Nitrogen-coordinated single-cobalt-atom electrocatalysts, particularly ones derived from high-temperature pyrolysis of cobalt-based zeolitic imidazolate frameworks (ZIFs), have emerged as a new frontier in the design of oxygen reduction cathodes in proton exchange membrane fuel cells (PEMFCs) due to their enhanced durability and smaller Fenton effects related to the degradation of membranes and ionomers compared with emphasized iron-based electrocatalysts. However, pyrolysis techniques lead to obscure active-site configurations, undesirably defined porosity and morphology, and fewer exposed active sites. Herein, a highly stable cross-linked nanofiber electrode was directly prepared by electrospinning using a liquid processability cobalt-based covalent organic polymer (Co-COP) obtained via pyrolysis-free strategy. The resultant fibers can be facilely organized into a free-standing large-area film with a uniform hierarchical porous texture and a full dispersion of atomic Co active sites on the catalyst surface. Focused ion beam-field emission scanning electron microscopy and computational fluid dynamics experiments confirmed that the relative diffusion coefficient was enhanced by 3.5 times, which can provide an efficient route both for reactants to enter the active sites, and drain away the produced water efficiently. As a result, the peak power density of the integrated Co-COP nanofiber electrode was remarkably enhanced by 1.72 times along with significantly higher durability compared with conventional spraying methods. Notably, this nanofabrication technique also maintains excellent scalability and uniformity, which are desired characteristics for assisting in membrane electrode assembly manufacturing for PEMFCs.
Article
Porous Covalent Organic Polymer Coordinated Single Co Site Nanofibers for Efficient Oxygen Reduction in Acids
https://doi.org/10.21203/rs.3.rs-1876198/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
The development of platinum-group-metal (PGM) free catalysts as acidic oxygen reduction reaction (ORR) cathodes in proton exchange membrane fuel cells (PEMFCs) is critical, as it is an important pathway towards realizing carbon neutralization1,2. Transition metal M-N-C (Fe, Co, Mn, etc.) catalysts have attracted much attention due to their special M-Nx active sites and low reaction energy barriers during ORR processes3–6. Compared with emphasized single-iron-based electrocatalysts, nitrogen-coordinated single-cobalt site catalysts exhibit more promising prospects owing to fewer concerns about the degradation of both the ionomer and the membrane by the Fenton reaction7,8. Atomically dispersed Co-N-C catalysts are usually synthesized from cobalt-based zeolitic imidazolate frameworks (ZIFs) through high-temperature pyrolysis, which presents challenges in maintaining the desired structure and synthetic control9–11. Moreover, the pyrolysis process leads to morphological destruction, causing many effective active sites to be buried due to undesired pore structures, thereby significantly reducing the catalyst activity.
Covalent organic polymers (COPs) have emerged as a promising electrocatalyst for ORR due to the atomic metrical control of the organic molecular components along with their robust architectural possibilities and chemical properties12–14. Additionally, the structure of COP materials obtained by irreversible kinetics is relatively stable, especially in alkaline systems, which is more conducive to the long-term operation of electrocatalytic ORR15,16. For example, Yang et. al.17 developed a simple yet effective pyrolysis-free approach to craft single-atom cobalt catalysts with high electrocatalytic ORR activity via judiciously wrapping an active porphyrin-based thiophene-sulfur site-containing PTS-COP shell around a highly conductive MWCNT core in-situ. Our group18 successfully developed a microwave-assisted synthetic method to produce a liquid processability Fe-based COPBTC catalyst with nitrogen-coordinated single-metal sites through an in-situ charged exfoliation approach. The obtained COP solution itself can be directly employed as a highly efficient Pt-replaced catalyst for zinc-air flow batteries, generating promising performance and outstanding stability. However, there are few studies on the performance of COP-based catalysts in acidic electrolytes, particularly in PEMFC devices, which still remains a great challenge.
