an overpotential of 1.93 eV (i.e., defined as the minimum energy required for the ORR on the surface). Obviously, the overpotential on the Co− G surface is large as compared to Pt-based catalysts.3,55 The preceding results indicate that the strong binding strength of ORR intermediates on the catalytic surface contributes to the O 2 activation in the initial reaction process but it thermodynamically hinders the formation of production species (e.g., H2O) in subsequent reactions. We now turn to investigate the ORR on the Co−N4−G surface. As shown in Figure 5a, O2 also prefers to bind on the top of the Co dopant with adsorption energy of −0.79 eV. The adsorption energy is 0.95 eV higher than that of O2* on Co−G but 0.66 eV lower than that of O2* on N−G. As discussed on the Co−G surface, the Co−O interaction weakens the O−O E


Research Article


Figure 6. Snapshots during the trajectories of AIMD for the oxygen reduction on (a) N−G, (b) Co−G, and (c) Co−N4−G surfaces at 300 K. The brown, cyan, red, blue, and pink balls represent C, H, O, N, and Co atoms, respectively.

3 and 5 → 6) are energetically uphill and the second and the third (H+ + e−) transfer steps (3 → 4 and 4 → 5) are energetically downhill. Unlike the ORR on Co−G, the ratelimiting step of oxygen reduction on Co−N4−G lies in the second (H+ + e−) transfer step (i.e., O2* → HOO*). The ratelimiting energy barrier (or overpotential) on the Co−N4−G surface is 1.00 eV that is far lower than that of Co−G, even lower than that of graphene-supported and unsupported Pt13 clusters (1.13−1.68 eV).55 This result suggests that the activity of graphene basal plane modified by the Co−N complex is responsible for the remarkable improvement of the rate-limiting energy barrier. 3.3. ORR Kinetics. Although the oxygen reduction on Co− N4−G is thermodynamically more favorable than that of the N−G and Co−G surfaces, we want to know whether the ORR on the Co−N4−G surface is also kinetically preferred. To shed light on the ORR kinetics of these graphene-based electrocatalysts, AIMD simulations were carried out at 300 K within 5 ps. Figure 6 presents a few snapshots of trajectories for the oxygen reduction on N−G, Co−G, and Co−N4−G surfaces. On the N−G surface, we introduced three N dopants into the graphene supercell in order to add the oxygen reduction probability. Nevertheless, the AIMD results indicate that the O2 molecule does not directly adsorb on the surface in the initial 2 ps (see Figure 6a), indicating a weak interaction of O2 with the surface. When the reaction time reaches 3 ps, an HOO intermediate is found but it still keeps a gaseous state in a finite time (5 ps). The formation of free HOO promotes the production of H2O2 due to the lack of O−O bond breaking, which agrees with the deduction of the free-energy diagram that the weak surface activity of N−G facilitates a two-electron reduction mechanism. On the Co−G surface, adsorption, bonding breaking, and rebonding of O2 molecule on top of the Co dopant can be found during the ORR (Figure 6b). Owing to the strong Co−O interaction, the further oxygen reduction process (e.g., 2O* → O* + HO*) has not been observed in a short reaction time (5 ps). In contrast, the adsorption of O2 molecule on the Co−N4−G surface binds with the Co dopant in the initial stage, then it transforms into HOO* by combining with an H atom in 2 ps, and finally a H2O molecule is released

