FeNxC Based Catalysts Prepared by the Calcination of Iron-Ethylenediamine@Polyaniline as the Cathode-Catalyst of Proton Exchange Membrane Fuel Cell

Calcinated tris(ethylenediamine)iron(III) chloride was used as a non-precious metal catalyst (NPMCs) for a proton exchanged membrane fuel cell (PEMFC) under the protection of polyaniline (PANI), which behaves as both nitrogen source and carbon supporter. The optimal ratio of FeCl3/EDA was found to be close to 1/3 under the consideration of the electrocatalytic performance, such as better oxygen reduction reaction (ORR) and higher power density. Two-stage calcination, one at 900 °C in N2 and the other at 800 °C in mixed gases of N2 and NH3, result in an FeNxC catalyst (FeNC-900-800-A) with pretty high specific surface area of 1098 m2·g−1 covered with both micro- and mesopores. The ORR active sites focused mainly on Fe–Nx bonding made of various pyridinic, pyrrolic, and graphitic N-s after calcination. The max. power density reaches 140 mW·cm−2 for FeNC-900-800-A, which is superior to other FeNxC catalysts, experiencing only one-stage calcination in N2. The FeNxC demonstrates only 10 mV half-wave-voltage (HWV) loss at 1600 rpm after 1000 redox cycles, as compared to be 27 mV for commercial Pt/C catalyst in the durability test.


Introduction
The expensive precious metals used as oxygen reduction reaction (ORR) catalysts restrain the mass-production of proton exchange membrane fuel cells (PEMFCs). Non-precious metal catalysts are suitable candidates to substitute Pt catalysts. The aim of non-precious metal catalysts (NPMCs) is to substitute high priced Pt-based catalysts. NPMCs require not only high ORR catalytic activity but also inexpensive components. The developing carbon based inexpensive non-precious metal ORR catalyst has turned out to be one of the most interesting topics in the area of fuel cells.
The chemical reaction of a fuel cell is based on the redox reaction between hydrogen and oxygen in the anode and cathode, respectively. However, ORR at the cathode is much slower than that of the hydrogen reduction reaction (HOR) at the anode, which needs more precious Pt-catalyst to pick up the ORR rate to fulfill the redox cycle to produce power. In other words, the ORR has turned out to be the neck-intermediate reaction during the redox reaction. And the main factor to cause the

Preparation of FeNxC Catalyst
1 g (0.0037 mol) of Iron(III) chloride hexahydrate was mixed with 1 M HCl (aq.) solution and stirred to become uniform. 1 mL (0.66 g:0.011 mol) of EDA was introduced into the mixture (the molar ratio Fe/EDA = 1/3) and stirred again, followed by the addition of 2 g aniline monomers. 4.9 g of ammonium persulfate (APS) dissolved in 1 M HCl (aq.) was slowly added into the previous mixture to start the polymerization until the mixed solution became dark green (about 1.5 h). After the polymerization, the solvents of the mixture were removed by vacuum oven at 80 • C and the precursor of the FeNxC catalyst was obtained. The precursor was heated to 700 • C (800, 900, 950 • C) at 10 • C min −1 and stayed for 1 h in the argon atmosphere, then cooled to room temperature (RT). These samples were named FeNxC-700 (-800, -900, -950). The impurities and magnetic materials of the obtained FeNxCs were removed by washing in 1 M H 2 SO 4 (aq.) at 80 • C for 3 h, followed by drying in a vacuum oven at 60 • C. These sample were named FeNxC-700-A (-800-A, -900-A, -950-A). FeNxC-900-A was further calcinated to 800 • C in N 2 and NH 3 atmospheres again at 10 • C min −1 (named FeNxC-900-800) and washed again in 1 M H 2 SO 4 (aq.) at 80 • C for 3 h, followed by drying in a vacuum oven at 60 • C. The sample was named FeNxC-900-800-A.

FTIR
The main functional groups of neat PANI and various FeNxCs prepared with different ratios of FeCl 3 /EDA were assigned from FTIR spectra which were recorded on an IFS3000 v/s FTIR spectrometer (Bruker, Ettlingen, Germany) at room temperature with a resolution of 4 cm −1 and 16 scanning steps.

X-ray Photoelectron Spectroscopy (XPS)
The different binding energy spectra of N1s of various FeNxCs were used to characterize the percentage of pyridinic, pyrrolic, graphitic Ns, etc. after calcination by an XPS instrument of Fison (VG)-Escalab 210 (Fison, Glasgow, UK) using Al Ka X-ray source at 1486.6 eV. The pressure in the chamber was maintained under 10 −6 Pa or lower during the measurement. A tablet sample was prepared by a stapler. The binding energies of the N1s around 400 eV were recorded.