Herein, a highly stable cross-linked nanofiber electrode was prepared based on a liquid processability cobalt-covalent organic polymer (Co-COP) by electrospinning to build an excellent active center structure. The hierarchical porous structure acquired by this nanofabrication method fully exposes Co-N4 sites, accelerates the transfer efficiency of reaction gases and products, and reduces the charge transfer resistance. Focused ion beam-field emission scanning electron microscopy (FIB-FESEM) was used to slice the samples before the real structures of the membrane electrodes were obtained by three-dimensional reconstruction technology. Computational fluid dynamics (CFD) results show that the cross-linked nanofiber electrode has high relative diffusion coefficient relative to the conventional spraying electrode (46.9% vs 13.4%), which provides an effective channel for diffusing reactants to the active center and removing the produced water efficiently. The results show that the power density of the cross-linked nanofiber electrode is significantly increased by 1.72 times compared with the conventional spraying electrode. Importantly, the cross-linked nanofiber electrode maintained 85% of the original voltage after a 150 hour durability test, compared to the conventional spraying electrode whose voltage dropped to 53% after 50 hours. In this regard, the cross-linked nanofiber electrode compares favorably with the hot-topic variety of monodispersed atomic metals embedded in nitrogen-doped carbon catalysts, which typically show ~60% of the original voltage after 100 hours, attesting to its significantly enhanced stability19–21.
Moreover, the performance and cycle life of the PEMFCs stack depends on the membrane electrode in practical application processes, as the failure of a single fuel cell leads to the failure of the whole fuel cell stack22. Therefore, the preparation of highly consistent membrane electrodes with the same physical and chemical parameters is the key to assembling fuel cell stacks with both high performance and long life. Although the catalytic activity of pyrolysis-free Co-COP in PEMFCs still needs improvement, this nanofabrication technique has good scalability and the performance of the obtained 200 cm2 film could maintain good consistency at a random position on the film, providing excellent processability for future commercialization. Benefiting from the advantages of Co-COP materials and nanofabrication technology, the three-phase reaction interface of the nanomaterial membrane electrode was designed and assembled, which allowed the atomic economy of the catalyst to be truly realized at the device level.
Preparation of free-standing hierarchically porous covalent organic polymer coordinated single Co site nanofiber electrode
A liquid processability Co-COP material with a Co-N4-C structure was synthesized by a pyrolysis-free method before being used in the Co-COP/AFGC catalyst, which was prepared by combining the Co-COP and acid functionalized graphite carbon (AFGC). Then, a highly stable cross-linked nanofiber electrode was prepared by electrospinning the ink prepared from Co-CoP/AFGC catalyst, ionomer, and polymer (polyacrylic acid (PAA) and polyvinyl alcohol (PVA)), followed by heating, and then subsequent esterification (The -COOH of PAA and -OH of PVA form an ester bond through condensation reaction) (Fig. 1a). Significantly, the cross-linked nanofibers have excellent scalability, rollability, bendability, and foldability, so they can be directly used as the free-standing cathode catalytic layer of PEMFCs (Fig. 1b). Instead of generating single-atom Co-N-C materials with a random distribution of Co on the carbon matrix by conventional pyrolysis methods, we obtained atomically dispersed Co-N-C catalysts through the interaction between COPs rich in Co-N-C and AFGC molecules.
Compared with XC-72, AFGC showed obvious highly graphitized lattice fringes in TEM images, which can improve the electrical conductivity and stability, and endows the catalyst with more rapid diffusion to assist the oxygen reduction reaction (Supplementary Fig. 1)23. The fluorescence intensity of the Co-COP/AFGC composites was significantly decreased compared with Co-COP (Supplementary Fig. 7). When the amount of AFGC was increased, the fluorescence intensity was further reduced, and even fluorescence quenching occurs, which can be ascribed to light-induced electron or energy transfer to the AFGC framework that could accumulate charges because of its conjugated network structure. When the Co-COP and AFGC were combined, the catalyst resistance was very close to the original AFGC (Fig. 2a).