bond of O2*, leading to an increase of O−O bond length. We find that the O−O bond length of O2* on Co−N4−G is 1.30 Å which lies between that of O2* on N−G and Co−G surfaces. So, Co−N4−G has a medium binding ability to activate O2 molecule as compared to N−G and Co−G surfaces. However, whether the O2* dissociation on the Co−N4−G surface priors to the (H+ + e−) transfer steps as appeared on Co−G surface. The energy comparison indicates that 2O* is 1.95 eV higher than O2*, suggesting that use of the medium-active Co−N4−G surface is difficult to directly dissociate O2* into 2O*. Therefore, the addition of H to the O2* in the first (H+ + e−) transfer step (2 → 3 in Figure 5a) induces the formation of HOO* on the surface. In the second (H+ + e−) transfer step, there are two possible reaction pathways (Figure 5b): H atom is attached at (i) the O site adjacent the Co dopant and (ii) the O site apart from the Co dopant, respectively. Interestingly, geometry-optimization results indicate that the addition of H atom leads to the formation of O* and desorption of a H2O molecule from the surface no matter which pathway the oxygen reduction follows. Such a result is different from the recent report from Kattel et al.39 in which the Co−N4 complex in graphene surface facilitates a two-electron reduction mechanism to form a peroxide intermediate. There are two essential preconditions for the production of H2O2 during the ORR: one is the stable H2O2 intermediate that is energetically more favorable than other intermediates, and the other is the weak surface activity of the catalyst that ensures the H2O2 desorption. To examine the possibility for the formation of H2O2 on the Co−N4−G surface, we performed a full geometry optimization for a H2O2 on the surface. However, the H2O2 on the Co−N4−G surface is unstable, and it spontaneously dissociates into 2HO* after the structural optimization (see Figure 5b). Moreover, the energy of 2HO* is 0.53 eV higher than that of the O* + H2O structure. Therefore, the Co−N4−G surface should facilitate the ORR toward the four-electron reduction mechanism. The free-energy diagram in Figure 5c shows that the reaction pathway is overall high exothermic at zero electrode potential (U = 0 V). By shifting the potential to equilibrium potential (U = 1.23 V), the first and the fourth (H+ + e−) transfer steps (2 → F


Research Article


down Co 3d states shifted to lower energy (Figure 7c) as compared to the electronic states of Co 3d on the Co−G surface. Therefore, the overlap of Co 3d and O 2p states is mainly located in the spin-down channel and less p−d hybridization appears in the spin-up channel, which contributes to a moderate binding ability of Co−N4−G to O2. The remarkable difference between the binding strength of oxygenated intermediates on Co−G and Co−N4−G can be examined by the d-band theory.56,57 For a TM-doped surface, the position of its d-band center closer to the Fermi level causes antibonding states to a higher energy, potentially making them more difficult to be filled, thereby inducing a stronger binding interaction between the surface and the adsorbed oxygenated species. The calculated d-band centers of Co dopant on Co−G and Co−N4−G are −1.64 and −1.84 eV, respectively (The details for calculating d-band centers are shown in section S5 of the SI). The smaller d-band center on Co−N4−G means a relatively weak interaction of the surface with oxygenated species, which is responsible for faster kinetics for the ORR on Co−N4−G. In order to further understand the interactions between graphene doping surfaces and oxygenated species, the chargedensity differences of N−G, Co−G, and Co−N4−G with adsorbed O2*, HOO*, and HO* + O* have been calculated, as shown in Figure 8. Upon the adsorption of oxygenated species,

from the surface leading the formation of O* structure (the real-time movie is shown in the Supporting Information Movie S1). It supports the thermodynamic conclusion that the medium-active Co−N4−G follows a direct four-electron reduction mechanism and exhibits superior ORR activity as compared to Co−G and N−G surfaces. 3.4. Mechanism of High ORR Activity on the Co−N4− G Surface. Based on the aforementioned analysis, it is now clear that the ORR activity of these doped graphene electrocatalysts depends on the binding strength of oxygenated species. The Co−N4−G surface with a moderate binding ability to oxygenated species is more active than those binding with oxygenated species too strong (e.g., Co−G) or too weak (e.g., N−G). In order to understand the mechanism of ORR activity difference among the graphene doping structures, electronic structures of N−G, Co−G, and Co−N4−G with oxygenated species have been investigated. Figure 7 shows calculated partial

Figure 8. Charge-density differences of (a) N−G, Co−G, and Co− N4−G surfaces with the adsorption of O2 and (b) these catalytic surfaces with the adsorption of HOO (or HO* + O*). The charge accumulation and depletion are colored in yellow and cyan, respectively. The numbers indicate charge transfer from catalytic surfaces to adsorbed oxygenated species. The charge transfer ΔQ is calculated by comparing atomic Bader charges on catalytic surfaces before and after the adsorption of oxygenated species.

