Porous Organic Polymer-Derived Fe2P@N,P-Codoped Porous Carbon as Efficient Electrocatalysts for pH Universal ORR.

A new porous organic polymer (CP-CMP) was designed and synthesized via the direct polymerization of pyrrole and hexakis(4-formyl-phenoxy)cyclotriphosphazene, skipping the tedious synthetic procedure of porphyrin-monomers containing special groups. This special porous organic polymer (POP) serves as an "all in one" precursor for C, N, P, and Fe. Direct carbonization of this special POP afforded Fe2P@N,P-codoped porous carbons with hierarchical pore structure and high graphitization. Finally, the optimal catalyst (CP-CMP-900) prepared by carbonization of CP-CMP at 900 °C exhibited high efficiency for oxygen electroreduction. Typically, CP-CMP-900 presented an oxygen reduction reaction half-wave potential (E 1/2) of 0.85, 0.73, and 0.65 V, respectively, in alkaline, neutral, and acidic media, close to those of commercial Pt/C in the same electrolyte (0.843, 0.71, and 0.74 V). Furthermore, it also displayed excellent methanol immunity and long-time stability in various electrolytes better than commercial Pt/C (20%).

Introduction

The depletion of traditional energy has compelled us to explore novel energy devices meeting the growing needs of social development.[1] Fuel cells, as eco-friendly energy storage and conversion devices with a high energy density and potential, have geared great research attention over the past few years.[2] However, the performance of these devices is mainly determined by the sluggish oxygen reduction reaction (ORR) that occurred at the cathode. Due to their scarcity and low sustainability, state-of-the-art Pt-based catalysts are undesirable.[3] For the development of the green energy technology, the availability of low-cost, efficient, and durable catalysts for ORR is a prerequisite.[4−6]

Owing to the features of low price, high earth abundance, and environmental friendliness, carbon-based materials have emerged as realistic alternatives, to which considerable efforts have been devoted.[7−9] The activity of these materials can be finely regulated by doping a transition metal (Fe, Mo, Co, Cu, Ni, Mn, etc.) or heteroatom (N, P, S, B, F, etc.) into the carbon skeleton.[10−13] For example, the association of heteroatoms could modify the electronic structure (charge and/or spin density redistribution) of carbon, creating active sites favorable for the adsorption capacity of oxygen and facilitating the ORR process.[14] And the introduction of transition-metal impurities is conducive to the formation of active sites, improving the catalytic performance significantly.[15,16]

Among the reported materials, due to their low cost and special stability, transition-metal phosphide-based catalysts have been widely studied.[17,18] Hitherto, numerous metal phosphide-based catalysts have been well designed and fabricated, including the linear, spherical, and flaky. And great progress has been achieved.[19] But for the synthesis of most catalysts, an extraneous phosphorus source that is hazardous, inflammable, and explosive is inevitable. And in most cases, to complete the phosphating reaction, the extraneous phosphorus source, such as organic triphenylphosphine, phosphonitrile, inorganic red phosphorus, phytic acid, and sodium hypophosphite, is in large excess, which is not conducive to study the formation mechanism of these materials. Furthermore, there are diversiform hypophosphites formed simultaneously during the phosphating process, for example, the formation of FeP and Fe2P.[3,20] Hence, it remains a huge challenge to obtain metal phosphides in a simple, safe, green, and controllable method.[21]

As a new series of multifunctional materials composed of pure organic units, porous organic polymers (POPs) have received great research interests in recent years.[22−25] The inherent nature, including the diversity of synthetic methods and the designability of structure, makes it available in numerous fields. Various monomers with special function and composition could be introduced into the porous skeleton facilely.[26,27] Furthermore, the organic structure of POPs enables them to have a definite element composition at the atom level (atomic ratio and atom spatial distance), which is beneficial to further research. Hitherto, thousands of POPs with different structures and functions have been prepared.[28−30] Recent reports have validated that direct pyrolysis of POPs is a simple method to prepare carbon-based catalysts for electrocatalysis.[31] However, there are almost no reports about the direct preparation of pure Fe2P nanoparticles-containing catalysts from POPs.

Owing to the special structure of metal porphyrin, catalysts derived from porphyrin-based POPs usually exhibited prominent activity toward ORR.[32] However, most of these polymers were synthesized via the polymerization of porphyrin-monomer containing special groups, in most cases, with a noble-metal catalyst under harsh conditions.[33] Apart from the high production cost, the synthetic procedure is tedious and the overall yield is usually low.[34] Hence, it would be significant to prepare porphyrin-based POPs in a simple way, skipping the complicated synthesis procedure of porphyrin-monomer containing a special group.

