Introduction

Nowadays, scarcity of fossil resources and deteriorating environment have become two increasingly serious problems. The development and use of sustainable clean energy have become critical. Among them, the fuel cell is a new energy technology that has received extensive attention because of its high efficiency, low pollution, high energy density (Zhu et al. 2015; Wu et al. 2018). Oxygen reduction reaction (ORR) of the fuel cell is a reduction reaction in which oxygen in the cathode of the fuel cell is reacted by a catalyst to obtain electrons. It is a very important electrochemical basic reaction, and it is the most important part of determining the fuel cell rate. Therefore, choosing a good ORR catalyst is a significant task in constructing fuel cell (Dresselhaus and Thomas 2001; Wang 2005; Qiu et al. 2011). Platinum is a good ORR electrocatalyst. The reaction process is a 4-electron transfer process that does not produce intermediate products. However, the low platinum reserves, the high cost, and the relatively short service life severely restrict the commercial production of fuel cells.

Therefore, to break through these bottlenecks, there are two main research directions in the search for alternative materials to take the place of expensive Pt-based catalysts (Yang et al. 2018). On the one hand, the metal-free heteroatom-doped carbon-based materials exhibit outstanding catalytic performance and favorable long-time stability (Li et al. 2013; Wang et al. 2009; Qu et al. 2010). However, these materials are active only under alkaline media, and their onset potentials are not excellent compared to those of platinum catalyst. Another focus of research is on non-precious metal-based catalysts (Lefevre et al. 2009; Jaouen et al. 2011; Mahmood et al. 2018). Although the onset potentials of these catalysts are almost the same of commercial Pt/C, the stability has always been a difficult problem.

In recent years, considerable attention has been poured into non-precious metal electrocatalysts which display remarkable performance in ORR, consisting of various transition metals species, heteroatom-doped carbon such as N, S, P which can change the electronic distribution of materials, thereby enhancing catalytic performance, and transition metal-nitrogen-doped carbons (M–N–C) (M=Fe, Co, Cu etc.) (Guo et al. 2018). Among them, transition metal-nitrogen-doped carbons are the most promising alternative electrocatalysts, and they have a similar mass activity to Pt/C and outstanding long-time stability. Although a large number of researches infer that the M–N–C plays an important role as an active site (Yang et al. 2018; Hu et al. 2017), the mechanism of ORR active sites in transition metal-nitrogen-doped carbon material has been still uncertain. So far, theoretical calculation and experimental verification have been carried out, and in order to synthesize promising nonprecious metal catalysts (NPMCs) with a Pt-like ORR performance, four critical factors need to be considered: firstly, the number and activity level of the active sites depends on the morphology and composition of the electrocatalyst; secondly, along with the increase of specific surface area and the formation of the pore structure, the degree of exposure of the active site during the oxygen reduction reaction process increases (Jaouen et al. 2011; Liang et al. 2013; Zhang et al. 2018a, b); besides, transition metal coordinated with N element serves as the active site; and finally, the doped heteroatoms (N, S, and B etc.) which are adjacent to the carbon sites change the distribution of electronic structures to enhance the ability to adsorb oxygen (Yang et al. 2017). Based on the above factors, the group of Prof. Baglio has successfully prepared carbon nanofibers modified with a combination of CoO and Co through a simple electrostatic spinning, showing the oxygen catalytic performance with a positive half-wave potential (0.87 V) for ORR (Alegre et al. 2017). Liang et al. have also synthesized a special structure of porous carbon nanofibers (HP-Fe-N/CNFs) via a facile strategy using electrospinning, exhibiting favorable long-time stability with about 87.8% current retention after 10,000 s for ORR (Zhao et al. 2017). Importantly, Fe–Co co-doped porous carbon nanotubes have not yet been reported. It is extremely meaningful via facile electrospinning to synthesize such nanostructured FeCo/N-C as the ORR electrocatalyst for a fuel cell or Zn-air batteries.

