Enhanced Oxygen Reduction Reaction by In Situ Anchoring Fe2N Nanoparticles on Nitrogen-Doped Pomelo Peel-Derived Carbon

The development of effective oxygen electrode catalysts for renewable energy technologies such as metal-air batteries and fuel cells remains challenging. Here, we prepared a novel high-performance oxygen reduction reaction (ORR) catalyst comprised of Fe2N nanoparticles (NPs) in situ decorated over an N-doped porous carbon derived from pomelo peel (i.e., Fe2N/N-PPC). The decorated Fe2N NPs provided large quantities of Fe-N-C bonding catalytic sites. The as-obtained Fe2N/N-PPC showed superior onset and half-wave potentials (0.966 and 0.891 V, respectively) in alkaline media (0.1 M KOH) compared to commercial Pt/C through a direct four-electron reaction pathway. Fe2N/N-PPC also showed better stability and methanol tolerance than commercial Pt/C. The outstanding ORR performance of Fe2N/N-PPC was attributed to its high specific surface area and the synergistic effects of Fe2N NPs. The utilization of agricultural wastes as a precursor makes Fe2N/N-PPC an ideal non-precious metal catalyst for ORR applications.


Introduction
According to the United Nations, almost 140 billion metric tons of biomass is generated globally from agricultural wastes. Therefore, the valorization of biomass can potentially reduce the use of fossil fuels while mitigating greenhouse gas emissions [1]. Biomass wastes, such as bananas [2,3], tea leaves [4], poplar catkins [5], gingko leaves [6], coconut shells [1], eggshell membranes [7], and sodium alginate [8], have been increasingly used for fabricating carbon materials. Biomass raw materials can be used to synthesize carbon materials with a high surface area and porous structure. This approach possesses several advantages (e.g., low cost, environmental friendliness, and convenience) over other routes employing different raw materials. Thus, biomass represents a nearly limitless resource for preparing a variety of new materials with novel structures and