Since the Bhaumik group reported the first example of porphyrin-based POPs that skipped the complicated synthesis procedure, a lot of work has been done. Herein, we propose a facile and modified one-step polymerization method.[35−38] A cyclotriphosphazene (CTP) and Fe-porphyrin (Fe-Por) constituent porous polymer with high specific areas (CP-CMP) and accurate ratio of elements was prepared in the solvent mixture of propionic acid and nitrobenzene with a high yield.[39−42] This special POP could serve as an “all in one” precursor of C, N, P, and Fe simultaneously. Direct carbonization of this special POP afforded N,P-codoped porous carbons embedded with well-defined Fe2P nanoparticles. Subsequent tests validated that the typical catalyst denoted as CP-CMP-900 exhibited excellent catalytic performance toward ORR with a 4e transfer pathway under various pH values.

Results and Discussion

As presented in Figure , CP-CMP was prepared via the copolymerization of hexa-(4-aldehyde-phenoxy)-cyclotriphosphazene and pyrrole in the presence of iron salts. And TB-CMP was prepared by the polymerization of 1,3,5-triformylbenzene and pyrrole under identical reaction conditions. CP-CMP-X was obtained through the direct carbonization of CP-CMP at various temperatures (800, 900, and 1000 °C). For comparison, TB-CMP-900 was fabricated by direct pyrolysis of TB-CMP at 900 °C.

Figure 1

Typical synthesis procedure for CP-CMP and schematic route for the preparation of CP-CMP-X catalysts.

The integration of CTP and porphyrin moieties was confirmed by Fourier transform infrared spectroscopy (FTIR, Figure S3) and solid-state 13C cross-polarization magic angle spinning (CP/MAS) NMR (Figure S4). As shown in Figure S3, characteristic vibration bands of CTP (P=N–P at 1218 and 1410 cm–1) and metal porphyrin (1602 cm–1 for C=C stretch of pyrrole together with 1000 and 1160 cm–1 for chelated M–N4 vibrations) could be clearly observed for all of these prepared polymers, indicative of the formation of porous skeletons.[43] Furthermore, the solid-state 13C NMR spectrum of CP-CMP and TB-CMP exhibits typical carbon signals (ca. 120, 132, and 155 ppm) assigned to the porphyrin macrocycles structure, further implying the successful construction of porous networks.[44]

TGA was performed to investigate the thermal stability of prepared polymers. As shown in Figure S5, CP-CMP presents excellent thermal stability with almost no weight loss until 210 °C, and the weight could maintain at 66% of the initial value at 800 °C, indicating a high thermal stability of prepared materials. Figure S6 presents the powder XRD pattern of prepared samples. Like other reported CMPs, no clear peaks could be found for all of these prepared polymers, implying the amorphous nature of the polymer precursors.[45]Figure a exhibits the PXRD pattern of CP-CMP-X. Different from the precursors, a prominent peak assignable to the diffraction of graphitic carbon appeared at 26.5°. Other peaks located at 35.3, 40.3, 44.3, 47.3, 53.0, 54.3, and 73.7° are ascribed to the characteristic peaks of Fe2P nanocrystalline (PDF #85-1725),[25] evidencing the formation of pure Fe2P nanoparticles, which is beneficial for the ORR in the carbonized samples (Figure a). The Raman spectrum was examined (Figure b) to investigate the graphitic degree of CP-CMP-X, from which a prominent D band (1335 cm–1) and G band (1580 cm–1) could be clearly observed. The ratio of intensities (ID/IG) was calculated according to the integral area. All of the catalysts derived from CP-CMP showed a high degree of graphitization, and CP-CMP-900 exhibits the highest graphitic degree (ID/IG = 0.84) among the prepared samples (ID/IG = 0.85 for CP-CMP-800 and ID/IG = 0.89 for CP-CMP-1000).

Figure 2

(a) Powder XRD pattern of CP-CMP-X catalysts (X is the carbonization temperature). (b) Raman spectrum of CP-CMP-X.