Herein, we have successfully synthesized Fe, Co, N co-doped carbon nanofibers (FeCo/N-C CNFs) via a facile method of high-temperature calcination pyrolysis of zinc, cobalt bimetallic zeolitic imidazolate framework (ZIF)-coated electrospun polyacrylonitrile fibers. Because of the addition of zinc acetate in the spinning solution, during the pyrolysis process, the self-etching process can be completed. Zn(Ac)2 and PAN can be respectively pyrolysis to ZnO and carbon at high temperature. After that, ZnO reacts with carbon to make ZnO deplete, thus achieving the purpose of etching carbon nanofibers (Chen et al. 2017). After pyrolysis, the bimetallic ZIF layer will be converted into N-doped porous carbon with high graphitization, and because of self-etching, the carbon nanofibers can form the interconnected hierarchically porous structures. Moreover, the high specific surface area and big pore volume provide electronic transmission channel as well as increase the mass transfer rate and make full use of active sites during the ORR process. Based on the above reasons, the FeCo/N-C CNFs catalysts show promising ORR catalytic performance in alkaline medium. In summary, this catalyst has the potential to be a commercial Pt/C alternative in practical applications.

Experimental section

Materials

Polyacrylonitrile (PAN) powder (average Mw 150,000) was purchased from Shanghai Titan Technology Co. Zinc acetate dehydrate (Zn(Ac)2·2H2O), cobalt(II) acetate tetrahydrate (Co(Ac)2·4H2O) and iron chloride hexahydrate (FeCl3·6H2O) were purchased from Sinopharm Chemical Reagent Co., Ltd. Dimethylformamide (DMF) was purchased from Shanghai Titan Technology Co., Ltd. 2-methylimidazole was purchased from Thermo Fisher Scientific (China) Co., Ltd.

Synthesis of FeCo/N-C CNFs

A total of 2.0 g of PAN powder was first dissolved into 10 mL of DMF with magnetically stirring for 12 h at room temperature (R.T.). Additionally, 1.9 g of Zn(Ac)2·2H2O, 0.1 g of Co(Ac)2·4H2O, and 0.2 g of FeCl3·6H2O were put into 10 mL DMF to be magnetically stirred together. By then, the obtained Co(Ac)2/Zn(Ac)2/FeCl3/DMF solution was slowly poured into the above PAN/DMF solution under stirring. After continuous mixing overnight, the fully homogeneous mixture solution of PAN/Co(Ac)2/Zn(Ac)2/FeCl3/DMF were obtained and subjected to electrospinning.

For electrospinning, the PAN/Co(Ac)2/Zn(Ac)2/FeCl3/DMF mixture solution was fed into a 10-mL plastic syringe with a 22-gauge stainless steel needle (0.7 * 32 mm). An electrospinning voltage of 15 kV was applied between the needle and the collector. The pushing flow rate of the needle was set as 0.02 mm/min. The rotation speed of the collector was 40 RPM, and the distance between the needle and the collector was set as 20 cm. Afterward, the PAN/Co(Ac)2/Zn(Ac)2/FeCl3 composite membranes obtained by electrospinning were directly immersed into 100 mL of 2-methylimidazole/ethanol solution (1 mol/L). After 12 h, PAN/Co(Ac)2/Zn(Ac)2/FeCl3 composite membrane was taken out and then washed several times with ethanol. Next, the composite membranes coated with a layer of bimetallic (Co2+ and Zn2+) ZIF were dried in an oven for 10 h at 80 °C. Then, for synthesis of FeCo/N-C CNFs, the dried composite membranes were first stabilized under an air atmosphere at 250 °C for 2 h with a heating rate of 1 °C min−1 and then carbonized at 800 °C for 20 h in a N2 flow with a heating rate of 5 °C min−1. In the next, the carbonized membranes were treated in hydrochloric acid (1 M) to remove the residual metal followed with water washing and drying. Finally, the FeCo/N-C CNFs were obtained. For comparison, Fe/N-C CNFs and Co/N-C CNFs were prepared by the same process. The metal contents for FeCo/N-C CNFs, Fe/N-C CNFs, and Co/N-C CNFs were the same by controlling the feed weight and molar ratios before electrospinning.