WXRD (Wide angle X-ray Diffraction: Powder X-ray Diffraction)
A copper target (Cu-Kα) Rigaku x-ray source (Rigaku, Tokyo, Japan) with a wavelength of 1.5402 Å was used for x-ray diffraction. The scanning angle (2θ) started from 10 • to 90 • with a voltage of 40 kV and a current of 30 mA, operated at 1 • min −1 .

Energy Dispersive X-ray Spectra (EDx)
The EDx spectra of various FeNxCs were obtained from an XL-40EFG by Philips (Amsterdam, Netherlands). The samples were coated with gold before measurement.

Scanning Electronic Microscopy (SEM)
The sizes and morphologies of FeNxCs were characterized by SEM (field emission gun scanning electron microscope, AURIGAFE, Zeiss, Germany).

Surface Area and Pore Size Measurement (BET Method)
Nitrogen adsorption-desorption isotherms (type IV) are obtained from an Autosorb IQ gas sorption analyzer (Manufacturer, city, country) (Quantachrome) at 25 • C. Samples are degassed at a temperature higher than 100 • C overnight. The surface area was calculated according to the BET equation when a linear BET plot with a positive C value was in the relative pressure range. Pore size distribution was determined by the Quenched Solid Density Functional Theory (QSDFT) methods based on a model of slit/cylinder pores. The total pore volumes were determined at P/P 0 = 0.95. The performance of the electrocatalyst support was tested with a three-electrode system. The round working electrode with an area of 1.5 cm 2 was prepared as follows: Ag/AgCl and carbon graphite were used as the reference and relative electrode, respectively. The electrochemical test was carried out in a potentiostat/galvanostat (Autolab-PGSTAT 30 Eco Chemie, Artisan Technology Group, IL, USA) in 0.5 M H 2 SO 4 solution and C-V curves were obtained with scanning potential from −0.2 to 1.0 V at a sweeping rate of 50 mV·s −1 . The catalyst ink was prepared by mixing 3mg support powder in isopropanol and stirred until it became uniform. Subsequently, 5% Nafion solution was added into the mixture as binder and the mixture was ultra-sonicated for 1 h and the obtained ink was uniformly spray-coated on the carbon paper for C-V test.
The current-potential polarization curves obtained from LSV of the various FeNxCs were measured using a rotating-disk electrode (RDE) operating at 900, 1200, 1600, 2500, and 3600 rpm in O 2 -saturated 0.5 M H 2 SO 4 , respectively. The reduced current densities of various FeNxCs, which were recorded at 1600 rpm within the measured voltage range (−1.2~0.5 V) were chosen to compare with each other.

MEA Preparation
A Nafion ® 212 sheet purchased from Ion Power Inc., New Castle, DE, USA was used as the proton exchange membranes. In order to remove the surface organic impurities and to convert the membranes into protonated (H + ) form, the Nafion-212 (4 × 4 cm 2 ) membrane was treated at 70 • C in 5 wt.% H 2 O 2 aqueous solution for 1 h, followed by submerging in 1 M H 2 SO 4 solution for 1 h and subsequently the treated membranes were dipped in distilled water for 15 min and stored in de-ionized water. The catalyst inks were prepared by mixing 20 mg of FeNxC powders in isopropanol and mechanically stirred until it became uniform before 5% Nafion solution was added. Eventually, the catalyst mixture was ultra-sonicated for 1h, followed by coating on both sides of the treated Nafion sheet dropwise as anode and cathode electrodes (2 × 2 cm 2 ), respectively, and hot-pressed at 140 • C with a pressure force of 70 kg cm −2 for 5min to obtain the MEA.

Single-Cell Performance Testing
The MEA was installed in a fuel cell test station for testing using the single-cell test equipment (model FCED-P50; Asia Pacific Fuel Cell Technologies, Ltd., Miaoli, Taiwan). The active cell area was 2 × 2 cm 2 . The temperatures of anode, cell and cathode and humidifying gas (90% RH) were all maintained at around 70 • C. The flow rates of anode input H 2 and the cathode input O 2 fuels were set at 200 and 100 mL·min −1 , respectively, based on stoichiometry. To test the electrochemical performance of FeNxC catalyst in the individual MEAs, both C-V and output powers were measured.