The contact angle of the PAA nanofibers membrane electrode was significantly improved compared with the conventional spraying method, which was related to the polymer PAA contained in ink prepared by spinning (Fig. 2b). The larger contact angle indicates that the cross-linked nanofibers membrane electrode has stronger hydrophobicity than the other two kinds of membrane electrodes. This may be due to the conversion of hydrophilic groups in the electrode (such as -COOH in PAA and -OH in PVA) into ester functional groups. To further confirm that the esterification reaction occurred after the heat treatment, the structure of the PAA nanofibers and cross-linked nanofibers was characterized by using FTIR spectroscopy (Fig. 2c). According to the molecular formulas of PAA and PVA, both have -OH groups and they appear at about 3282 cm-1. After the heat treatment, the characteristic peak of C=O stretching vibration shifted right from the original 1719 cm-1 to 1710 cm-1, which indicated that an esterification reaction occurred. The peaks at 1139 cm-1 and 1063 cm-1 also confirmed the formation of the ester bond24. Interestingly, AFGC has a large adsorption capacity at any pressure during the BET analysis via N2 adsorption, and their corresponding specific surface areas were reduced from 58 to 19 m2 g-1, indicating that the porosity of AFGC was reduced after the combination of AFGC and Co-COP (Fig. 2d). Additionally, compared with Co-COP@AFGC, the specific surface area of the cross-linked nanofibers increased significantly (19 m2 g-1 vs 676 m2 g-1). The adsorption capacity increased significantly at low pressure, which showed that the number of micropores and mesopores increased. Further, the adsorption capacity at high pressure was greater than that of Co-COP, suggesting that there were many macroporous structures as well25. It can be seen from the pore size distribution that the cross-linked nanofibers were rich in microporous, mesoporous, macroporous, and super-porous structures, which directly improves the performance of PEMFCs (Fig. 2e).
As shown in Supplementary Fig. 9, the catalyst was uniformly distributed on a Nafion membrane, but some of the particles were agglomerated. The particle size of the catalyst was about 100 nm, and many of these particles were aggregated under the action of van der Waals forces to form a large cluster. During the ORR in PEMFCs, gas diffusion in the bulk catalytic layer will produce an additional bending pathway, resulting in mass transfer resistance within each cluster. Supplementary Fig. 9c, d show SEM images of the PAA nanofibers obtained by direct electrospinning of a catalyst ink containing only PAA polymer. The morphology of this sample was the same as the cross-linked nanofiber’s network structure, which is to say there were many large voids all over the fibrous film along with some nodes on the fibers (Fig.. 2f, g). When these nodes were enlarged, it was found that they were also composed of many nanofibers, which contain sufficient micropores, mesopores, and macropores. Additionally, larger agglomerated particles were not visible in the nanofibers, indicating that the shearing and mixing events during the electrospinning process increased the fracture of large catalyst agglomerated particles. During electrospinning, the larger catalyst particles were sheared under the action of the external electric field, and the aggregated structure of the fibers was formed under the combined action of solvent evaporation and electrostatic crushing. Recent studies have shown that the smaller the aggregation of the catalyst on the conventional spraying membrane electrode, the smaller the pores contained in the electrode structure, thus lowering the effective diffusion coefficient26. However, the key aspect here is that although electrospinning exhibits smaller catalyst aggregation, this integrated approach allows for large continuous microporous networks in the entire electrode structure, which was absent with conventional sprayed membrane electrodes. The unique network structure facilitates mass transfer between interconnected nanofibers through microporous voids. The TEM image of single cross-linked nanofibers was obtained, and a nanofiber diameter of about 300 nm was observed, along with a certain degree of porosity in the fibers. Moreover, each nanofiber contains homogeneous and densely dispersed catalytic sites (Fig. 2h). Co single atoms were observed by atomic-resolution high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) (Fig. 2i). Element mapping images were obtained to evaluate the fiber's ionomer distribution, which was prepared by electrospinning with Co-COP/AFGC catalyst: PAA: Nafion (50:15:30 wt.%). The elements F and S were completely distributed on the nanofiber electrode, indicating a uniform and continuous ionomer phase (Fig. 2j).
The high-resolution N 1s spectrum of Co-COP/AFGC can be divided into three peaks at 398.8, 399.5, and 400.5 eV, representing pyridinic N, metal-N bonding, and pyrrolic N, respectively (Supplementary Fig. 10c). The pyridinic N doping can reduce the adsorption energy barrier of O2 on adjacent carbon atoms and accelerate the rate-limiting first electron transfer process in ORR. For the Co 2p fine spectrum, the two peaks at 779.8 eV and 794.8 eV belong to the characteristic peaks 2p3/2 and 2p1/2 of Co3+, and the two peaks at 782.5 eV and 996.5 eV are associated with the characteristic peaks 2p3/2 and 2p1/2 of Co2+ (Supplementary Fig. 10d)27. In Co-COP/AF-GC, the nuclear energy level spectra of Co 2p3/2 and Co 2p1/2 shifted to 709.8 eV and 797.2 eV, indicating that the electron cloud density on Co was changed by the combination of Co-COP and AF-GC (Supplementary Fig. 10f).