Instruments and characterization

The morphologies of the samples were observed by scanning electron microscopy (JEOL JSM-7800F) and transmission electron microscopy (JEOL JEM-2100F). The crystal structures were examined by X-ray diffraction (XRD; Bruker D8 advance). Raman spectroscopy was collected by using Lab-RAM HR800 (Horiba Jobin Yvon). Nitrogen adsorption-desorption isotherms and pore-size distribution date were collected using a Micrometrics ASAP2020 analyzer at 77 K. X-ray photoelectron spectroscopy (XPS) were tested with a Kratos Axis UltraDLD spectrometer, using a source gun of Al Kα as the excitation source.

Electrochemical measurements

All the electrochemical measurements were carried out in 0.1 M KOH solution at ambient temperature on a CHI660E electrochemistry workstation with a three-electrode system. A platinum foil was used as counter electrodes, and an Ag/AgCl electrode-saturated KCl was used as reference electrodes. A catalyst film was formed by dropping catalyst ink onto RDE as working electrode. The FeCo/N-C CNFs catalyst ink was prepared by ultrasonically dispersing 4.0 mg of the catalyst powder in a solution containing 800 μL isopropanol, 150 μL deionized water, and 50 μL 0.05wt% Nafion. After that, a certain amount of catalyst ink was added dropwise onto the RDE, giving a catalyst loading of 0.20 mg/cm2.

All potentials in this study were converted to potentials relative to a reversible hydrogen electrode (RHE) by the following equation:

$$ E\left(\mathrm{vs}\ \mathrm{RHE}\right)=E\left(\mathrm{vs}\ \mathrm{Ag}/\mathrm{AgCl}\right)+0.197+0.059\mathrm{pH} $$

The electron transfer number during the ORR was determined by the Koutechy-Levich equation:

$$ \frac{1}{J}=\frac{1}{J_{\mathrm{L}}}+\frac{1}{J_{\mathrm{K}}}=\frac{1}{B{\omega}^{1/2}}+\frac{1}{J_{\mathrm{K}}} $$
$$ \mathrm{B}=0.2\mathrm{n}F{C}_0{D}_0^{2/3}{\nu}^{-1/6} $$

where J, JL, and JK are the measured current density and the limiting and kinetic current densities, respectively; ω is the rotating speed, n is the electron transfer number, F is the Faraday constant (96,485 C mol−1), C0 is the bulk concentration of O2 (1.2 × 10−3 mol L−1), D0 is the diffusion coeffcient of O2 (1.9 × 10−5 cm2 s−1 for 0.1 M KOH solution), and ν is the kinematic viscosity of the electrolyte (0.01 cm2 s−1 for 0.1 M KOH solution).