FTIR Spectra
Regularly, Fe 3+ are coordinated with water in the form of Fe 3+ (H 2 O) 6 (FeCl 3 .6H 2 O) and the coming of EDAs was able to replace the water molecules. Significant green precipitation was seen when the molar ratio of FeCl 3 /EDA is close to 1/3, whose structure is thought to be Fe 3+ (NH 2 C 2 H 4 NH 2 ) 3 . When more EDA was added, no further precipitation can be seen. It is believed that an insoluble salt of tri(ethylenediamine)iron(III) chloride was created as the three EDA ligands occupied the six coordination positions of ferric ion, resulting in the formation of octahedral crystal demonstrated in Scheme 1. The Fe-EDA complex was prepared with different FeCl3/EDA ratios before covering by PANI, which were characterized by IR-spectra and illustrated in Figure 1. Major functional groups (doped and undoped states of quinoid, benzenoid, -C-N-groups) of PANI are assigned directly as seen in Figure 1, indicating PANI (ES) was successfully obtained. However, when the FeCl3/EDA ratio was over 1/3 (1/4), the spectrum demonstrated more undoped -B-NH-B-than doped -Q-NH + -B-groups (Q: Quinoid; B: Benzenoid) due to the neutralization effect of the surplus EDA which behaves as a base chemical and can induce the dedoping of the -Q-NH + -B-group. The sulfonic groups assigned at 1020 cm −1 were derived from the persulfate initiator, which is converted to sulfuric group after the initial reaction.
A blank test without the protection of PANI produced a high magnetic FeNxC catalyst after calcinating at 900 °C. Therefore, even under the protection of PANI, the calcinated FeCl3/EDA@PANI with FeCl3/EDA ratio equal to or higher than 1/4 also still demonstrated a magnetic attraction property due to the strong reducing force of the excess EDA, which made the dispersion in the solvent very difficult during preparation of a slurry ink for the MEA. Although the magnetic compounds of the calcinated FeNxC products could be easily removed by acid-leaching when the FeCl3/EDA ratio was below 1/4, it could only be partly removed at a higher ratio than 1/3 (1/4). In other words, a higher FeCl3/EDA ratio than 1/3 would lead to a magnetic FeNxC catalyst during calcination and could not be totally removed by acid-leaching. The Fe-EDA complex was prepared with different FeCl 3 /EDA ratios before covering by PANI, which were characterized by IR-spectra and illustrated in Figure 1. Major functional groups (doped and undoped states of quinoid, benzenoid, -C-N-groups) of PANI are assigned directly as seen in Figure 1, indicating PANI (ES) was successfully obtained. However, when the FeCl 3 /EDA ratio was over 1/3 (1/4), the spectrum demonstrated more undoped -B-NH-B-than doped -Q-NH + -B-groups (Q: Quinoid; B: Benzenoid) due to the neutralization effect of the surplus EDA which behaves as a base chemical and can induce the dedoping of the -Q-NH + -B-group. The sulfonic groups assigned at 1020 cm −1 were derived from the persulfate initiator, which is converted to sulfuric group after the initial reaction.
A blank test without the protection of PANI produced a high magnetic FeNxC catalyst after calcinating at 900 • C. Therefore, even under the protection of PANI, the calcinated FeCl 3 /EDA@PANI with FeCl 3 /EDA ratio equal to or higher than 1/4 also still demonstrated a magnetic attraction property due to the strong reducing force of the excess EDA, which made the dispersion in the solvent very difficult during preparation of a slurry ink for the MEA. Although the magnetic compounds of the calcinated FeNxC products could be easily removed by acid-leaching when the FeCl 3 /EDA ratio was below 1/4, it could only be partly removed at a higher ratio than 1/3 (1/4). In other words, a higher FeCl 3 /EDA ratio than 1/3 would lead to a magnetic FeNxC catalyst during calcination and could not be totally removed by acid-leaching.

XPS
The active sites of FeNxC can absorb O2 gas and form a peroxide structure which would dissociate in the presence of protons during reduction (Scheme 2). The formed peroxide would dissociate in two approaches [31,32], one becoming diol and following the four-e route, directly converting into water. The other went on consecutive reduction with two electrons involved in the following stage with H2O2 as the intermediate product (Scheme 2). The catalytic mechanism followed the traditional six-coordination catalytically reaction. Theoretically, the O2 gas with two lone pairs could be absorbed by the active sites of FeNxC through the coordination, or trapped in the porous holes with various N-related bonds whose lone paired electrons can improve the O2 absorbing capability of FeNxC and behave as active sites like transitional metals (Fe). The active -N can be pyrrolic -N, graphitic -N (Scheme 1), but not pyridinic -N [33][34][35][36][37][38], whose lone-paired electrons were easily covered by the acidic proton through acid-base neutralization (-N: + H + => -NH + ). The first calcination at high temperatures in the N2 atmosphere can create various -N containing covalent bonds, behaving as active sites and the second calcination in the mixed gases of N2 and NH3 would engrave on the FeNxC surfaces with lots of micro-or

XPS
The active sites of FeNxC can absorb O 2 gas and form a peroxide structure which would dissociate in the presence of protons during reduction (Scheme 2). The formed peroxide would dissociate in two approaches [31,32], one becoming diol and following the four-e route, directly converting into water. The other went on consecutive reduction with two electrons involved in the following stage with H 2 O 2 as the intermediate product (Scheme 2). The catalytic mechanism followed the traditional six-coordination catalytically reaction.