Figure 7. Calculated partial density of states of (a) N−G, (b) Co−G, and (c) Co−N4−G surfaces with the adsorption of O2. For C 2p states, only the C atom adjacent to the dopant (e.g., N or Co) is considered. The Fermi level is set to energy zero.

densities of states (PDOSs) of three catalytic surfaces with an adsorbed O2 by GGA + U, in which the U value is set to be 5 eV. To facilitate their comparison, PDOSs are plotted for the same energy window from −12 to 4 eV. The O2 adsorption on the N−G surface induces extremely localized states in both spin-up and spin-down channels, as shown in Figure 7a. There are few overlaps between O 2p states and electronic states of C and N atoms, indicating a weak interaction of adsorbed O2 with the surface. On the Co−G surface, the 2p states of adsorbed O2 are delocalized and have a strong hybridization with Co 3d states at −7 eV to the Fermi level (Figure 7b), which is responsible for a large binding strength of O2 and the activation of O2 dissociation on Co−G. In contrast, the p−d hybridization between Co and N dopants on the Co−N4−G surface causes the spin-up Co 3d states shifted to higher energy and the spin-

it can be clearly found the charge depletion (in cyan) around the Co dopant and the charge accumulation (in yellow) at O− O and Co−O bonds on Co−G and Co−N4−G, suggesting the charge transfer from these catalytic surfaces to oxygenated species. In contrast, there are few charge depletions on the N− G surface, in particular for the O2 adsorption. Based on Badercharges analysis,47,48 we find that the charge transfer ΔQ is 0.17 e, 0.64 e, and 0.41 e for the O2 adsorption on N−G, Co−G, and Co−N4−G surfaces, respectively. The ΔQ values display that the binding ability of Co−N4−G to an O2 should lie between Co−G and N−G. The medium binding ability of Co−N4−G to oxygenated species is also demonstrated after the first (H+ + e−) transfer step, as shown in Figure 8b. This originates from G