Results and Discussion
With the aim of identifying the nanocrystalline phases present in the PPC, N-PPC, Fe-PPC, and Fe-N-PPC materials, X-ray diffraction (XRD) measurements were carried out ( Figure 1). PPC and N-PPC only showed the two typical (002) and (100) peaks of graphitic carbon at 26.2 and 43.3 • . Fe-PPC showed a minor amount of Fe (PDF#87-0721). The XRD patterns of Fe-PPC and Fe-N-PPC showed a broad peak at 23 • a small shoulder peak at 43.3 • assigned to the (002) (100) planes of graphite, respectively. A Fe 2 N crystalline phase (PDF#73-2102) was also detected in Fe-N-PPC. Moreover, the Fe-based NPs were accompanied by strong C signals, thereby demonstrating that these NPs were successfully loaded on the PPC material. The (002) diffraction peak shifted to lower angles (23 • ) for Fe-PPC and Fe-N-PPC as compared to PPC and N-PPC (26.2 • ), revealing an larger interplanar distance [31].
The I D /I G intensity ratio values obtained from the Raman spectra were 0.85, 0.89, 0.84, and 0.88 for the PPC, N-PPC, Fe-PPC, and Fe-N-PPC samples, respectively (Figure 2a). The D and G Raman bands are indicative of sp 3 defect sites of the graphite layers and the E 2g vibrational mode of sp 2 -bonded pairs, respectively. Thus, the I D /I G intensity ratio values obtained revealed that the introduction of Fe and N generated defects and disorderliness in the porous carbon material. These defects and disorder favored the formation of active sites on the surface of the catalysts. The intensity of the G-band increased upon addition of Fe 3+ . The I D /I G intensity ratio values remained nearly unchanged for all the samples, which resulted in excellent ORR activity. Nanomaterials 2017, 7, 404 3 of 10 of the G-band increased upon addition of Fe 3+ . The ID/IG intensity ratio values remained nearly unchanged for all the samples, which resulted in excellent ORR activity.  XPS measurements were performed to analyze in detail elemental composition and chemical valence of PPC, N-PPC, N-PPC, and Fe-N-PPC. As shown in Figure 2b and Table 1, the XPS spectrum of Fe-N-PPC showed the presence of C (95.86%), N (1.60%), O (2.12%), and Fe (0.43%). The N 1s   XPS measurements were performed to analyze in detail elemental composition and chemical valence of PPC, N-PPC, N-PPC, and Fe-N-PPC. As shown in Figure 2b and Table 1, the XPS spectrum of Fe-N-PPC showed the presence of C (95.86%), N (1.60%), O (2.12%), and Fe (0.43%). The N 1s XPS measurements were performed to analyze in detail elemental composition and chemical valence of PPC, N-PPC, N-PPC, and Fe-N-PPC. As shown in Figure 2b and Table 1, the XPS spectrum of Fe-N-PPC showed the presence of C (95.86%), N (1.60%), O (2.12%), and Fe (0.43%). The N 1s spectra was deconvoluted into five bands centered at 398.5, 399.2, 399.7, 400.9, and 405.8 eV (Figure 2c) assigned to pyridinic-N, N-Fe compounds, pyrrolic-N, graphitic-N, and oxidized-N, respectively [32]. PPC, N-PPC, and Fe-N-PPC showed pyridinic-N contents of 25.9 at. %, 31.9 at. %, and 32.2 at. %, respectively. Other N species contents are also listed in Table 2. Remarkably, the ORR active sites in the N-doped carbon nanomaterials were identified to involve carbon atoms with Lewis basicity and located in the vicinity of pyridinic-N [19]. Thus, the high content of pyridinic-N of the N-PPC and Fe-N-PPC sample is important in enhancing the ORR performance of these materials. The N-Fe compounds (9.0 at. %) showed the presence of iron nitride species. The Fe 2p spectra of Fe-N-PPC ( Figure 2d) contained weak Fe 2p 3/2 and Fe 2p 1/2 doublet signals at 710.8 eV and at 724.9 eV, respectively, suggesting the presence of slightly metallic species [29,31,[33][34][35][36]. The scanning electron microscopy (SEM) images of PPC are shown in Figure 3a,b. As indicated above, Fe-N-PPC was prepared from pomelo peel via a two-step carbonization process followed by calcination at 900 • C under an NH 3 atmosphere. This calcination process resulted in Fe 3+ /PPC being converted into a Fe-N-PPC. As shown in Figure 4a,d, transmission electron microscopy (TEM) and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) observations revealed the presence of Fe 2 N NPs distributed inside or anchored on the N-doped porous carbon with little agglomeration. The size of the Fe 2 N NPs was in the 20-30 nm range. Figure 3c shows the N 2 physisorption isotherms of the PPC and Fe-N-PPC catalysts. Fe-N-PPC exhibited a typical type IV curve characteristic of mesoporous materials. The pore size distribution curves (Figure 3d) of PPC and Fe-N-PPC revealed average pore sizes 2.60 and 3.48 nm, respectively. As shown in Table 3, Fe-N-PPC exhibited higher Brunauer-Emmett-Teller (BET) surface area than PPC (1103.90 versus 353.05 m 2 ·g −1 ). We concluded that increasing the temperature of the calcination treatment to 900 • C was effective in increasing the surface area of the materials by generating numerous defects. Larger BET surface area and accompanying defects are important in improving the ORR catalytic activity [5,34]. Thus, Fe-N-PPC showed superior ORR performance.