The coordination of metals and nitrogen is critical for ORR performance. X-ray absorption near edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) was used to explore the coordination environment of Fe and N. The XANES at the Co K-edge of Co-COP/AF-GC clearly shows that the black line peak position was located between Co foil and CoO (marked in the green area in Fig. 3a), indicating that the average valence state of Co single atoms was higher than that of metal Co (0) and lower than that of oxidized Co (ІІ). As shown in Fig. 3b, the Co-COP/AF-GC EXAFS curve has a prominent peak at 1.46 Å corresponding to Co-N, which was significantly different from the main characteristic peak of Co-Co coordination (2.18 Å, 2.67 Å) of Co foil and CoO, indicating that Co atoms were atomically dispersed. The matching of calculated and measured spectra with Co-N two-body backscattering χ2 signals in K space suggests that the primary coordination environment of cobalt was Co-N (Supplementary Fig. 11). To quantitatively characterize the structural parameters of Co in Co-COP/AF-GC, least-squares EXAFS fitting was carried out by using a tetrahedral geometric structure model. The fitting curve was shown in Fig. 3c, and the detailed fitting parameters were listed in Table S3, further revealing that the coordination number of Co in Co-COP/AF-GC was about 4. Additionally, the bond length was about 1.9 Å, corresponding to the Co-N4-C structure. To visualize the dispersion of Co atoms in Co-COP/AF-GC, the wavelet transform (WT) of Co K-edge EXAFS analysis was further shown in Fig. 3d. The maximum value near 7.2 Å-1 was related to the Co-O interaction in CoO and Co3O4, and the maximum value around 8.1 Å-1 corresponds to the Co-Co bond in the Co foil. However, Co-COP/AF-GC has an obvious maximum tensile value near 5.0 Å-1, which was different from Co-O and Co-Co. From the above analysis, it can be seen that the cobalt atoms in Co-COP/AF-GC were atomically dispersed and coordinated with four N atoms.
Hierarchically porous nanofiber electrode for PEMFCs
The ORR performance of Co-COP/AF-GC catalyst and cross-linked nanofibers in an O2-saturated 0.1 M HClO4 electrolyte solution was firstly studied under a three-electrode system. As shown in Supplementary Fig. 12, the linear sweep voltammetry (LSV) curves of Co-COP/AF-GC and cross-linked nanofibers catalysts were nearly coincident. After careful analysis, the half-wave potential difference between the Co-COP/AF-GC catalyst and the cross-linked nanofibers catalyst was only 10 mV, and the kinetic current density (Jk) at 0.8 V was also not significantly different, which indicates that electrospinning technology does not affect the performance of the catalyst on the three electrodes.
Then, we explored the structure and PEMFC performance of nanofibers electrodes as a function of ionomer content and relative humidity. Understanding and optimizing the distribution of ionomers was especially important for improving the performance of the non-precious metal catalyst supported PEMFCs, as changes in the configuration and distribution of ionomers within the catalyst layer affect the proton transport. However, to achieve high fuel cell performance, it was necessary to combine the developed electrode structure with the electron conduction network of carbon-based electrocatalysts, the proton transport of ionized membranes, and efficient reactant gas supply to the electrode pores28,29. Therefore, measuring the performance of these nanofibers electrodes at the device level was very important to further understand their intrinsic activity, and to establish the dynamics and relationships between the schema, structure, and performance on the coverage of the ionomer. Supplementary Fig. 14c shows the polarization and power density curves of different ionomer contents at a relative humidity (RH) of 100%. The performance of PEMFCs was significantly improved when the ionomer content was increased from 25% to 30%, and the peak power density was increased from 195 mW cm-2 to 305 mW cm-2. Notably, as the content of the ionomer continues to increase, the power density decreases. It can be seen from the SEM images that the catalytic layer was composed of many fibers when the content of the ionomer was relatively small (Fig. 4b and Supplementary Fig. 15). As the content reached 30%, there were many nodes on the nanofibers, which also contained abundant micropores, mesopores, and macropores. At a 35% ionomer content, the resulting framework between the fibers of the catalytic layer was stacked, and many of the catalyst particles were agglomerated together. Additionally, under the condition of 75% RH the performance variation trend of PEMFCs with different ionomer content was consistent with that of 100% RH (Supplementary Fig. 14b). More importantly, the peak power density of PEMFCs at 100% RH was higher than that at 75% RH, which indicates that the water content of the electrolyte membrane affects the conductivity of the membrane during the fuel cell operation and determines the output performance of the cell to greatly. The corresponding Tafel slopes of different ionomer contents were calculated, and the 30% ionomer content catalyst showed the lowest Tafel slope, suggesting that it possesses the fastest ORR kinetics. Conversely, the 25% ionomer content catalyst demonstrated a much higher Tafel slope (280 mV dec-1) at low currents, suggesting slower kinetics. Additionally, the Tafel slopes also prove that different RH has a major influence on the performance of PEMFCs. (Supplementary Fig. 14d, e).