Results and discussion

As shown in Fig. 1, the preparation process of FeCo/N-C CNFs can be divided into three steps. Firstly, using the Co(Ac)2, Zn(Ac)2, FeCl3, and PAN as start materials, the electrospun nanofiber membranes embedded with metal ions (Co2+, Zn2+, Fe3+) were prepared. Then, the PAN/Co(Ac)2/Zn(Ac)2/FeCl3 composite membranes obtained by electrospinning were directly immersed into 2-methylimidazole/ethanol solution. In this process, some metal ions (Co2+ and Zn2+) were continuously dissolved from the inside of the electrospun nanofiber and formed ZnCo-ZIF on the surfaces of nanofibers. Figure 2 a shows the morphologies of PAN/Co(Ac)2/Zn(Ac)2/FeCl3 composite fibers. The surfaces of the fibers are very clean and smooth. After immersing into the 2-methylimidazole/ethanol solution, the surface morphologies of the fibers obviously changed and became very rough (Fig. 2b). In the meantime, the mean diameter of the fiber obviously increased from ~ 352.7 to ~ 444.8 nm with an increase of 92.1 nm. The metal ions (Co2+ and Zn2+) dissolved out from the original PAN/Co(Ac)2/Zn(Ac)2/FeCl3 composite fibers and formed ZnCo-ZIF in the presence of the ligand of 2-methylimidazole. After that, the FeCo/N-C CNFs were prepared by carbonization of BM-ZIF@PAN nanofibers at 800 °C for 20 h with N2 flow. Observing from the SEM images of as-prepared FeCo/N-C CNFs (Fig. 2c), the morphology of the fiber with the mean diameter of ~ 280 nm is well maintained after carbonization. The shrink in the diameter is mainly due to the volatilization of non-carbon components in PAN during pyrolysis process. TEM images were used to further investigate the structures of FeCo/N-C CNFs. From Fig. 2d–f, it can be clearly seen that the surfaces of FeCo/N-C CNFs exhibit porous fibrous structure derived from BM-ZIF after carbonization with a mean pore size of ~ 52.75 nm. In fact, the introduction of Zn(Ac)2 can facilitate pore generation due to its self-etching in the process of carbonization. Meantime, the ZnCo-ZIFs covered on the surfaces of fibers turned into the porous carbon frameworks wrapped on the surfaces of nanofibers. Additionally, it is worth noting that there is an individual carbon tube in Fig. 2g growing on the surface of CNFs. It is mainly due to that Fe can be used as catalyst for the formation of carbon nanotubes (Hou et al. 2003; Lee et al. 2002; An et al. 2018). From the HRTEM image, the Fe nanoparticle was encapsulated in the carbon shell with the clear lattice fringe spacing of 0.202 nm corresponding to the (110) planes of Fe (Fig. 2h). The lattice spacing for the outer carbon shell is about 0.343 nm, which is close to that of graphite, 0.34 nm. The Fe NPs wrapped in PAN can catalytically graphitize the carbon precursor to form graphite phase during carbonization process. It is also demonstrated that the Fe and Co doped into the structures of carbon shell did not form nanoparticles anchored on the surfaces of the nanofiber. From the EDS mapping images (Fig. 2i), the elements of nitrogen, iron, and cobalt are well distributed in the FeCo/N-C CNFs. From EDX analysis, the Co and Fe account for about 15.0 wt% and 16.8 wt% of the FeCo/N-C CNFs, respectively.

Fig. 1
figure 1

Schematic diagram for the preparation process of FeCo/N-C CNFs

Full size image
Fig. 2
figure 2

a SEM image of PAN/Co(Ac)2/Zn(Ac)2/FeCl3 nanofibers. b SEM image of BM-ZIF@PAN nanofibers, c SEM image of FeCo/N-C CNFs. d–g TEM images of FeCo/N-C CNFs. f High-resolution TEM image of FeCo/N-C CNFs. g HAADF STEM images and elemental mapping of FeCo/N-C CNFs

