The design and development of high-activity and low-cost catalysts can effectively improve the slow kinetics of the oxygen reduction reaction (ORR), which is a fundamental reaction for some related technologies, such as fuel cells and metal-air batteries [1,2]. Platinum (Pt) and its alloys are considered to be the most potential oxygen reduction catalysts due to their high activity. However, the high cost and poor stability of Pt seriously hinder its large-scale application in commercial fuel cells. In addition, Pt has other drawbacks, such as the poor durability and the ease with which it can be poisoned [3–5]. Currently, transition metal supported on nitrogen-doped carbon (M-N-C) materials have been extensively investigated [6–8], and are considered more promising candidates that can replace Pt or other precious metal-based ORR catalysts [9,10]. Compared to noble metal-based ORR electrocatalysts, M-N-C catalysts generally have some advantages, such as low raw materials cost, simple synthesis route, and excellent ORR performance [11–13]. The conventional method for preparation of these materials consists of pyrolyzing the mixed precursor containing metallic elements (M = Fe, Co, etc.), carbon elements, and doping with nitrogen atoms [14–17].

It is generally accepted that not only the amount and type of doped-N atoms but also the porosity, surface area and oxygen transport resistance of M-N-C can also affect the ORR performance [18,19]. The morphology and structure of M-N-C catalysts can be tuned by choosing the appropriate carbon source and nitrogen source to achieve an excellent oxygen reduction performance [20–22]. In previous studies, polyaniline (PANI) [23,24], dicyandiamide (DCDA) [25,26], and melamine (MLMN) [27–29] have been the most commonly used nitrogen sources. DCDA is an aminonitrile dimer, PANI is an aniline polymer, and MLMN contains three amino groups, while some C atoms in the benzene ring are also replaced by N atoms. There exist conjugated structures in PANI, DCDA, and MLMN molecules, and the delocalized π electrons they provided can facilitate the N atoms doped carbon skeleton during pyrolysis [30]. In addition, DCDA and MLMN have high ratio of nitrogen to carbon, PANI is a polymer with good mechanical properties, and the three-dimensional structure is conducive to improve the distribution density of catalytically active sites [31]. Although PANI, DCDA, or MLMN alone have previously been used as nitrogen precursors for M-N-C synthesis, in-depth investigation of M-N-C catalysts prepared by choosing dual nitrogen sources has been rarely reported.

In this work, Fe-PD/PGO (Fe-PANI/DCDA/graphene oxide modified by p-phenylenediamine) nanosheets were synthesized in ice-water bath using PGO as carbon carrier, PANI, and DCDA as dual nitrogen sources, FeCl₃·6H₂O as single metal source, and Fe-PD/PGO catalytic materials were prepared by pyrolysis. In addition, the effects of different PANI/DCDA addition ratios, metal content, and pyrolysis temperature on the ORR activity of Fe-PD/PGO catalysts were analyzed. The result shows that the structural defects in the Fe-PD/PGO catalytic increase, and the degree of graphitization decreases, which further increase the content of N atoms, thus improving the distribution density of catalytic active sites in the material.

Graphite oxide (GO) was prepared by utilizing the modified Hummers method, and through an amide reaction to synthesis the p-phenylenediamine (PPD) modified GO, denoted as PGO. First, 1.0 g of the GO prepared was transferred to 1.0 L ultrapure water and dispersed for 0.5 h under ultrasonic conditions. Then 16 mmol of PPD and 16 mmol of H₂SO₄ were added to the GO dispersion subsequently. Meanwhile, 16 mmol of NaNO₂ was slowly dissolved in 10 mL of ultrapure water, and the reaction was carried out at 60 °C for 4.0 h. NaNO₂ was used to stabilize acidic conditions and under GO an azo reaction with aniline, which favors the reaction of aniline with graphene oxide to form PGO. After centrifuged the suspension, it was washed with ethanol, and then freeze-dried to obtain PGO for storage. Before use, the PGO was immersed in 0.1 mol/L of hydrochloric acid for two hours to remove any metal impurities that may be present in the PGO.