Research Article


(2) Gong, K.; Du, F.; Xia, Z.; Durstock, M.; Dai, L. Nitrogen-Doped Carbon Nanotube Arrays with High Electrocatalytic Activity for Oxygen Reduction. Science 2009, 323, 760−764. (3) Nørskov, J.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J.; Bligaard, T.; Jonsson, H. Origin of the Overpotential for Oxygen Reduction at a Fuel-Cell Cathode. J. Phys. Chem. B 2004, 108, 17886− 17892. (4) Jiao, Y.; Zheng, Y.; Jaroniec, M.; Qiao, S. Z. Design of Electrocatalysts for Oxygen- and Hydrogen-Involving Energy Conversion Reactions. Chem. Soc. Rev. 2015, 44, 2060−2086. (5) Greeley, J.; Stephens, I. E. L.; Bondarenko, A. S.; Johansson, T. P.; Hansen, H. A.; Jaramillo, T. F.; Rossmeisl, J.; Chorkendorff, I.; Nørskov, J. K. Alloys of Platinum and Early Transition Metals as Oxygen Reduction Electrocatalysts. Nat. Chem. 2009, 1, 552−556. (6) Shao, M.; Peles, A.; Shoemaker, K. Electrocatalysis on Platinum Nanoparticles: Particle Size Effect on Oxygen Reduction Reaction Activity. Nano Lett. 2011, 11, 3714−3719. (7) Yu, X.; Ye, S. Recent Advances in Activity and Durability Enhancement of Pt/C Catalytic Cathode in PEMFC: Part II: Degradation Mechanism and Durability Enhancement of Carbon Supported Platinum Catalyst. J. Power Sources 2007, 172, 145−154. (8) Wang, J. X.; Inada, H.; Wu, L.; Zhu, Y.; Choi, Y. M.; Liu, P.; Zhou, W. P.; Adzic, R. Oxygen Reduction on Well-Defined Core-Shell Nanocatalysts: Particle Size, Facet, and Pt Shell Thickness Effects. J. Am. Chem. Soc. 2009, 131, 17298−17302. (9) Xing, T.; Sunarso, J.; Yang, W.; Yin, Y.; Glushenkov, A. M.; Li, L. H.; Howlett, P. C.; Chen, Y. Ball Milling: A Green Mechanochemical Approach for Synthesis of Nitrogen Doped Carbon Nanoparticles. Nanoscale 2013, 5, 7970−7976. (10) Qu, L.; Liu, Y.; Baek, J.-B.; Dai, L. Nitrogen-Doped Graphene as Efficient Metal-Free Electrocatalyst for Oxygen Reduction in Fuel Cells. ACS Nano 2010, 4, 1321−1326. (11) Zhang, C.; Mahmood, N.; Yin, H.; Liu, F.; Hou, Y. Synthesis of Phosphorus-Doped Graphene and Its Multifunctional Applications for Oxygen Reduction and Lithium Ion Batteries. Adv. Mater. 2013, 25, 4932−4937. (12) Geng, D.; Chen, Y.; Chen, Y.; Li, Y.; Li, R.; Sun, X.; Ye, S.; Knights, S. High Oxygen-Reduction Activity and Durability of Nitrogen Doped Graphene. Energy Environ. Sci. 2011, 4, 760−764. (13) Yang, D. S.; Bhattacharjya, D.; Inamdar, S.; Park, J.; Yu, J. S. Phosphorus-Doped Ordered Mesoporous Carbons with Different Lengths as Efficient Metal-Free Electrocatalysts for Oxygen Reduction Reaction in Alkaline Media. J. Am. Chem. Soc. 2012, 134, 16127− 16130. (14) Sheng, Z.-H.; Gao, H.-L.; Bao, W.-J.; Wang, F.-B.; Xia, X.-H. Synthesis of Boron Doped Graphene for Oxygen Reduction Reaction in Fuel Cells. J. Mater. Chem. 2012, 22, 390−395. (15) Yang, Z.; Yao, Z.; Li, G. F.; Fang, G. Y.; Nie, H.; Liu, Z.; Zhou, X. M.; Chen, X.; Huang, S. M. Sulfur-Doped Graphene as an Efficient Metal-Free Cathode Catalyst for Oxygen Reduction. ACS Nano 2012, 6, 205−211. (16) Jin, Z.; Nie, H.; Yang, Z.; Zhang, J.; Liu, Z.; Xu, X.; Huang, S. Metal-Free Selenium Doped Carbon Nanotube/Graphene Networks as A Synergistically Improved Cathode Catalyst for Oxygen Reduction Reaction. Nanoscale 2012, 4, 6455−6460. (17) Jeon, I.-Y.; Choi, M.; Choi, H.-J.; Jung, S.-M.; Kim, M.-J.; Seo, J.M.; Bae, S.-Y.; Yoo, S.; Kim, G.; Jeong, H. Y.; Park, N.; Baek, J.-B. Antimony-Doped Graphene Nanoplatelets. Nat. Commun. 2015, 6, 7123. (18) Zheng, Y.; Jiao, Y.; Ge, L.; Jaroniec, M.; Qiao, S. Z. Two-Step Boron and Nitrogen Doping in Graphene for Enhanced Synergistic Catalysis. Angew. Chem., Int. Ed. 2013, 52, 3110−3116. (19) Liang, J.; Jiao, Y.; Jaroniec, M.; Qiao, S. Z. Sulfur and Nitrogen Dual-Doped Mesoporous Graphene Electrocatalyst for Oxygen Reduction with Synergistically Enhanced Performance. Angew. Chem., Int. Ed. 2012, 51, 11496−11500. (20) Gao, F.; Zhao, G.-L.; Yang, S. Catalytic Reactions on the OpenEdge Sites of Nitrogen-Doped Carbon Nanotubes As Cathode Catalyst for Hydrogen Fuel cells. ACS Catal. 2014, 4, 1267−1273.