Full size image

From Fig. 3a, the diffraction peak at 2θ = 24.6° which is indexed to C (002) crystal plane can be observed in XRD spectra of FeCo/N-C CNFs, Fe/N-C CNFs, and Co/N-C CNFs. The broad peaks at 2θ = 43.8° are originated from the C(100) crystal plane. For FeCo/N-C CNFs, there is a sharp peak located at 2θ = 44.5° and a tiny peak at 2θ = 65.2 which is ascribed to the Fe (110), Fe (200) crystal plane diffraction peak (JCPDS 87-0722). The crystallite size was determined by using the Debye-Scherrer formula given by Dhkl = kλ/βcosθ, where Dhkl is the crystallite size estimated from the (hkl) line, k is the Scherrer constant, β is the half-width, λ is the X-ray wavelength, and θ is the diffraction angle (Guruvammal et al. 2016). The crystallite size for Fe NPs is ~ 11.135 nm, which is consistent with high-resolution TEM image. Furthermore, Fig. 3b gives the Raman spectra of all the samples to evaluate the degree of graphitization of these catalysts. All spectra of samples show two peaks at about 1350 and 1579 cm−1, which can be attributed to D band representing structural defects and G band representing graphite, respectively (Yang et al. 2015; Shanmugam and Osaka 2011; Lu et al. 2017). The intensity ratios of these two bands (ID/IG) of FeCo/N-C CNFs, Fe/N-C CNFs, and Co/N-C CNFs are near ~ 2.1, which reflects the high degree of disorder and defects arise from the dopant of Fe, Co, and N elements. The nitrogen adsorption and desorption isotherm and pore size distribution of FeCo/N-C CNFs are shown in Fig. 3c, d. According to the original IUPAC classification, FeCo/N-C CNFs exhibit type IV isotherms, and its hysteresis loop is clear H4 type hysteresis loops. The total pore volume of FeCo/N-C CNFs is about 115.5 cm3/g STP estimated from the amount adsorbed at P/P0 = 0.99. After converting, the pore volume is about 0.178 cm3/g STP. Meanwhile, the specific surface area is calculated to be 370.8 m2 g−1. Also, the FeCo/N-C CNFs catalyst has a microspore structure with size distribution concentrate on ~ 1 nm. Therefore, FeCo/N-C CNFs had the rational micropore structure with a large specific surface area, which was beneficial to the mass transport of ORR relevant species and more active sites exposed.

Fig. 3
figure 3

a XRD patterns and b Raman spectra of FeCo/N-C CNFs, Fe/N-C CNFs, and Co/N-C CNFs. c Nitrogen adsorption-desorption isotherms of FeCo/N-C CNFs. d Corresponding pore-size distribution curve of FeCo/N-C CNFs

Full size image

Moreover, the elemental chemical composition and binding energy of the FeCo/N-C CNFs were provided by the X-ray photoelectron spectroscopy (XPS). As shown in Fig. 4a, the spectrum proved the coexistence of C, N, O, Fe, and Co elements in FeCo/N-C CNFs. As exhibited in Fig. 4b, the high-resolution Co 2p XPS spectra exhibited three prominent bands at binding energies of 779.4/794.6, 781.6/796.6, and 786.7/802.9 eV; these peaks can prove the existence of metallic Co, CoCxNy, and Co-Ny, Co-O, respectively (Deng et al. 2013; Jiang et al. 2013; Zhang et al. 2018a, b). The peak binding energies at 781.6/796.6 eV which had two extra peaks located at 786.7/802.9 eV called satellite peak reveal the oxidized Co species (Zhang et al. 2016). Meanwhile, the high-resolution Fe 2p XPS spectra shown in Fig. 4c were deconvoluted into three distinct peaks. Zero-valent Fe (Fe0) corresponded to the peak located at 706.6 eV, which confirmed the presence of metallic Fe. The other two peaks at 709.7 and 713.4 eV could be ascribed to the 2p3/2 orbitals of Fe2+ and Fe3+ species, respectively. Moreover, the shake-up satellite of 718.0 eV also confirmed the presence of Fe3+ species (An et al. 2018; Lin et al. 2014; Guan et al. 2018; Peng et al. 2013; Singh et al. 2015; Su et al. 2017; Yamashita and Hayes 2008). Furthermore, the high-resolution N 1s XPS spectra (Fig. 4d) displays four types of peaks: pyridinic-N (398.4 eV), pyrrolic-N (399.7 eV), graphitic-N (400.9 eV), and oxidized-N (403.4 eV) (Lin et al. 2014; Liu et al. 2010; Chen et al. 2015). Also, pyridinic-N bonded to two carbon atoms at the edge of the graphite plane can provide a lone pair of electrons and therefore can absorb oxygen and intermediate molecules. And graphitic-N bonded to three carbon atoms in the graphite plane is conducive to raising the limiting current density. So, these two types of N have great influence on ORR activity. Besides, owing to pyridinic-N having electron-donating properties, Fe and Co atoms can be more easily anchored on the material (Lin et al. 2014).