XPS
The active sites of FeNxC can absorb O2 gas and form a peroxide structure which would dissociate in the presence of protons during reduction (Scheme 2). The formed peroxide would dissociate in two approaches [31,32], one becoming diol and following the four-e route, directly converting into water. The other went on consecutive reduction with two electrons involved in the following stage with H2O2 as the intermediate product (Scheme 2). The catalytic mechanism followed the traditional six-coordination catalytically reaction. Theoretically, the O2 gas with two lone pairs could be absorbed by the active sites of FeNxC through the coordination, or trapped in the porous holes with various N-related bonds whose lone paired electrons can improve the O2 absorbing capability of FeNxC and behave as active sites like transitional metals (Fe). The active -N can be pyrrolic -N, graphitic -N (Scheme 1), but not pyridinic -N [33][34][35][36][37][38], whose lone-paired electrons were easily covered by the acidic proton through acid-base neutralization (-N: + H + => -NH + ). The first calcination at high temperatures in the N2 atmosphere can create various -N containing covalent bonds, behaving as active sites and the second calcination Theoretically, the O 2 gas with two lone pairs could be absorbed by the active sites of FeNxC through the coordination, or trapped in the porous holes with various N-related bonds whose lone paired electrons can improve the O 2 absorbing capability of FeNxC and behave as active sites like transitional metals (Fe). The active -N can be pyrrolic -N, graphitic -N (Scheme 1), but not pyridinic -N [33][34][35][36][37][38], whose lone-paired electrons were easily covered by the acidic proton through acid-base neutralization (-N: + H + => -NH + ). The first calcination at high temperatures in the N 2 atmosphere can create various -N containing covalent bonds, behaving as active sites and the second calcination in the mixed gases of N 2 and NH 3 would engrave on the FeNxC surfaces with lots of micro-or mesopores, resulting in a drastically increasing surface area and exposing more active -N and -Fe sites to the O 2 gas. Therefore, FeNxC-900-800, which experiences two stages of calcination in different atmospheres, gives the highest LSV current and max. power density, which will be discussed.
For PANI itself, it would crosslink into ladder-like polymers with each other in the initial stage of thermal heating and higher temperature pyrolysis allows the carbonation between the ladder-like polymers, which would form networks of graphitic -N, pyridinic -N, and pyrrolic -N (Scheme 3). Both pyridinic and pyrrolic -Ns are on the edges of calcinated PANI while graphitic -N are located inside. The conjugated, carbonized structure of pyrolyzed EB can act as a conducting medium as well, allowing the flowing in of electrons from the anode. It saves the trouble of adding XC-72 in the preparation of the cathode ink. Besides, the tri(ethylenediamine)iron(III) complexes can successfully convert into a covalent network under the protection of enclosed PANI during calcination and Fe would capture more -N, becoming the active Fe-Nx centers (Scheme 4). mesopores, resulting in a drastically increasing surface area and exposing more active -N and -Fe sites to the O2 gas. Therefore, FeNxC-900-800, which experiences two stages of calcination in different atmospheres, gives the highest LSV current and max. power density, which will be discussed. For PANI itself, it would crosslink into ladder-like polymers with each other in the initial stage of thermal heating and higher temperature pyrolysis allows the carbonation between the ladder-like polymers, which would form networks of graphitic -N, pyridinic -N, and pyrrolic -N (Scheme 3). Both pyridinic and pyrrolic -Ns are on the edges of calcinated PANI while graphitic -N are located inside. The conjugated, carbonized structure of pyrolyzed EB can act as a conducting medium as well, allowing the flowing in of electrons from the anode. It saves the trouble of adding XC-72 in the preparation of the cathode ink. Besides, the tri(ethylenediamine)iron(III) complexes can successfully convert into a covalent network under the protection of enclosed PANI during calcination and Fe would capture more -N, becoming the active Fe-Nx centers (Scheme 4). The N1s XPS spectra of FeNxCs treated at high temperatures after acid-leaching are stacked together and demonstrated in Figure 2(a). The compositions of each type of -Ns are illustrated in Figure 2(b). The neat and FeNxC-900-800-A demonstrate less pyridinic -Ns. More pyridinic -Ns are found for FeNxCs experiencing only one stage of calcination in N2 atmosphere according to Figure 2. Briefly, some of the pyridinic -Ns were converted to pyrrolic -Ns at the second calcination in the presence of NH3 and N2. Two features can be seen for the N1s XPS spectrum of FeNxC-900-800-A in Figure 2(b). One is the competing high concentration of pyrrolic and graphitic -Ns and the other is that it owns the significant Fe-N covalent-bonding. High power density of the single cell prepared with a catalyst of FeNxC-900-800-A containing more graphitic, pyrrolic -Ns, and active Fe-N bonds will be discussed.  Figure 2b. The neat and FeNxC-900-800-A demonstrate less pyridinic -Ns. More pyridinic -Ns are found for FeNxCs experiencing only one stage of calcination in N 2 atmosphere according to Figure 2. Briefly, some of the pyridinic -Ns were converted to pyrrolic -Ns at the second calcination in the presence of NH 3 and N 2 . Two features can be seen for the N1s XPS spectrum of FeNxC-900-800-A in Figure 2b. One is the competing high concentration of pyrrolic and graphitic -Ns and the other is that it owns the significant Fe-N covalent-bonding. High power density of the single cell prepared with a catalyst of FeNxC-900-800-A containing more graphitic, pyrrolic -Ns, and active Fe-N bonds will be discussed.