the catalytic sites modified by the hybridization between Co 3d and N 2p states (Figure 7c). As discussed earlier, a moderate binding ability of the catalyst to oxygenated species facilitates the activation of O2 dissociation and the H2O desorption from the catalytic surface. Therefore, the further optimization for the coupling between TM (TM = Co, Fe, and Ni) and nonmetal elements (e.g., N, P, As, and S) may induce graphene-based electrocatalysts with an optimum surface activity to interact with oxygenated species, realizing high catalytic activity of the catalysts.

4. CONCLUSIONS In summary, we performed systematic theoretical investigations on the geometry structures and pathways, kinetics, and mechanism of ORR on N−G, Co−G, and Co−N4−G surfaces. Our results demonstrate that N-doped graphene with a weak surface activity has difficulty breaking the O−O bond of O2 and HOO*, thus facilitating a series two-electron reduction pathway and the formation of peroxide, while the strong binding ability of surface sites on Co−G to oxygenated species easily activates O2 but it brings a large ORR overpotential (1.93 eV). Introducing Co−N4 complex into the graphene basal plane strongly mediates its surface activity, which contributes to the activation of O2 dissociation and the H2O desorption from the catalytic surface. Hence, the ORR on the Co−N4−G surface has a smaller overpotential (∼1.00 eV) that is 0.93 eV smaller than that of the Co−G surface. The careful optimization for the coupling between transition-metal and nonmetallic dopants in the graphene basal plane can lead to the further improvement of ORR activity and durability of graphene-based electrocatalysts.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.5b09169. Models and energetics of graphene doping structures, details and the related data of free-energy diagram calculations, the test of the Hubbard U value for the Co 3d states, four-electron reduction pathway of ORR on the N−G surface, and calculations of d-band centers for Co−G and Co−N4−G (PDF) Movie of oxygen reduction on the Co−N4−G surface (MPG)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +86-0571-86875622. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported in part by the National Natural Science Foundation of China (Grant Nos. 11404309 and 11347109) and the Zhejiang Provincial Natural Science Foundation of China (Grant No. LQ13A040001). Computational resources from the Shanghai Supercomputer Center are acknowledged.

■

REFERENCES

(1) Wee, J. H. Applications of Proton Exchange Membrane Fuel Cell Systems. Renewable Sustainable Energy Rev. 2007, 11, 1720−1738. H