Fig. 4c depicts the relationship between the RH and the capacitance of both the conventional spraying and cross-linked nanofiber electrodes. The normalized capacitance was defined as the capacitance at a certain RH divided by the capacitance at 100% RH (the maximum capacitance). The specific formula was as follows:
where v represents the scanning rate, 
and 
are the current density in the positive and negative scanning direction at 0.4 V, respectively. The normalized capacitance of conventional spraying membrane electrodes depends on RH, which is because the micropores and mesopores of electrocatalyst and electrode get filled with liquid water for improved interactions between ionomer and catalysts. In contrast, the normalized capacitance of the cross-linked nanofiber electrode was largely unrelated to RH (in the range of 50 to 100% RH), which means that the ionomer was well distributed on the electrocatalyst.
Next, we further explored the performance of conventional spraying and cross-linked nanofiber electrodes by using a 5 cm2 hydrogen-oxygen MEA fuel cell. Fig. 4a shows the conceptual structure of the cross-linked nanofiber and conventional spraying electrodes to better understand the configuration of these two electrodes. The typical configuration of traditional catalysts can be described as Co-COP@AFGC encapsulated within Nafion ionomer; however, this orientation hinders the diffusion of gas molecules to the active sites, greatly decreasing the catalyst utilization30,31. Compared with the electrode structure of the traditional configuration, the cross-linked nanofiber catalytic layer has an ultrahigh catalyst utilization because both the interior and exterior surfaces are available to the reactants. Fig. 4d shows the polarization and power density curves. The peak power density of the cross-linked nanofiber electrode reached 305 mW cm-2, while that of the conventional spraying electrode only achieved 178 mW cm-2. Under completely humid conditions (100% RH), the performance of the cross-linked nanofiber membrane electrode was dramatically improved compared to the conventional spraying membrane electrode. To the best of our knowledge, this is the highest Pmax to be reported in H2-O2 PEMFCs so far with the pyrolysis-free macrocyclic compound as cathode catalysts or proton exchange membrane (Fig. 4e and Table S5). The resulting catalyst was subsequently evaluated under actual H2-air conditions, and the peak power density of the cross-linked nanofiber electrode reached 305 mW cm-2 (Supplementary Fig. 18a). The higher performance index of the membrane electrode can potentially be attributed to enhanced porosity provided by electrospinning, making the active sites of the electrochemical reaction be more fully exposed. Using BTC and CoCl2 as raw materials, the Co-COP/AFGC electrocatalyst synthesized by the combination of Co-COP and acid-functionalized XC-72 showed excellent power density. Further, the power density of the cross-linked nanofiber electrode obtained by electrospinning was greatly improved. Compared with the traditional high-temperature carbonization technology, the catalyst based on this pyrolysis-free strategy not only has higher repeatability in the preparation process, but also can achieve a high density of accessible active centers within defined structures.