The morphology and inner structure of prepared samples are visualized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As exhibited in Figures a and S7, bulks stacked by intergrown spheroidal particles could be clearly observed for all of the prepared polymers. Figures b,c and S8 present the TEM image of prepared materials, from which a large number of pores could be observed. Figure d shows the SEM images of CP-CMP-900. Like the polymer precursor, spheroidal particles stacked loosely could be obviously observed, further implying the excellent thermal stability of the prepared polymers. Figure e,f displays the TEM and HR-TEM images of CP-CMP-900. As indicated in Figure f, spherical Fe2P nanoparticles with a diameter of 10 nm are distributed uniformly in the carbon skeletons. Clear lattice fringes belonging to the Fe2P specie could also be observed in Figures f and S9. As marked in Figures f and S9f, a regular interplanar spacing of about 0.22 nm ascribing to the (210) plane of the Fe2P nanocrystalline could be detected.[20] To further confirm the composition of prepared catalysts, TEM–energy-dispersive spectroscopy (EDS) elemental mapping was examined. As presented in Figure g, uniformly distributed C, N, P, O, and Fe over the porous skeletons could be clearly detected. Furthermore, the EDS of HR-TEM (Figure S10) also evidenced the coexistence of Fe, N, C, and P in the porous skeletons. The inset in Figure g exhibited the corresponding SAED pattern of CP-CMP-900.

Figure 3

SEM and HR-TEM images of prepared materials: (a) SEM image of CP-CMP at a scale bar of 200 nm; (b, c) TEM and HR-TEM images of CP-CMP at a scale bar of 100 and 10 nm, respectively; (d) SEM image of CP-CMP-900 at a scale bar of 200 nm; (e, f) HR-TEM images of CP-CMP-900 at a scale bar of 20 and 5 nm, respectively; (g) TEM and corresponding EDS layered images of CP-CMP-900 at a scale bar of 50 nm.

The pore features of prepared samples were investigated by low-temperature nitrogen absorption and desorption measurements at 77 K. As indicated in Figure , both the polymers and catalysts exhibit typical isotherm combining the characteristics of type I and type IV, featured by a sharp uptake in the low-pressure region (micropore) and a large hysteresis loop (mesopore) in the high-pressure range. Furthermore, continued growth could also be found in the high-pressure range, indicative of the existence of macropore.[46,47] In general, a high specific surface area could increase the exposure of active sites and facilitate mass transport, which is beneficial to the ORR process. And the rich mesopore distribution is very significant for the migration of electrolyte ions, improving the catalytic performance. After the pyrolysis, the BET surface areas increased greatly. The calculated BET surface areas are 207 and 866 m2 g–1 for CP-CMP and CP-CMP-900, respectively. Furthermore, the BET surface area of TB-CMP-900 reaches 598.3 m2 g–1, much higher than that of the uncarbonized sample. Hierarchical pore ranging from micropore to mesopore could also be found from the pore width distribution curve (Figure b,d). For example, as exhibited in Figure d, a sharp peak at around 3.8 nm and other two secondary peaks (1.2 and 1.4 nm) were observed from the pore size distribution of CP-CMP-900. And this indicated that CP-CMP-900 holds a large amount of mesoporous structure with a relatively narrow pore size distribution. And the detailed information is summarized in Table S1.

Figure 4

BET and pore size distribution of prepared polymers and typical catalysts. (a) Low-temperature N2 adsorption and desorption isotherm of CP-CMP and TB-CMP; (b) low-temperature N2 adsorption and desorption isotherm of CP-CMP-X and TB-CMP-900; (c) pore size distribution of CP-CMP and TB-CMP; and (d) pore size distribution of CP-CMP-X and TB-CMP-900.

X-ray photoelectron spectroscopy further validates the coexistence of C, N, P, and Fe for CP-CMP-X samples (Figures a and S11–S13). The high-resolution N 1s spectrum of CP-CMP-900 (Figure b) could be divided into five peaks located at 398.4 eV (0.55 atom %), 399.3 eV (0.34 atom %), 399.8 eV (0.09 atom %), 400.9 eV (1.70 atom %), and 403.1 eV (0.32 atom %), ascribed to the pyridinic-, metallic-, pyrrolic-, graphitic-, and oxidized-N, respectively.[48] And among these deconvoluted N species, the content of pyridinic-N and graphitic-N, which could improve the activity toward ORR significantly, occupied a larger proportion in surface nitrogen types.[49] Furthermore, the total content of pyridinic and graphitic-N for CP-CMP-900 is higher than those of CP-CMP-800 and CP-CMP-1000. Figure c presents the 2p spectra of P. As exhibited in Figure c, the peaks of P 2p are distributed at 129.4, 132.1, 133.6, and 134.3 eV, assigned to the Fe–P, P–C, P–O, and P–O–Fe, respectively.[3,20] The existence of the P–C bond further demonstrated the doping of P into the N-doped porous carbon. Furthermore, the Fe-Nx and Fe2P species could also be proved by XPS (Figure d). The high-resolution Fe 2p spectra of CP-CMP-900 could be deconvoluted into seven peaks. The peaks appearing at 707.2 and 711.0 eV are assigned to the Fe–P and Fe–N bonds, respectively. Other peaks attributed to Fe3+ 2p1/2, Fe2+ 2p1/2, Fe3+ 2p3/2, Fe2+ 2p3/2, and satellite peak are situated at 726.2, 723.3, 712.6, 709.8, and 717.6 eV, respectively.[50] The detailed elemental analysis by XPS is shown in Table S2. Taken all of these together, a prominent electrocatalytic activity toward ORR is anticipated for the Fe2P- and FeNx-decorated N, P-doped carbon.[51]