Research Article


(41) Cramer, C. J. Essentials of Computational Chemistry Theories and Models, 2nd ed.; John Wiley & Sons: West Sussex, England, 2004. (42) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953−17979. (43) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (44) Payne, M. C.; Teter, M. P.; Allan, D. C.; Arias, T. A.; Joannopoulos, J. D. Iterative Minimization Techniques for Ab Initio Total-Energy Calculations: Molecular Dynamics and Conjugate Gradients. Rev. Mod. Phys. 1992, 64, 1045−1097. (45) Kresse, G.; Joubert, D. From Ultrasoft Pseudopentials to the Projector Augmented-Wave Methods. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, 1758−1775. (46) Cococcioni, M.; de Gironcoli, S. Linear Response Approach to the Calculation of the Effective Interaction Parameters in the LDA+U Method. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 71, 035105. (47) Bader, R. F. W. A Quantum Theory of Molecular Structure and Its Applications. Chem. Rev. 1991, 91, 893−928. (48) Tang, W.; Sanville, E.; Henkelman, G. A Grid-Based Bader Analysis Algorithm without Lattice Bias. J. Phys.: Condens. Matter 2009, 21, 084204. (49) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901. (50) Zabet-Khosousi, A.; Zhao, L.; Pálová, L.; Hybertsen, M. S.; Reichman, D. R.; Pasupathy, A. N.; Flynn, G. W. Segregation of Sublattice Domains in Nitrogen-Doped Graphene. J. Am. Chem. Soc. 2014, 136, 1391−1397. (51) Tison, Y.; Lagoute, J.; Repain, V.; Chacon, C.; Girard, Y.; Rousset, S.; Joucken, F.; Sharma, D.; Henrard, L.; Amara, H.; Ghedjatti, A.; Ducastelle, F. Electronic Interaction between Nitrogen Atoms in Doped Graphene. ACS Nano 2015, 9, 670−678. (52) Zhao, L.; Levendorf, M.; Goncher, S.; Schiros, T.; Pálová, L.; Zabet-Khosousi, A.; Rim, K. T.; Gutiérrez, C.; Nordlund, D.; Jaye, C.; Hybertsen, M.; Reichman, D.; Flynn, G. W.; Park, J.; Pasupathy, A. N. Local Atomic and Electronic Structure of Boron Chemical Doping in Monolayer Graphene. Nano Lett. 2013, 13, 4659−4665. (53) Jiao, Y.; Zheng, Y.; Jaroniec, M.; Qiao, S. Z. Origin of the Electrocatalytic Oxygen Reduction Activity of Graphene-Based Catalysts: a Roadmap to Achieve the Best Performance. J. Am. Chem. Soc. 2014, 136, 4394−4403. (54) Zhang, L.; Xia, Z. Mechanisms of Oxygen Reduction Reaction on Nitrogen-Doped Graphene for Fuel Cells. J. Phys. Chem. C 2011, 115, 11170−11176. (55) Lim, D.-H.; Wilcox, J. Mechanisms of the Oxygen Reduction Reaction on Defective Graphene-Supported Pt Nanoparticles from First-Principles. J. Phys. Chem. C 2012, 116, 3653−3660. (56) Hammer, B.; Nørskov, J. K. Electronic Factors Determining the Reactivity of Metal Surfaces. Surf. Sci. 1995, 343, 211−220. (57) Hammer, B.; Norskov, J. K. Why Gold is the Noblest of All the Metals. Nature 1995, 376, 238−240.