The current research on PEMFCs catalysts is mainly focused on the non-precious metal-based M-N-C materials obtained by high-temperature calcination. In this work, we have developed a unique pyrolysis-free strategy for the synthesis of a liquid processability Co-COP/AFGC catalyst with a defined structure. The Co-COP/AFGC catalyst and two kinds of polymers (PAA and PVP) were electrospun together to obtain the nanofiber network film with high specific surface area and abundant porosity. Then, the nanofiber film was heated to esterify the carboxyl and hydroxyl groups in the two polymers to obtain the final cross-linked nanofiber electrode, which can greatly improve its chemical stability. Supplementary Fig. 18b shows the voltage and current density curves, and the Tafel slopes of the conventional spraying and cross-linked nanofibers membrane electrodes were 203 and 149 mV dec-1, respectively. In the high potential region, the polarization curve was mainly kinetic overpotential, which is closely related to the activity of the catalyst. Therefore, the cross-linked nanofiber membrane has higher ORR kinetic activity than conventional spraying catalysts.
The performance of conventional spraying and cross-linked nanofiber electrodes was further evaluated by an electrochemical impedance spectroscopy (EIS) test under the frequency range of 105 to 0.1 Hz at a current density of 0.5 A cm-2 (Fig. 4f). The two semicircles of the cross-linked nanofibers membrane electrode were both shrunk compared with the conventional spraying membrane electrode, indicating faster reaction kinetics and better mass transfer for the former32,33. The analyzer software from Scribner Company was used to fit the EIS data to quantify the impedance of each part of the fuel cell and draw the equivalent circuit model diagram. The charge transfers resistance (Rct) and mass transfer resistance (Rmt) of the cross-linked nanofibers membrane electrode were lower than that of the conventional catalyst. Cyclic voltammetry (CV) curves were produced under the condition of hydrogen and nitrogen gas flow at the anode and cathode, respectively; the electrochemical specific surface area (ECSA) of each membrane electrode was obtained by calculating the hydrogen desorption region of CV Curves (Supplementary Fig. 18d). The results show that the cross-linked nanofiber electrode has a larger ECSA than the traditional spraying electrode (51.7 m2 g-1 vs. 42.0 m2 g-1). The increase in the ECSA value indicates that the utilization rate of the Co-COP/AF-GC catalyst was significantly improved. This is understood to be the result of the membrane electrode prepared by the conventional spraying nanoparticle Co-CoP/AFGC catalyst having the characteristics of an ultra-compact, large flow resistant, and discontinuous electron transport network. However, the macropores in the single fibers obtained by electrospinning Co-COP/AFGC catalyst were conducive towards enhancing the diffusion of reactants to the active site. The micron-sized secondary macropores between the fiber skeleton allow a large amount of gas diffuse into the catalyst layer under lower pressure, forming a high-speed electron transport network. Additionally, the highly graphitized carbon matrix in Co-COP/AF-GC catalyst proved to be beneficial in increasing the electronic conductivity of the catalyst. Therefore, Co-COP/AFGC cathode has a unique porous fibers morphology and well-designed layered porous structure, which makes the membrane electrode have both a higher active site utilization as well as efficient mass transfer performance, improving the performance of PEMFCs.
Uniformity and durability of hierarchically porous nanofiber electrode
Moreover, to verify that the catalyst was evenly distributed in the obtained cross-linked nanofiber membrane, we conducted a performance evaluation on different parts of the membrane, which showed good consistency (Fig. 4g). One of the main reasons hindering the commercial application of catalysts is that the performance of the catalysts obtained by the scale-up preparation are very different from that obtained by laboratory tests, which is due to many uncertain factors in the macro preparation processes. Conversely, we use electrospinning technology to evenly distribute the liquid processability COP-based catalyst in the carbon-supported nanofibers membrane, which provides a good direction for the commercial application of these catalysts. In the conventional spraying and cross-linked nanofiber electrode, the relationship between catalyst, proton accessibility and RH distribution were shown in Supplementary Fig. 20. For traditional spraying electrodes, it can be seen that with the increase of RH, the total pore size and void volume of the electrode decrease, and the gas path to the active site becomes less direct. However, for the cross-linked nanofiber electrode the connectivity of the ion network increases due to the existence of the three-dimensional porous framework, which is more conducive towards proton conduction through the catalytic layer.