Figure 5

XPS spectrum of CP-CMP-900. (a) Survey spectrum of CP-CMP-900; (b) high-resolution N 1s XPS spectra of CP-CMP-900; (c) high-resolution P 2p XPS spectra of CP-CMP-900; and (d) high-resolution Fe 2p XPS spectra of CP-CMP-900.

The catalytic activities toward the ORR of pyrolyzed porous organic polymers (CP-CMP-X and TB-CMP-900) were assessed in various electrolytes including KOH (0.1 M), HClO4 (0.5 M), and PBS (0.1 M) via cyclic voltammetry (CV) and linear sweep voltammetry (LSV) on rotating disk electrode (RDE) or rotating ring-disk electrode (RRDE). The onset potential was defined as the potential at the current density of 0.1 mA cm–2. The methanol tolerance test was performed by the current–time (i–t) plots with the addition of methanol to the corresponding electrolyte at the time of 400 s. TB-CMP-900 was a reference catalyst free of P and Fe2P.

As exhibited in Figures S14–S16, the electrical signals are virtually featureless in the Ar-saturated KOH solution within the entire potential range, while in the O2-saturated electrolyte, a well-defined cathodic peak corresponding to ORR was obviously observed in the case of all catalysts.[52,53] We optimize the carbonization temperature via testing the polarization curve of CP-CMP-X. Figure a presents the LSV curves of the as-prepared catalysts together with commercial Pt/C measured on an RDE with a scan rate of 5 mV s–1 at 1600 rpm. Notably, CP-CMP-900 afforded the most positive onset potential of 0.997 V (vs RHE), close to the value of commercial Pt/C (0.996) and higher than that of other CP-CMP-X samples (0.983 V for CP-CMP-800 and 0.986 V for CP-CMP-1000). Furthermore, the Eonset of CP-CMP-900 is also positive compared to that of TB-CMP-900 (Eonset = 0.990 V), indicating that CP-CMP-900 was more electrocatalytically active than the control catalysts. Accordingly, CP-CMP-900 showed the highest ORR activity with a half-wave potential (E1/2) of 0.85 V, 11 mV and 7 mV positive than that of commercial Pt/C (E1/2 = 0.843 V) and TB-CMP-900 (E1/2 = 0.839 V), comparable to the performance of other similar reference catalysts previously reported (Table S3).[54−56] Furthermore, one could observe clearly that CP-CMP-900 exhibited excellent catalytic activity toward ORR with a diffusion-limited current density (JL) of 4.78 mA cm–2 (vs 5.19 mA cm–2 of Pt/C). And all of these validated the formation of Fe2P, and the introduction of P into the N-doped porous carbon could enhance the ORR activity. To gain insights into the reaction kinetics of ORR catalyzed by CP-CMP-900, LSV curves at various rotation speeds (from 400 to 2500 rpm) were recorded (Figure b) and fitted according to the Koutecký–Levich (K–L) equation (eqs S1 and S2). As shown in Figure c, almost parallel fitting lines were obtained from the Koutecký–Levich (K–L) plots of CP-CMP-900, implying a first-order reaction kinetics toward the O2 concentration. And the calculated electron transfer number was 3.97 (0.2 V vs RHE) matching pretty well with the results calculated from the RRDE measurements (Figure d) on the basis of eqs S3 and S4. In addition, only low yields of H2O2 (less than 8.0%) were detected, indicative of a four-electron transfer pathway (Figure d) for ORR. In terms of practical applications, a high stability is a prerequisite. Hence, a chronoamperometric test (i–t) was conducted. As shown in Figure e, after a continuous constant potential cycling of 20 000 s, the current of CP-CMP-900 could maintain 84% of the initial value, while a decrease of about 40% was found for commercial Pt/C. Different from Pt/C, the polarization curve measured after the i–t test almost coincided with the previous curve (Figure S15b), validating that CP-CMP-900 has better long-cycle durability than commercial Pt/C.[57,58]Figure f shows the methanol crossover effects of CP-CMP-900. Only a slight current change could be detected on the CP-CMP-900 loaded electrode after the injection of methanol (3.0 M), and it reverts to the previous state with increasing time. However, a dramatic current decrease was observed for Pt/C, indicating that CP-CMP-900 possesses excellent methanol immunity. And this could also be evidenced by the LSV curve obtained before and after the injection of methanol (Figures S14f and S15c). The detailed information about the contrast catalysts is given in the Supporting Information (Figures S14–S16 and Table S2).