0.2 g of acid-soaked PGO was added to 0.2 L of hydrochloric acid solution (0.5 mol/L) and dispersed in an ultrasonic stirrer for 1 h to make it homogeneous. Then, 1 mL aniline and a certain amount of DCDA were added to the FeCl₃·6H₂O mixture which was put into an ice water bath to bring the temperature of the reaction system below 5 °C. Thereafter, 0.5 mol/L of HCl was used as the solvent and ammonium persulfate as the solute to prepare a solution (0.2 g/mL) which was slowly dripped into the mixture system using a constant-speed dropping funnel. The mixture was stirred continuously for some time to allow the aniline to polymerize completely. Afterwards, solid-liquid separation was conducted by rotary evaporation, and then dried again. The sample dried was heated at 5 °C/min to 900 °C for 1 h under nitrogen conditions with a flow rate of 0.2 L/min. Finally, the product was immersed in 0.5 mol/L H₂SO₄ and heated at 80 °C for 12 h to remove unstable materials. The precursors were then washed and dried several times by filtration and deionized water, and heated for the second time at 5 °C/min to 900 °C for 3 h, with a nitrogen flow rate of 0.2 L/min. After grinding, the final product was Fe-PD/PGO. Fe-DCDA/PGO and Fe-PANI/PGO were synthesized by adding only DCDA or aniline.

The structure and degree of crystallization of Fe-DCDA/PGO, Fe-PANI/PGO, and Fe-PD/PGO samples were first tested by X-ray diffraction (XRD). In Fig.1(a), the Fe-DCDA/PGO, Fe-PANI/PGO, and Fe-PD/PGO samples show characteristic peaks located around 26.1°, and 42.9°, well assigned to the (002) and (101) planes of carbon, respectively [32]. The height of the peak can represent the perfection of the crystal type, the more perfect the grain arrangement, the higher the peak. Among them, the (002) peak of the Fe-DCDA/PGO sample is the sharpest, indicating that the carbon materials have a higher degree of graphitization, while the broader (002) peak of Fe-PD/PGO indicates a decrease in its degree of graphitization [33]. Furthermore, the diffraction peak of the Fe-DCDA/PGO sample centered at 29.5° belongs to the (111) plane of Fe₃C (JCPDS: 035-0772) [21,34], while the peaks of Fe-PANI/PGO and Fe-PD/PGO samples at 35.2° are attributed to the (200) planes of Fe₃C, indicating that there is some Fe element in the form of Fe₃C in the samples obtained. The Fe/Fe₃C nanocrystals boost the activity of Fe-N_x, which is essential for the high ORR performance [35,36]. In addition, Raman spectroscopy analysis was used to characterize the carbon structure of the samples (Fig.1(b)). The I_D/I_G value^†

¹⁾I_D/I_G represents the intensity ratio of D to G band (I_D/I_G), which has been used as an effective method to characterize the graphitization degree in carbon materials.

of Fe-PD/PGO is 1.02, which is significantly higher than those of Fe-PANI/PGO (0.91), Fe-DCDA/PGO (0.84), and PGO (0.85), (cf., Fig. S1 in Electronic Supplementary Material), indicating that the Fe-PD/PGO sample has the highest amount of carbon structural defects, the highest disorder, and the lowest degree of graphitization among the three samples, which is in agreement with previous XRD results [37]. Therefore, it can be concluded that by choosing the right nitrogen sources, the product obtained can have a low degree of graphitization, along with a large number of carbon structural defects, which are highly desirable to achieve an excellent ORR performance [38].