XRD
The optimal ratio of tri(ethylenediamine)iron(III)@PANI is also checked with XRD patterns which demonstrate a significant crystalline pattern of FeS after calcination at 900 °C for all ratios in Figure 3(a). However, the FeS crystalline can be removed by leaching with acid when the ratio is between 1/1 and 1/3 as shown in Figure 3(b). For a sample prepared with

XRD
The optimal ratio of tri(ethylenediamine)iron(III)@PANI is also checked with XRD patterns which demonstrate a significant crystalline pattern of FeS after calcination at 900 • C for all ratios in Figure 3a. However, the FeS crystalline can be removed by leaching with acid when the ratio is between 1/1 and 1/3 as shown in Figure 3b. For a sample prepared with tri(ethylenediamine)iron(III)@PANI equal to 1/4, the X-ray diffraction pattern reveals the dominating FeS crystalline structure. The initiation of aniline monomers can convert the peroxysulfate into sulfate ions and the formed FeSO 4 can easily be converted into FeS with high temperature calcination.
XRD patterns are also used to characterize the optimal calcination (preparing) temperature of the FeNxCs. The XRD spectra and standard patterns of each specific crystalline compound are illustrated in Figure 3c,d to demonstrate the crystalline pattern of FeNxCs after treating at various temperatures and acid-leaching. All FeNxCs which experience only one-stage calcination demonstrated a high concentration of FeS crystals (Figure 3c). Only after second stage calcination at 800 • C (FeNxC-900-800) in mixed gases (N 2 + NH 3 ) the FeS can be replaced with Fe 3 O 4 (Figure 3c). However, the high magnetic property of Fe 3 O 4 makes further characterization by SEM and TEM impossible. Besides, strong attractive magnetic forces make the dispersion of FeNxC-900-800 in the isopropanol solvent very difficult, leading to no MEA that can be successfully prepared. Therefore, the acid-leaching procedure is necessary and can remove most of the FeS or Fe 3 O 4 crystallines for all FeNxCs except that one prepared with FeCl 3 /EDA equal to 1/4 according to Figure 3d. After removing the FeS or Fe 3 O 4 compounds by acid-leaching, the spaces, once occupied by them, become many nanopores and mesopores on the surfaces of FeNxCs, which can expose more active sites of Fe-N or create more empty, tiny spaces to accommodate more oxygen gas for better electrocatalytic properties. tri(ethylenediamine)iron(III)@PANI equal to 1/4, the X-ray diffraction pattern reveals the dominating FeS crystalline structure. The initiation of aniline monomers can convert the peroxysulfate into sulfate ions and the formed FeSO4 can easily be converted into FeS with high temperature calcination. XRD patterns are also used to characterize the optimal calcination (preparing) temperature of the FeNxCs. The XRD spectra and standard patterns of each specific crystalline compound are illustrated in Figure 3(c) and (d) to demonstrate the crystalline pattern of FeNxCs after treating at various temperatures and acid-leaching. All FeNxCs which experience only one-stage calcination demonstrated a high concentration of FeS crystals (Figure 3(c)). Only after second stage calcination at 800 °C (FeNxC-900-800) in mixed gases (N2 + NH3) the FeS can be replaced with Fe3O4 (Figure 3(c)). However, the high magnetic property of Fe3O4 makes further characterization by SEM and TEM impossible. Besides, strong attractive magnetic forces make the dispersion of FeNxC-900-800 in the isopropanol solvent very difficult, leading to no MEA that can be successfully prepared. Therefore, the acid-leaching procedure is necessary and can remove most of the FeS or Fe3O4 crystallines for all FeNxCs except that one prepared with FeCl3/EDA equal to 1/4 according to Figure 3(d). After removing the FeS or Fe3O4 compounds by acid-leaching, the spaces, once occupied by them, become many nanopores and mesopores on the surfaces of FeNxCs, which can expose more active sites of Fe-N or create more empty, tiny spaces to accommodate more oxygen gas for better electrocatalytic properties.

EDx
The atomic ratios of C, N, S, and Fe ( obtained from EDX spectra for all FeNxCs treated at various temperatures after acid-leaching ) are demonstrated in Figure 4 and listed in Table 1. The carbon composition increased with temperature in the first stage of calcination for FeNxCs and slightly

EDx
The atomic ratios of C, N, S, and Fe ( obtained from EDX spectra for all FeNxCs treated at various temperatures after acid-leaching ) are demonstrated in Figure 4 and listed in Table 1. The carbon composition increased with temperature in the first stage of calcination for FeNxCs and slightly decreased after second calcination (FeNxC-900-800-A) in an ammonia atmosphere, indicating NH 3 did etch out some of the surface carbon and allowed for the formation of more Fe-N bonding [34].