As revealed by energy-dispersive energy dispersive X-ray spectroscopy (EDS), the content of elemental Fe in the Fe-N-PPC varied significantly after the double carbonization treatment (Figure 4b,e) [29]. EDS mapping analysis of Fe-N-PPC indicate the presence of C, O, N, and Fe atoms. As shown by high resolution TEM (HRTEM) (Figure 4b), graphitized carbon protected these NPs from agglomerating and the thickness of the carbon layer is about 3 nm [29,37], thereby improving the ORR activity. The HRTEM image of Fe 2 N NPs revealed a fringe spacing of 0.21 nm, which corresponded to the (−1−11) plane of the Fe 2 N crystal phase (Figure 4c).        As shown in Figure 5a, Fe-N-PPC exhibited a noticeable reduction peak in the O 2 -saturated 0.1 M KOH solution, and this peak was absent in the N 2 -saturated 0.1 M KOH solution. These results are indicative of the high ORR electrocatalytic activity of Fe-N-PPC. RDE tests were performed to obtain the linear scan voltammetry (LSV) curves and further study the ORR performance of the samples. Fe-N-PPC displayed an onset potential of 0.966 V (Figure 5b), which was higher than those of PPC (0.89 V), N-PPC (0.899 V), Fe-PPC (0.940 V), and 20% Pt/C (0.952 V) catalysts. The prominent ORR activity of Fe-N-PPC was confirmed by its positive half-wave potential (0.891 V) and large current density (5.077 mA·cm −2 ). Pt/C showed a positive half-wave potential (0.865 V) and a large current density (4.92 mA·cm −2 ). Fe-N-PPC showed superior ORR activities as compared to other Fe 2 N-based catalysts reported in the literature. The large limited current density of Fe-N-PPC may result from its porous structure with anchored Fe 2 N particles and the broad diffusion pathways for solvated O 2 [38]. The superior onset potential of Fe-PPC versus porous carbon and N-PPC was originated from the presence of Fe and Fe-based systems. The more positive potential of Fe-N-PPC was produced by the presence of Fe and N co-doping systems, which increased the number of ORR active sites of carbon-based electrocatalysts, as confirmed by the XRD and Raman data. Besides, assessment of preparation process of Fe-N-PPC is compared to recently reported Fe 2 N/C catalysts and pomelo peel-derived carbon materials for ORR. As shown in Figure 5a, Fe-N-PPC exhibited a noticeable reduction peak in the O2-saturated 0.1 M KOH solution, and this peak was absent in the N2-saturated 0.1 M KOH solution. These results are indicative of the high ORR electrocatalytic activity of Fe-N-PPC. RDE tests were performed to obtain the linear scan voltammetry (LSV) curves and further study the ORR performance of the samples. Fe-N-PPC displayed an onset potential of 0.966 V (Figure 5b), which was higher than those of PPC (0.89 V), N-PPC (0.899 V), Fe-PPC (0.940 V), and 20% Pt/C (0.952 V) catalysts. The prominent ORR activity of Fe-N-PPC was confirmed by its positive half-wave potential (0.891 V) and large current density (5.077 mA·cm −2 ). Pt/C showed a positive half-wave potential (0.865 V) and a large current density (4.92 mA·cm −2 ). Fe-N-PPC showed superior ORR activities as compared to other Fe2N-based catalysts reported in the literature. The large limited current density of Fe-N-PPC may result from its porous structure with anchored Fe2N particles and the broad diffusion pathways for solvated O2 [38]. The superior onset potential of Fe-PPC versus porous carbon and N-PPC was originated from the presence of Fe and Fe-based systems. The more positive potential of Fe-N-PPC was produced by the presence of Fe and N co-doping systems, which increased the number of ORR active sites of carbonbased electrocatalysts, as confirmed by the XRD and Raman data. Besides, assessment of preparation process of Fe-N-PPC is compared to recently reported Fe2N/C catalysts and pomelo peel-derived carbon materials for ORR. Table 4 reveals that Fe-N-PPC showed excellent electrochemical property and was prepared only with pomelo peels, ammonia and FeCl3·6H2O.   The K-L plots (Figure 5d) of Fe-N-PPC, obtained from the LSV experiments by varying the rotation rates and a scan rate of 10 mV/s (Figure 5c), exhibited a good linearity and allowed us to determine the n of this catalyst. Fe-N-PPC showed n values of 4.06, 4.06, and 4.07, suggesting that a four-electron (4e − ) hybrid pathway prevailed from −0.6 to −0.8 V during the ORR process.
Durability tests were performed for the Fe-N-PPC and commercial Pt/C catalysts at −0.25 V vs. Ag/AgCl in an O 2 -saturated 0.1 M KOH aqueous solution at a rotation rate of 400 rpm (Figure 5e). After 20,000 s of reaction, the cathodic current decreased to 79.6% and 54.6% of the initial baseline value for Fe-N-PPC and Pt/C, respectively. Fe-N-PPC protected by a carbon shell on Fe 2 N nanoparticles, which can show the better stability [29]. The tolerance to methanol crossover effects was also tested for Fe-N-PPC and Pt/C (Figure 5f). After the addition of a 3 M methanol solution, commercial Pt/C instantaneously decreased its current as a result of the oxidation of methanol on the electrode surface. In contrast, Fe-N-PPC retained 97.2% of its initial baseline activity after methanol addition, revealing a good tolerance to methanol crossover.