A comparison of Figs. S3 and S2 indicates that the pyrolyzed catalyst has a stacked lamellar structure. The three-dimensional porous structure in Fe-PD/PGO materials is more pronounced, and the layers are interconnected through porous carbon structures to form a network structure. As shown in Fig.2, the Fe-DCDA/PGO material exhibits a silk-like microstructure, and the Fe-PANI/PGO catalytic material contains some fibrous structures. The morphology of Fe-PD/PGO differs significantly from that of Fe-DCDA/PGO and Fe-PANI/PGO materials. The silk-like structure is no longer obvious, and many fibrous structures appear, which is consistent with the three-dimensional network structure observed in SEM.

As illustrated in Fig.3(a) and Fig.3(b), a typical N₂ adsorption–desorption isotherms type-IV with a hysteresis loop can be observed in the section of relatively high pressures, suggesting that all catalysts contain the mesoporous structure. In the case of Fe-PD/PGO, the volume adsorbed demonstrates a rapid increase at a relatively low pressure (P/P₀ = 0–0.01), indicating the existence of abundant micropores. The pore size distribution of Fe-PD/PGO catalytic materials is concentrated below 2.0 nm. The surface areas of Fe-DCDA/PGO, Fe-PANI/PGO, and Fe-PD/PGO samples were calculated to be 305.4, 359.8, and 639.1 m²/g, respectively. Moreover, the pore volumes were 0.33, 0.48, and 0.86 cm³/g, respectively, indicating that the use of DCDA and PANI as dual nitrogen sources can effectively enlarge the specific surface area and cumulative pore volume. In fact, the large specific surface area and the high proportion of micropores in the catalyst favor the exposure of more catalytic active sites, and improve the probability of their contact with O₂, thus improving the ORR performance of the materials [39,40]. In addition, voltametric tests were performed on Fe-DCDA/PGO, Fe-PANI/PGO, and Fe-PD/PGO samples in 0.1 mol/L of KOH to measure the electrochemical active surface area [41]. Compared with the specific surface area of the catalytic material itself, its corresponding electrochemically active surface area (EASA) is lower, indicating that the specific surface area possessed by the catalytic material cannot be fully utilized in the electrochemical reactions. Therefore, using EASA (converted from cm² to m²/g by dividing the EASA area by the catalysts mass loading) to study the ORR activity of catalytic materials is straightforward. The EASA of Fe-DCDA/PGO, Fe-PANI/PGO, and Fe-PD/PGO catalytic materials are 191.4, 283.4, and 317.5 m²/g, respectively (Fig.3(c)). Combined with the ORR performance tests, it is found that the EASA size and the ORR activity level have a similar trend, i.e., the ORR activity of the catalyst increases with the increase of EASA [42].

The high-resolution XPS spectra of N 1s for Fe-PANI/PGO, Fe-DCDA/PGO, and Fe-PD/PGO were displayed in Fig.4. The N 1s can be roughly divided into four peaks with binding energies centered at 398.7 ± 0.2, 400.1 ± 0.2, 401.5 ± 0.3, and ≈ 403.5 eV, corresponding to pyridinic N (N-P), pyrrolic N (N-Pr), graphitic N (N-G), and oxidized pyridinic N (N-OP) [28]. The content of different types of N is shown in Tab.1. Compared with Fe-DCDA/PGO and Fe-PANI/PGO, the N atom content in Fe-PD/PGO catalytic materials is further increased, making them have a higher N-P/N-G value (0.84). This may be due to the synergistic promotion of doping with the N atom by the presence of DCDA and PANI. Meanwhile, PANI can form a structurally stable six-membered ring structure containing two N atoms under the action of oxidants during the polymerization process, effectively promoting the formation of N-P and N-G, among which N-P is believed to favor the performance of the catalysts [12,30,43]. From Fig. S4, it can be seen, that Fe-DCDA/PGO and Fe-PANI/PGO materials have emerging emission peaks at 724.3 and 710.9 eV, with the Fe 2p_3/2 peak, in the high-energy region (724.3 eV), caused by Fe (III); and the Fe 2p_1/2 peak, in the low-energy region (710.9 eV), caused by Fe²⁺ or Fe³⁺ bound to N atoms [44]. The Fe-N_x structure is widely recognized as an active site with a catalytic ability [45,46]. The synergy interaction at the Fe-N_x nanoparticles/N-C interface, certainly redistribute the electron density of the active sites [47,48].