(21) Kurak, K. A.; Anderson, A. B. Nitrogen-Treated Graphite and Oxygen Electroreduction on Pyridinic Edge Sites. J. Phys. Chem. C 2009, 113, 6730−6734. (22) Duan, J.; Chen, S.; Jaroniec, M.; Qiao, S. Z. Heteroatom-Doped Graphene-Based Materials for Energy-Relevant Electrocatalytic Processes. ACS Catal. 2015, 5, 5207−5234. (23) Xing, T.; Zheng, Y.; Li, L. H.; Cowie, B. C. C.; Gunzelmann, D.; Qiao, S. Z.; Huang, S. M.; Chen, Y. Obseveration of Active Sites for Oxygen Reduction Reaction on Nitrogen-Doped Multilayer Graphene. ACS Nano 2014, 8, 6856−6862. (24) Yu, D.; Zhang, Q.; Dai, L. Highly Efficient Metal-Free Growth of Nitrogen-Doped Single-Walled Carbon Nanotubes on PlasmaEtched Substrate for Oxygen Reduction. J. Am. Chem. Soc. 2010, 132, 15127−15129. (25) Naughton, M. S.; Brushett, F. R.; Kenis, P. J. A. Carbonate Resilience of Flowing Electrolyte-Based Alkaline Fuel Cells. J. Power Sources 2011, 196, 1762−1768. (26) Sun, T.; Wu, Q.; Che, R.; Bu, Y.; Jiang, Y.; Li, Y.; Yang, L.; Wang, X.; Hu, Z. Alloyed Co-Mo Nitride as High-Performance Electrocatalyst for Oxygen Reduction in Acidic Medium. ACS Catal. 2015, 5, 1857−1862. (27) 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. (28) Zhu, Y.; Zhang, B.; Liu, X.; Wang, D.-W.; Su, D.-S. Unravelling the Structure of Electrocatalytically Active Fe−N Complexes in Carbon for Oxygen Reduction Reaction. Angew. Chem., Int. Ed. 2014, 53, 10673−10677. (29) Kong, A.; Zhu, X.; Han, Z.; Yu, Y.; Zhang, Y.; Dong, B.; Shan, Y. Ordered Hierarchically Micro- and Mesoporous Fe−Nx Embedded Graphitic Architectures as Efficient Electrocatalysts for Oxygen Reduction Reaction. ACS Catal. 2014, 4, 1793−1800. (30) Jiang, S.; Zhu, C.; Dong, S. Cobalt and NitrogenCofunctionalized Graphene as a Durable Non-Precious Metal Catalyst with Enhanced ORR Activity. J. Mater. Chem. A 2013, 1, 3593−3599. (31) Liu, H.; Zheng, Y.; Wang, G. X.; Qiao, S. Z. A ThreeComponent Nanocomposite with Synergistic Reactivity for Oxygen Reduction Reaction in Alkaline Solution. Adv. Energy Mater. 2015, 5, 1401186. (32) Zhou, R. F.; Qiao, S. Z. Fe/N Co-Doped Graphitic Carbon Bulb for Highperformance Oxygen Reduction Reaction. Chem. Commun. 2015, 51, 7516−7519. (33) Liang, J.; Zhou, R. F.; Chen, X. M.; Tang, Y. H.; Qiao, S. Z. Fe− N Decorated Hybrids of CNTs Grown on Hierarchically Porous Carbon for High-Performance Oxygen Reduction. Adv. Mater. 2014, 26, 6074−6079. (34) Duan, J. J.; Chen, S.; Dai, S.; Qiao, S. Z. Shape Control of Mn3O4 Nanoparticles on Nitrogen-Doped Graphene for Enhanced Oxygen Reduction Activity. Adv. Funct. Mater. 2014, 24, 2072−2078. (35) Zhou, R.; Qiao, S. Z. Silver/Nitrogen-Doped Graphene Interaction and Its Effect on Electrocatalytic Oxygen Reduction. Chem. Mater. 2014, 26, 5868−5873. (36) 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. (37) Liang, W.; Chen, J.; Liu, Y.; Chen, S. Density-FunctionalTheory Calculation Analysis of Active Sites for Four-Electron Reduction of O2 on Fe/N-Doped Graphene. ACS Catal. 2014, 4, 4170−4177. (38) Kattel, S.; Atanassov, P.; Kiefer, B. A Density Functional Theory Study of Oxygen Reduction Reaction on Non-PGM Fe-Nx-C Electrocatalysts. Phys. Chem. Chem. Phys. 2014, 16, 13800−13806. (39) Kattel, S.; Atanassov, P.; Kiefer, B. Catalytic Activity of Co-Nx/ C Electrocatalysts for Oxygen Reduction Reaction: a Density Functional Theory Study. Phys. Chem. Chem. Phys. 2013, 15, 148−153. (40) Dai, L.; Xue, Y.; Qu, L.; Choi, H.-J.; Baek, J.-B. Metal-Free Catalysts for Oxygen Reduction Reaction. Chem. Rev. 2015, 115, 4823−4892. I


Atomic Mechanism of Electrocatalytically Active CoâN Complexes in

Recommend Documents

Atomic Mechanism of Electrocatalytically Active CoâN Complexes in