SEM
The optimal ratio of FeCl3/EDA was also confirmed by taking the SEM micrographs at various ratios and illustrated in Figure 5a-e. Samples with ratios from 1/1 to 1/3 (Figure 5b-d) demonstrated micro-and mesopores on the surfaces after calcination at 900 °C. However, all the pores were covered with tiny Fe elements (Figure 5e) when the ratio was increased to 1/4, revealing the optimal ratio to be equal to the stoichiometric one (1/3).
The micrographs of all FeNxCs treated at different temperatures are illustrated in Figure 5f-i, which reveals the decreasing pore-size from large pores to mesopores with temperature for one-stage calcination. And the second calcination (two-stage) could have further shrunk the pore-size but still remained within the mesoporous range (Figure 5i). The surface area could have increased tremendously after second calcination in NH3 at 800 °C as well, which will be discussed in the following section.  Sample before acid-leaching demonstrates high S % in Figure 4a for FeNxC-700. And the low S % for all acid-treated samples reveal (Figure 4b-d) most of the crystallized FeS are removed. The absence of oxygen % for FeNxC-900-800-A indicates that magnetic Fe 3 O 4 found before acid-treatment from X-ray diffraction has been removed by acid treatment (Figure 4e). N composition decreases with temperature and slightly increases after NH 3 join in the second stage of calcination (Table 1). The Fe reaches the highest (3.84%) for FeNxC-900-800-A (Table 1), attributing to the formation of more Fe-N bonds in the presence of NH 3 , acting as both etching materials and N sources.

SEM
The optimal ratio of FeCl 3 /EDA was also confirmed by taking the SEM micrographs at various ratios and illustrated in Figure 5a-e. Samples with ratios from 1/1 to 1/3 (Figure 5b-d) demonstrated micro-and mesopores on the surfaces after calcination at 900 • C. However, all the pores were covered with tiny Fe elements (Figure 5e) when the ratio was increased to 1/4, revealing the optimal ratio to be equal to the stoichiometric one (1/3).

BET Surface Area
The type IV isotherm, which is the characteristic N2-absorption and desorption curves of mesopores, can be clearly seen in Figure 6. FeNxC-900-800-A (Figure 6a) demonstrates much higher specific volume (>250 cm 3 ·g −1 ) than either -900-A or -950-A (<200 cm 3 ·g −1 ) at all relative pressures. Furthermore, the surface area (specific volume) became less and less with increasing calcination temperature during one-stage calcination in the N2 atmosphere. The collapsing destruction at a high The micrographs of all FeNxCs treated at different temperatures are illustrated in Figure 5f-i, which reveals the decreasing pore-size from large pores to mesopores with temperature for one-stage calcination. And the second calcination (two-stage) could have further shrunk the pore-size but still remained within the mesoporous range (Figure 5i). The surface area could have increased tremendously after second calcination in NH 3 at 800 • C as well, which will be discussed in the following section.

BET Surface Area
The type IV isotherm, which is the characteristic N 2 -absorption and desorption curves of mesopores, can be clearly seen in Figure 6. FeNxC-900-800-A (Figure 6a) demonstrates much higher specific volume (>250 cm 3 ·g −1 ) than either -900-A or -950-A (<200 cm 3 ·g −1 ) at all relative pressures. Furthermore, the surface area (specific volume) became less and less with increasing calcination temperature during one-stage calcination in the N 2 atmosphere. The collapsing destruction at a high temperature can result in the loss of more Fe and N elements and surface area, as already seen in Table 1 and Figure 6a, respectively. The specific area was significantly increased from 395 to 1098 m 2 ·g −1 according to Table 2 and Figure 6a when FeNxC was calcinated again in an NH 3 atmosphere at 800 • C. Additionally, the pore size distribution illustrated in Figure 6b was obtained by Barrett-Joyner-Halenda (BJH) method, revealing the presence of both micro-and mesopores. The re-rising surface area found in Figure 6a and Table 2 for FeNxC-900-800-A could be attributed to the etching power of NH3 which caused only insignificant damages on the surfaces but created more micro-and mesopores in addition to compensating for loss of the N element. The graded distribution of pore sizes indicates FeNxCs are very useful to improve ORR because the presence of micropores can uncover the FeNxC active sites, and mesopores are able to play the roles of confining the O2 inside, effectively shortening the diffusion distances [39]. Table 2. the specific surface area and averaged pore size of FeNxC.