Figure 6

Electrochemical performance of prepared catalysts in alkaline (0.1 M KOH) conditions: (a) Polarization curve of prepared catalysts and commercial Pt/C at 1600 rpm in O2-saturated KOH solution with a sweep rate of 5 mV s–1; (b) LSV curve of CP-CMP-900 at various rotation speeds in O2-saturated KOH solution with a sweep rate of 5 mV s–1; (c) K–L plots for CP-CMP-900 at various potentials; (d) percentage of hydrogen peroxide yield and the electron transfer number (n) of CP-CMP-900 at different potentials; (e) durability evaluation from the i–t chronoamperometric responses of the CP-CMP-900 electrodes in aqueous solution of KOH (0.1 M) saturated with O2; and (f) methanol crossover of CP-CMP-900.

Moreover, CP-CMP-900 presents a high catalytic activity in neutral conditions (0.1 M PBS). Figure a shows the CV curve of CP-CMP-900 in the solution saturated with or without oxygen. Obvious cathodic peaks could be observed from the CV curves (black lines), but a featureless peak could be detected in the Ar-saturated solution (red line). As exhibited in Figure b, the Eonset value of CP-CMP-900 is 0.906 V, comparable to that of commercial Pt/C (0.911 V) and higher than that of the phosphorus-free TB-CMP-900 (0.901 V). And the E1/2 value (0.73 V) of CP-CMP-900 (Figure c) is more positive than that of Pt/C and other prepared catalysts (Figures S17 and S19). The calculated electron transfer number (n) is 3.95 at 0.1 V (vs RHE) according to the K–L plots (Figure d), validating a 4e reduction process for the ORR. And the electron transfer number (n) obtained from RRDE measurements further confirmed the 4e pathway for the ORR with a negligible yield of H2O2 (less than 5.0%) in the potential range of 0.1–0.8 V (Figure e). Besides, CP-CMP-900 also exhibits much better long-term stability than Pt/C (Figures f and S18a). A slight peak current decrease of 4.5% (vs 20% of Pt/C) could be observed after a long cycle of 20 000 s. As exhibited in Figure S18b, CP-CMP-900 also possesses excellent methanol immunity far beyond commercial Pt/C in the PBS solution. The details are listed in Table S3 in SI.

Figure 7

Electrochemical performance of prepared catalysts in neutral (0.1 M PBS) conditions: (a) CV of CP-MP-900 in 0.1 M PBS saturated with O2 at a sweep rate of 50 mV s–1; (b) polarization curve of prepared catalysts and commercial Pt/C at a rotation speed of 1600 rpm in O2-saturated PBS solution with a sweep rate of 5 mV s–1; (c) LSV curves of CP-MP-900 at various rotation speeds from 400 to 2500 rpm in O2-saturated 0.1 M PBS solution; (d) K–L plots for CP-CMP-900 at various potentials; (e) percentage of hydrogen peroxide yield and the electron transfer number (n) of CP-MP-900 at different potentials; and (f) long-time stability curves of CP-MP-900 together with the commercial Pt/C in O2-saturated 0.1 M PBS solution.