Synthesis of Samples
Raw pomelo peel was air-dried and ground into a fine powder (80-mesh sieve). The powder was carbonized at 800 • C for 5 h under an Ar flow. The resulting porous carbon material was washed with deionized water and an HCl solution, and finally oven dried under air at 80 • C overnight to produce PPC. 1 g of PPC was subsequently dispersed under stirring in a saturated FeCl 3 solution overnight to allow adsorption saturation. A small amount of FeCl 3 ·6H 2 O was encapsulated in the pores of the carbon material. The FeCl 3 -saturated PPC material was collected by filtration and oven dried at 80 • C for 12 h and subsequently calcined by heating it from room temperature to 900 • C (5 • C·min −1 ) for 2 h under an NH 3 atmosphere to finally produce Fe-N-PPC (i.e., Fe 2 N/N-PPC). For comparison, the porous carbon material was carbonized at similar conditions to obtain N-PPC. Fe-PPC was obtained by heating FeCl 3 -saturated porous carbon in an Ar atmosphere.

Characterization
A Hitachi SU8010 microscope (Hitachi, Tokyo, Japan) was used for the SEM observations. TEM, EDS and HRTEM images of the samples were obtained on JEM-2100 (JEOL, Tokyo, Japan) and Tecnai G2 F20 (FEI, Hillsboro, OR, USA) microscopes, respectively. XRD experiments were conducted on a Bruker D8 Advance X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å, Bruker, Karlsruhe, Germany). An AMICUS/ESCA 3400 device (Kratos, Manchester, UK) was used to collect the XPS data using Mg Kα (12 kV, 20 mA) radiation. The C1s line at 284.8 eV served as the reference. A Micromeritics ASAP 2020 N 2 adsorption/desorption apparatus (Micromeritics, Atlanta, GA, USA) was used for performing the Barrett-Joyner-Halenda (BJH) pore structure analysis, and for measuring the BET specific surface area (SSA) at 77.35 K. Elemental mapping was conducted on a JEM-ARM 200F microscope operating at 200 kV (JEOL, Tokyo, Japan). The Raman spectra were obtained on a Renishaw instrument (Renishaw, London, UK) with an Ar ion laser (514.5 nm excitation wavelength).

Electrochemical Measurements
The ORR catalytic activity of the samples was evaluated on a CHI760D instrument (CH Instruments, Shanghai, China). The powder (5 mg) sample was mixed with a Nafion solution (5%) and ultrasonically mixed for 30 min before being loaded on a glassy carbon (GC) electrode (0.153 mg cm −2 ). The GC electrode containing a thin powder layer was fixed on a rotating disk electrode (RDE) that served as the working electrode (WE). Pt foil and Ag/AgCl served as counter and reference electrodes (CE and RE, respectively). A 0.1 M KOH aqueous solution served as the electrolyte, and the rotation speed varied from 400 to 2025 rpm. The potentials (versus Ag/AgCl) were converted to a RHE scale by the equation: Koutecky-Levich (K-L) plots were subsequently obtained from the linear sweep voltammetry (LSV) data with the aim of measuring the electron transfer number (n) in the ORR process. The electron transfer number per O 2 molecule in the ORR process can be calculated according to the following equations [39]: 1/J = 1/J L + 1/J K = 1/Bw 1/2 + 1/J K where F is the Faraday constant (96,485 C), C o is the oxygen bulk concentration (1.2 × 10 −3 mol·cm −3 ), D o is the O 2 diffusion coefficient (ca. 1.9 × 10 −5 cm 2 ·s −1 ), ν is the electrolyte kinematic viscosity (1.1 × 10 −2 cm 2 ·s −1 ), J is the measured current density, ω = 2πN with N being the linear rotation speed, J L is the diffusion-limiting current density, and J K is the kinetic current density.

Conclusions
In summary, N-doped highly graphitized porous carbon-supported Fe 2 N NPs (Fe-N-PPC) were successfully synthesized from pomelo peel, ammonia and FeCl 3 ·6H 2 O via a two-step pyrolysis processes. The resulting Fe-N-PPC catalyst exhibited an outstanding ORR activity in alkaline media with onset and half-wave potentials of 0.966 and 0.891 V, respectively. These values were more positive than those of a commercial Pt/C catalyst. Fe-N-PPC also showed excellent electrochemical stability and methanol tolerance compared to Pt/C. These good characteristics resulted from the high specific surface area and the synergistic effect of Fe 2 N. In summary, the Fe-N-PPC catalyst represents a promising alternative to expensive commercial Pt/C catalysts for practical applications of alkaline fuel cells.