Oxygen Reduction Reaction (ORR) Performance
To evaluate the electrocatalytic activity of FeNxCs prepared with different FeCl3/EDA ratios and treated temperatures after acid-leaching, various LSV curves are measured and shown in Figure 7. After treated at 900 °C, neat PANI did not demonstrate any ORR current in Figure 7a though it is an N containing compound and calcination could also improve its conductivity. If ferric ions were introduced in the absence of EDA, the reduced current at zero potential increased to −0.95 mA·cm −2 in Figure 7a, indicating the electrocatalytic capability of iron transitional metal for ORR, depicted in Scheme 2, whose empty d-orbitals were able to accommodate (adsorb) oxygen for ORR. The presence of coordinated N is similar to the co-catalyst of Ziegler Natta, which behaves as the absorbing sites  Additionally, the pore size distribution illustrated in Figure 6b was obtained by Barrett-Joyner-Halenda (BJH) method, revealing the presence of both micro-and mesopores. The re-rising surface area found in Figure 6a and Table 2 for FeNxC-900-800-A could be attributed to the etching power of NH 3 which caused only insignificant damages on the surfaces but created more micro-and mesopores in addition to compensating for loss of the N element. The graded distribution of pore sizes indicates FeNxCs are very useful to improve ORR because the presence of micropores can uncover the FeNxC active sites, and mesopores are able to play the roles of confining the O 2 inside, effectively shortening the diffusion distances [39].

Oxygen Reduction Reaction (ORR) Performance
To evaluate the electrocatalytic activity of FeNxCs prepared with different FeCl 3 /EDA ratios and treated temperatures after acid-leaching, various LSV curves are measured and shown in Figure 7. After treated at 900 • C, neat PANI did not demonstrate any ORR current in Figure 7a though it is an N containing compound and calcination could also improve its conductivity. If ferric ions were introduced in the absence of EDA, the reduced current at zero potential increased to −0.95 mA·cm −2 in Figure 7a, indicating the electrocatalytic capability of iron transitional metal for ORR, depicted in Scheme 2, whose empty d-orbitals were able to accommodate (adsorb) oxygen for ORR. The presence of coordinated N is similar to the co-catalyst of Ziegler Natta, which behaves as the absorbing sites of unsaturated monomers owning double bonds like oxygen (Scheme 2). Likewise, with those Fe-Nx with certain empty coordinated sites, they are able to accommodate oxygen molecules with the right incoming angles that allow them to fit into the empty coordinated sites of iron centers and be captured by the FeNxC catalysts.
However, without the cooperation (complexation) of -Nx (in Fe-Nx form) from the EDA, the ORR can only occur in a lower scale. The joining of equal molar ratio of EDA (1/1) before calcination at 900 • C could slightly improve the reduced current to 1 mA·cm −2 in Figure 7a. A significant increase of reduced current to 3 mA·cm −2 when the molar ratio was elevated to 1/3 which is the stoichiometric ratio to form tri(ethylenediamine)iron(III) complex (Scheme 1). We believed the effective Fe-Nx bonding in the form of pyridinic, graphitic, or pyrrolic N were formed at this ratio by calcination at 900 • C (Scheme 1). The six coordinated Ns belonging to three EDAs chelated to Fe 3+ will be reduced to less than five (Fe-N 2 , Fe-N 3 , or Fe-N 4 ) after high temperature calcination and part of the Ns of EDA would join into the carbonization structure of PANI (Scheme 4).
Excess EDA did not further improve the ORR but almost deterred the ORR according to Figure 7a. The reduced current almost disappeared in its LSV curve due to the covered active sites (drastically decreased surface area, will be discussed in BET section). Therefore, the stoichiometric ratio 1/3 is the optimal ratio for obtaining the highest degree of ORR.
FeNxCs with FeCl 3 /EDA ratio of 1/3 were calcinated at various temperatures and the results were illustrated in Figure 7b. When the temperature was increased from 700 to 900 • C, it brought the improvement of reduced current according to Figure 7b. However, the current decreased when the temperature was over 950 • C due to the loss of Fe and N. The reduced current can be regained if it experiences second calcination in mixed gases of N 2 and NH 3 . As expected, the reduced current was increased to 3.4 mA·cm −2 after second calcination (Figure 7b), which can create more Fe-Nx covalent-bondings and surface area as already discussed.
LSV curves were obtained from RDE at different rotating speeds to calculate the number of e-transferred from K-L equations and compared with commercial catalyst Pt/C ORR in Figure 7c-e. The LSV curves and K-L lines (insets) of FeNxC-900-A, FeNxC-900-800-A and commercial Pt/C were obtained and illustrated in Figure 7c-e, respectively. The numbers of e-transferred for each catalyst at different potentials are revealed in Figure 7f which shows the e-transferred numbers are all close to 3.7 but the two-stage calcinated FeNxC (900-800) own higher e-transferred number of 3.75.