Fig. 4
figure 4

a XPS survey spectrum of FeCo/N-C CNFs. b High-resolution of Co 2p, c Fe 2p, and d N 1s XPS spectrum

Full size image

Subsequently, the ORR performances of all the samples were accurately investigated by detailed electrochemical tests. Cyclic voltammetry (CV) curves in N2- and O2-saturated 0.1 M KOH, respectively, of all the catalysts, were first used to investigate the ORR performance of all catalysts. As shown in Fig. 5a, it was self-evident that there were oxygen reduction current peaks for all the samples appearing in the O2-saturated electrolyte, whereas in the N2-saturated electrolyte, there was no obvious oxygen reduction current peak. The ORR peak potential of FeCo/N-C CNFs is ~ 0.83 V which is close to the Pt/C (0.90 V) and more positive than that for Fe/N-C CNFs (0.76 V) and Co/N-C CNFs (0.71 V), indicating that FeCo/N-C CNFs shows the good ORR performance. Meantime, from the CV curves obtained in N2-saturated 0.1 M KOH, the double-layer capacitance of FeCo/N-C CNF catalysts was much higher than of other contrast samples, suggesting FeCo/N-C CNF catalysts had a larger solid-electrolyte interface due to their higher BET surface areas. Electrochemical impedance spectroscopy (EIS) measurements were conducted to investigate the electroconductivity of FeCo/N-C CNFs, Fe/N-C CNFs, and Co/N-C CNFs. As shown in Figure S2 and Table S1, it can been seen that CoFe/N-C CNFs exhibit the lower charge-transfer resistance (Rct = 27.52 Ω) than Co/N-C CNFs (Rct = 37.71 Ω), Fe/N-C CNFs(Rct = 33.90 Ω), indicating that CoFe/N-C CNFs had the highest electron conductive ability among them (Gao et al. 2017 and He et al. 2017). Further, linear sweep voltammetry (LSV) is used to investigate the ORR performance of the catalysts. As revealed in Fig. 5b, FeCo/N-C CNFs display a satisfactory Pt/C like ORR activity. Although the half-wave potential (E1/2, 0.834 V) and onset potential (Eonset, 0.990 V) of FeCo/N-C CNFs are more negative than Pt/C (0.863 V, 1.034 V), the limiting current density (4.33 mA cm−2) is comparable to commercial 20% Pt/C (4.20 mA cm−2), moreover, which outperforms that of the other two contrast catalysts.

Fig. 5
figure 5

a CV curves in N2- or O2-saturated 0.1 M KOH solution of Pt/C, FeCo/N-C CNF, Fe/N-C CNF, and Co/N-C CNF catalysts (scan rate 50 mV s−1). b ORR LSV curves of Pt/C, FeCo/N-C CNFs, Fe/N-C CNFs, and Co/N-C CNFs catalysts. c LSV curves of FeCo/N-C CNFs at different rotation speeds (inset figure is the K-L plots of FeCo/N-C CNFs with different electrode potentials). d Tafel plots of Pt/C, FeCo/N-C CNF, Fe/N-C CNF, and Co/N-C CNF catalysts