As shown in Fig.5(a), there is no obvious redox peak in the cyclic voltammetry (CV) curves of the different catalysts in the N₂-saturated system. However, when tested in the O₂-saturated system, the Fe-DCDA/PGO catalyst produces a reduction peak at 0.795 V vs. reversible hydrogen electrode (RHE). Combined with CV tests under N₂ saturated conditions, this peak can be determined to be the O₂ reduction peak. Compared with Fe-DCDA/PGO materials, the reduction peak potentials of Fe-PANI/PGO, Fe-PD/PGO, and Pt/C were shifted positively by 20, 41, and 39 mV, respectively, indicating that Fe-PD/PGO has a comparable ORR performance. The onset potentials (E_onset) were determined at the current density of 0.1 mA/cm². The E_onset, E_1/2, and J_k@0.9V of Fe-PD/PGO materials reaches 0.951 V (vs. RHE), 0.886 V (vs. RHE), and 3.178 mA/cm², respectively, much higher than those of Fe-DCDA/PGO and Fe-PANI/PGO materials, and even exceeded the E_onset (0.932 V vs. RHE) and E_1/2 (0.849 V vs. RHE) of Pt/C catalysts (Fig.5(b) and Fig.5(d)). The corresponding Tafel slope was calculated based on the LSV curve mentioned above, and the specific slope is shown in Table S1. The Tafel slope calculated for Fe-PD/PGO is 76.8 mV/dec¹, which is lower than that of other catalysts, indicating relatively faster ORR kinetics with Fe-PD/PGO [24].

At the same time, to study the changes in the number of transferred electrons (n) of the catalytic materials mentioned above during the ORR process, LSV tests were performed on Fe-PANI/PGO, Fe-PD/PGO, and Pt/C materials at different rotation rates under the same testing conditions, and the corresponding Koutecky–Levich (K–L) equations were calculated (Fig. S5). The number of transferred electrons of the Fe-PD/PGO catalytic material is 3.98, which is higher than that of Fe-PANI/PGO (3.62), and is equivalent to the number of transferred electrons of Pt/C catalyst (3.98), indicating that the Fe-PD/PGO catalytic material has a high electron transfer efficiency in the ORR process, which agrees with the conclusion obtained by Tafel analysis.

The effect of different PD molar ratios on the ORR performance of Fe-PD/PGO catalytic materials were also analyzed. Figure S6 shows the CVs, linear sweep voltammetries (LSVs), Tafel, and performance parameter plots of Fe-PD (1:1/2:3/1:2/2:5/1:3)/PGO catalysts, respectively. According to the CV test in Fig. S6(a), there is a significant oxygen reduction peak at 0.786 V (vs. RHE) for Fe-PD (1:1)/PGO in an O₂-saturated electrolyte.

Compared with the Fe-PD (1:1)/PGO catalyst, the peak potentials of Fe-PD (2:3/1:2/2:5/1:3)/PGO were positively shifted by 30, −17, 37, and 34 mV, respectively, indicating that the Fe-PD (2:5)/PGO catalytic material has the best ORR activity under alkaline conditions. To accurately evaluate the ORR performance of the different catalytic materials, further LSV tests were performed. As shown in Figs. S6(b) and S6(d), the Fe-PD (2:5)/PGO catalytic material shows a significant improvement in E_onset, E_1/2, and J_k@0.9 V, with E_onset and E_1/2 reaching 0.951 V (vs. RHE) and 0.886 V (vs. RHE), respectively. Therefore, when the PANI/DCDA molar ratio is 2:5, the Fe-PD/PGO catalytic material prepared has the best ORR activity and relatively faster ORR kinetics (Table S2).