Meanwhile, the ORR performance in acidic electrolyte is more important for the potential application in PEMFC. CP-CMP-900 also exhibits prominent catalytic performance in acidic electrolyte (0.1 M HClO4). An obvious reduction peak could be found from the CV curves in oxygen-saturated solution (Figure a and Figures S20 and S22) but not in the Ar-saturated media for all of the samples. As shown in Figure b, the Eonset value of CP-CMP-900 is 0.856 V approaching that of Pt/C (1.011 V) and more positive than that of other prepared catalysts. In addition, the E1/2 value of CP-CMP-900 is 0.65 V (vs 0.74 V of Pt/C, Figure S20) surpassing that of the catalyst free of P and Fe2P. As determined from the K–L plots (inset of Figure c), the electron transfer number (n) of CP-CMP-900 was found to be approximately 3.79, indicating that CP-CMP-900 favored a four-electron pathway for the acidic ORR. Consistent well with the results (Figure d) achieved from the RDE measurements, RRDE tests further validated a four-electron pathway for ORR in acidic media (3.92 at 0.5 V). In addition, the generated percentage of 2e product H2O2 is below 13.1% at the tested potential range of 0–0.7 V (Figure d). The long-term stability of ORR catalysts is also a major concern in fuel cell technology. So, the long-term stability of CP-CMP-900 is shown in Figure e; a continuous O2 reduction for 20 000 s on the CP-CMP-900-coated electrode resulted in only 33.9% loss of the value of current density. While Eonset and E1/2 of Pt/C decreased greatly under the same conditions (Figure S21a). The morphology and elements distribution of CP-CMP-900 after the i–t tests were visualized in Figures S23 and S24. The morphology was almost unchanged after the i–t test, and all elements are still evenly distributed over the carbon skeleton, indicative of the excellent stability of prepared catalysts. In addition, the role of Fe2P in catalyzing the ORR was further detected by leaching off Fe2P in the acid media.[38,39] As shown in Figure S23, the electrochemical activity of CP-CMP-900 toward ORR is higher than that after the acid treatment. The Eonset value of CP-CMP-900 decreased to 0.82 V, which clearly testifies the importance of Fe2P. As seen in Figure f, after the injection of methanol (t = 400 s), the cathodic current of commercial Pt/C decreased dramatically. However, as for CP-CMP-900, only a slight change could be detected. Besides, the loss of E1/2 for CP-CMP-900 is less than that of Pt/C (49 vs 72 mV, Figure S21b) after the injection of methanol. All of these convincingly testified that CP-CMP-900 preceded Pt/C for ORR in direct methanol fuel cells. The details are given in Table S4 in SI. To further prove the stability of prepared catalysts, the change of morphology and elemental distribution of CP-CMP-900 after the durability test in 0.1 M HClO4 was visualized by the SEM and corresponding element mapping. As shown in Figures S24 and S25, the morphology of CP-CMP-900 was almost unchanged after the i–t test. And all elements are still evenly distributed over the carbon skeleton.

Figure 8

Electrochemical performance of prepared catalysts in 0.1 M HClO4: (a) CV of CP-CMP-900 in 0.1 M HClO4 saturated with O2 at a sweep rate of 50 mV s–1; (b) LSV curves of CP-CMP-X and commercial Pt/C at 1600 rpm; (c) LSV curves of CP-CMP-900 at various rotation speeds from 400 to 2500 rpm in O2-saturated solution; (d) percentage of hydrogen peroxide yield and the electron transfer number (n) of CP-CMP-900 at different potentials; (e) long-time stability curves of CP-CMP-900 together with the commercial Pt/C at a constant voltage of 0.8 V (vs RHE) in O2-saturated 0.1 M HClO4 solution; and (f) methanol tolerance test of CP-CMP-900 and Pt/C in O2-saturated 0.1 M HClO4.

The excellent catalytic activity of CP-CMP-900 in various electrolytes could be attributed to the synergistic effect of the homogeneous distribution of Fe2P nanoparticles and FeNx species in the N,P-rich porous skeletons.[59−62] The doping of P and N could change the charge distribution of carbon materials, activating the O2 adsorption capacity.[63−66] Furthermore, the high porosity and hierarchical pore structure are good for mass transfer, improving the ORR ability. And the contents of pyridinic-N and graphitic-N, which have been proved favorable for the ORR by both theory and experimental results, are higher than those of the other N species.

Conclusions

In summary, we develop a facile and controllable method for the preparation of metal porphyrin-based POPs introducing C, N, P, and Fe simultaneously. Through direct carbonation of this special POP, a highly stable carbon-based catalyst (CP-CMP-900) embedded with well-defined Fe2P and FeNx species was obtained. A finishing point was achieved by doping of P and Fe2P into the pyridinic-N- and graphitic-N-dominated carbon. The synergistic effect of hierarchical pore structure, high porosity, and abundant Fe2P and FeNx species enable CP-CMP-900 present excellent ORR catalytic activity in the whole pH range. This work may provide a convenient, controllable, and efficient method for the design and preparation of more efficient carbon-based ORR catalysts for direct fuel cells and metal–air batteries.