Full size image

To investigate the kinetics mechanism of the ORR during the reaction, we recorded the ORR polarization curves of FeCo/N-C CNFs (Fig. 5c) and commercial Pt/C at various rotation speeds (400−2025 rpm). Higher rotational speed reduces the diffusion distance and further increase the current density. According to the Koutecky-Levich (K-L) equation, we can calculate the electron transfer number (n). The corresponding K–L plots of FeCo/N-C CNFs (inset of Fig. 5c) and commercial Pt/C exhibit good linear correlation indicating the first-order reaction kinetics toward the dissolved oxygen concentration (Liang et al. 2011). From the inset K–L plot in Fig. 5c, the slope of the K–L plot is about ~ 3.57 ranging from + 0.3 to + 0.6 V which is close to an ideal four-electron pathway (n = 4), meaning that the FeCo/N-C CNFs can catalyze a 4-electron oxygen reduction reaction (Xu et al. 2017 and Chai et al. 2019). Determined from electrochemical double-layer capacitance (Cdl) (Nai et al. 2018 and Deng et al. 2016), the electrochemically active surface areas (ECSAs) of FeCo/N-C CNFs is about 145 cm−2 (Supporting information). Moreover, the corresponding Tafel plots are shown in Fig. 5d the FeCo/N-C CNFs were observed a Tafel plot slope of 80 mV dec−1, which was smaller than that of commercial Pt/C (92 mV dec−1), Fe/N-C CNFs (106 mV dec−1), and Co/N-C CNFs (167 mV dec−1). Based on the above observations, the FeCo/N-C CNFs had a more superior electron transfer rate and kinetic reaction than the other three catalyst samples.

Besides the ORR activity of catalysts, we also tested its methanol tolerance and long-term stability. The methanol tolerance test was investigated by chronoamperometry with injection of a certain amount of methanol into the electrolyte at a specific time. The long-term stability performance was also investigated by chronoamperometric measurements (Fig. 6a). After the electrochemical stability test for 10,000 s at 1600 rpm, about 87.5% of the initial current density can be maintained, which is a lot better than that for Pt/C (59.4%), suggesting the excellent stability of FeCo/N-C CNFs. From Fig. 6b, after injecting methanol, we can see there was no apparent fluctuation of current density for FeCo/N-C CNFs, while an obvious current density reduction for the Pt/C could be observed, suggesting the better tolerance performance of FeCo/N-C CNFs to crossover effect.

Fig. 6
figure 6

a Electrochemical stability test. b Methanol tolerance test by current-time curves of FeCo/N-C CNFs and 20% Pt/C at 0.65 V in O2-saturated 0.1 M KOH at 1600 rpm

Full size image

Conclusion

In general, the porous FeCo/N-C CNF electrocatalysts were successfully fabricated by using electrospinning, immersion, and simple pyrolysis. In the process of immersion, the ZnCo-ZIF can be generated on the surfaces of the electrospun composite fibers, facilitating the formation of interconnected porous structure in pyrolysis. The porous FeCo/N-C CNFs catalysts exhibited high ORR performance. Remarkably, in alkaline media, the FeCo/N-C CNFs exhibited a satisfactory Pt/C like ORR activity. It is also proved that FeCo/N-C CNFs displayed not only more outstanding long-time durability but also methanol crossover effect tolerance than those of commercial Pt/C in alkaline solution. The high ORR catalytic performances of FeCo/N-C CNF catalysts are mainly derived from the following important aspects: (i) the doping Fe, Co, and N led to inhomogeneous charge distribution which facilitates the capture of oxygen species by the catalysts during the reaction (Tong et al. 2017; Wang et al. 2017; Gong et al. 2009); (ii) the porous structure gives a large specific surface area, derived from etching by Zn2+ in the fiber, improving the stability of the reaction and accelerates mass transport of reactants; and (iii) the doping of Fe-Nx and Co-Nx enhances the electron transfer ability and acts as an ORR active site (Tan et al. 2018; Wu et al. 2012; Zhou et al. 2017; Meng et al. 2016). The outcomes of this study may bring some new synthesis strategy for the preparation of non-precious carbon-based electrocatalyst and have a high potential for fuel cells.