Single Cell Testing
The single cell made of PANI-catalyst demonstrated no significant power or current density although it already owned carbonized after calcination according to Figure 8a. When only FeCl3 was

Single Cell Testing
The single cell made of PANI-catalyst demonstrated no significant power or current density although it already owned carbonized after calcination according to Figure 8a. When only FeCl 3 was enclosed by PANI in the absence of EDA, the obtained single cell demonstrated a slight increase in power and current densities, indicating the active Fe-N covalent bonding cannot be easily created with the covering PANI, although it is also an N source. The max. power (Pmax) and reduced current densities increased with increasing EDA from ratio 1/2 to 1/3 as seen in Figure 8a. The increase of these two types of densities revealed the importance of the complexation between FeCl 3 and EDA which allow the Fe 3+ to firmly capture the N of EDA through the coordinated bonding before calcination and converted into FeNxC covalently networks after calcination. However, when the ratio of FeCl 3 /EDA is over 1/3 (the stoichiometric ratio), additional EDA (1/4) did not bring the increase of either power or current density, but almost destroyed the entire power generation capability as shown in Figure 8a. We understand the excess EDA would reduce more tiny Fe elements which can seal all the micro-and mesopores as already seen in SEM micrographs in Figure 5e.

Stability Test
A simple stability testing was performed by cycling the electrode between 0.0 and 0.9 V for 1000 cycles (about 3 h) in O2 saturated, 0.5 M H2SO4 (aq.) and forced the corrosion of the FeNxC catalyst, leading to the decrease of its reduced current. The half-wave-voltage (HWV) was measured after 1000 The Pmax can further increase to 140 mW·cm −2 when FeNxC (FeNC-900-800-A) went through a second stage of calcination in NH 3 in accordance with Figure 8b, which also demonstrates that Pmax increased with treated temperatures at one-stage calcination, similar to the results obtained from LSVs (Figure 7b).

Stability Test
A simple stability testing was performed by cycling the electrode between 0.0 and 0.9 V for 1000 cycles (about 3 h) in O 2 saturated, 0.5 M H 2 SO 4 (aq.) and forced the corrosion of the FeNxC catalyst, leading to the decrease of its reduced current. The half-wave-voltage (HWV) was measured after 1000 cycles at 1600 rpm for FeNC-900-800-A ( Figure 9a) and compared with commercial Pt/C catalyst in Figure 9b. The HWV loss for FeNxC catalyst is only 10 mV compared to 27 mV for Pt/C, which proves the non-precious FeNxC catalyst is more resistant to the harsh acid electrolyte.
Polymers 2019, 11, x FOR PEER REVIEW 19 of 22 cycles at 1600 rpm for FeNC-900-800-A ( Figure 9a) and compared with commercial Pt/C catalyst in Figure 9b. The HWV loss for FeNxC catalyst is only 10 mV compared to 27 mV for Pt/C, which proves the non-precious FeNxC catalyst is more resistant to the harsh acid electrolyte. Figure 9. Accelerated durability test of (a) FeNC-900-800-A (b) commercial Pt/C catalyst.

Conclusions
In order to create and enhance the covalent bonding between Fe and N, FeCl3 was complexed (neutralized) with EDA in the beginning, followed by the protection (covering) of PANI before thermal calcination to become an FeNxC catalyst.

Conclusions
In order to create and enhance the covalent bonding between Fe and N, FeCl 3 was complexed (neutralized) with EDA in the beginning, followed by the protection (covering) of PANI before thermal calcination to become an FeNxC catalyst. FeCl 3 was found to be able to complex with three EDAs (ligands) which occupied the six coordinated sites of the octahedral structure of Fe 3+ . The complexation brought the firm association of N with Fe 3+ , leading to the formation of Fe-N-C covalent bonding after calcination under the protection of PANI. The optimal ratio of FeCl 3 /EDA for better electrocatalytic performance of the resultant FeNxC catalyst is 1/3, as confirmed by IR-spectra, X-ray diffraction, LSV curves, SEM micrographs, and the higher power density of the obtained single cell.
Most of the FeNxC catalysts generated FeS crystals after calcination due to the persulfate initiator that applied to prepare PANI shell, which can be removed by acid-leaching according to X-ray diffraction. However, if FeNxC experiences second calcination in NH 3 , magnetic Fe 3 O 4 is the dominant product that can also be removed by acid.
It was also found that the FeNxC, which experiences two-stage calcination (FeNC-900-800-A), one in N 2 and the other in mixed gases of N 2 and NH 3 , demonstrated the best ORR performance and obtained a higher power density due to the presence of high concentrations of both micro-and mesopores as characterized by SEM micrographs. The higher surface area of FeNC-900-800-A measured by BET method also contributed to the better electrocatalytic performance. The active sites of the FeNxCs were confirmed by XPS as the pyridinic, pyrrolic, and graphitic N centers, which can also accommodate the O 2 and improve ORR.
The FeNxC based catalysts showed excellent performance and stability for ORR, which implies that it is a promising candidate as an electrocatalyst in PEMFC. In the future, we would try to substitute Fe with Co (Cobalt) and provides other types of amino-complexes with Co to further increase the electrocatalytic performance of the non-precious metal catalysts.

Conflicts of Interest:
The authors declare no conflict of interest.