The ORR activity of Fe-PD/PGO catalytic material first increases and then decreases with increasing PD ratio, which is mainly caused by the fact that the addition of DCDA can synergize with PANI, which increases the defects in the catalytic material during pyrolysis, and then allows more N atoms to be incorporated into the carbon structure, improving the distribution density of the catalytic active site. However, an excess of DCDA can significantly reduce the PANI content in the precursor, and PANI can form a stable six-membered ring structure containing two N atoms during the polymerization process, which can effectively promote N-P and N-G formation. Therefore, an excess of DCDA can weaken the ORR performance of the catalytic material.

The durability of the catalysts was evaluated, as shown in Fig.6(a) and Fig.6(b). After 5000 and 10000 CV cycles, the E_onset of the LSV curve of the Fe-PD/PGO catalytic material measured had almost no attenuation, while E_1/2 only decreased by 6 and 7 mV, respectively; under the same testing conditions, after 10000 cycles of CV cycles, the E_onset and E_1/2 of the LSV curves of the Pt/C catalyst measured decreased by 19 and 25 mV, respectively. Additionally, the durability of Fe-PD/PGO was also tested by chronoamperometry. The current attenuation of the catalyst is very slow, and the current density still maintains 95.6% after 30000 s (Fig.6(c)). Therefore, the Fe-PD/PGO material shows a good electrochemical stability performance.

In addition, the catalyst performance in terms of resistance to methanol tolerance was tested by chronoamperometry, j(t), at 0.7 V (vs. RHE in an O₂-saturated electrolyte, without rotation) CH₃OH (3.0 mol/L) was added at 100 s to evaluate the methanol tolerance resistance. As illustrated in Fig.6 (d), compared with Fe-PD/PGO, Pt/C catalyst shows a higher current density at the initial stage of the reaction. However, with the addition of CH₃OH, its current density sharply decreases and cannot return to its original state. The main reason for this is that CH₃OH undergoes an oxidation reaction on the catalyst surface, and the oxidation current generated interacts with the ORR reduction current, resulting in a mixed electrode potential formation leading to a decrease. The current densities of Fe-PD/PGO changes only slightly after experiencing brief fluctuations, indicating that under the same conditions, Fe-PD/PGO presents a better methanol tolerance than Pt/C catalysts.

In this work, the Fe-PD/PGO catalyst was prepared by two-step pyrolysis using PGO as carbon carrier, DCDA and PANI as dual nitrogen sources, and FeCl₃·6H₂O as a single metal source. In material synthesis, PANI preferentially generates a coarser carbon planar structure with a relatively lower degree of graphitization, a higher content of nitrogen and iron, and dominant micropores. On the other hand, the graphite produced by DCDA has a finer structure, a higher degree of graphitization, a lower nitrogen and iron content, and a higher mesoporous/microporous ratio. Therefore, the dual nitrogen sources not only enhance the doping amount of N atoms in the Fe-PD/PGO catalytic material, but also maintain a high N-P/N-G value (0.84), favorable in increasing the distribution density of catalytic active sites in the material. In addition, the high specific surface area of Fe-PD/PGO and EASA can fully expose active sites on the surface of the material, which improves the contact probability with O₂ molecules. The E_onset, E_1/2 and, J_k@0.9V of Fe-PD/PGO reach 0.951 V (vs. RHE), 0.886 V (vs. RHE), and 3.178 mA/cm, respectively. Compared with Pt/C, E_onset and E_1/2 have a positive shift of 19 and 37 mV, respectively, and the electron transfer number (n) of Fe-PD/PGO in the ORR process reaches 3.98. Moreover, the Fe-PD/PGO electrocatalysts show a better electrochemical stability and methanol resistance than Pt/C catalyst.