To investigate the durability of the electrode, the long-term stability of the three catalysts was tested using an H2-air fuel cell at a constant current density of 0.4 A cm-2 (80 °C, 0.1 L min-1 of gas flow at 1 bar back pressure and 100% RH). As shown in Fig. 4i, compared with the traditional spraying and PAA nanofibers membrane electrode, the cross-linked nanofiber electrode maintained 85% of the original voltage after the 150-hour durability test, demonstrating the cross-linked nanofiber electrode’s excellent stability. The cross-linked nanofiber electrode, attesting to much improved catalyst stability, even compared with the popular monodispersed atomic metal embedded in nitrogen-doped carbon catalysts, which usually decreases to around 60% after 100 hours operation19–21. In the transient state, the dynamic cycle conditions operated sequentially from 50 to 700 mA cm-2, and the results showed that no severe water flooding occurred during the cycle operation (Fig. 4h). The durability test was further conducted via cyclic voltammetry. After 5000 cycles, the performance of cross-linked nanofiber electrode decreased by 6% (from 304 to 286 mW cm-2). Meanwhile, the maximum power density of the conventional spraying electrode was reduced by 54% (from 178 to 82 mW cm-2). Even after 10,000 cycles, the maximum power density of cross-linked nanofiber electrode was still 274 mW cm-2, exhibiting a performance retention rate of 90% (Supplementary Fig. 21b-d). The EIS of the cross-linked nanofiber electrode was studied at the current density of 0.5 A cm-2 during the durability test. After 5000 cycles, the changes in Rct and Rmt were very small (Rct increased from 0.02 Ω cm2 to 0.028 Ω cm2, and the Rmt varied from 0.18 Ω cm2 to 0.21 Ω cm2), indicating that the degradation of the catalytic layer was negligible. However, for the conventional spraying electrode, the value of Rmt was enlarged by 2.16 times after 5000 cycles (Supplementary Fig. 21e, f). The increase of Rmt reflects the degradation of the Co-COP/AFGC catalytic layer, which seriously impacts the mass transfer. Consistent with the decline of fuel cell performance, the ECSA loss of the cathode shows a similar trend (Supplementary Fig. 22). For instance, after 5000 cycles, the ECSA of the cross-linked nanofibers decreased by 10.6%, while a significant loss of 51.7% was observed for the conventional spraying electrode after 5000 cycles (ECSA values changed from 42-20.3 m2g-1). Furthermore, despite undergoing 10000 cycles, the ECSA loss of cross-linked nanofibers was only 14.5% (from 51.7-44.2 m2g-1).
Generally, one of the reasons for the degradation of catalyst stability is the occurrence of the Fenton reaction. To assess levels of reactive oxygen species (ROS) production, UV/visible absorption spectrometry was performed using 2, 20-azophytes (3-ethylbenzylthiazoline-6-sulfonate) (ABTS) as substrates. The absorbance value of Co-COP/AFGC in Supplementary Fig. 23 was like that without a catalyst, which verifies the proposed the inhibition effect of the Co-COP/AFGC catalyst on ROS in the reaction process. In addition, demetallation was another major degradation mechanism that was often recognized. We used Co-COP/AFGC and Fe-COP/AFGC catalysts to perform an electrochemical cycle of 0.6 - 1.0 V (relative to RHE) in 0.1 M HClO4 for 30 min, and then conducted an LSV test (Supplementary Fig. 24). The results show that the electrocatalytic performance of Fe-COP/AFGC decreases significantly compared with Co-COP/AFGC.
To understand the observed trend of Co-COP/AFGC stability, we conducted first-principles density functional theory (DFT) calculations to predict the demetallation trend of M-N4 active sites. The Co-COP and Fe-COP models and the AFGC composite models were constructed, respectively. Accordingly, only van der Waals interaction was observed during the electron localization function (ELF) analysis in the Co-COP/AFGC catalyst (Supplementary Fig. 26). As confirmed by the comparison of charge density differences under the same isosurface levels, the electrons on the Co active site of the Co-COP/AFGC catalyst were transferred to AFGC (Supplementary Fig. 27). Combined with the previous demetallation calculation results, it indicates that there was a stronger interaction between Co-COP and AFGC, leading to a high resistance to demetallation from the Co-N4 active sites. The ability to demetallize was verified by calculating the change in Gibbs free energy (ΔG) before and after demetallation (Table S6). The results show that the ΔG (-6.881 eV) of Co-COP/AFGC was more negative than that of Fe-COP/AFGC (-6.758 eV), which shows that Co-N-C is more resistant to demetallation than Fe-N-C, which is consistent with the above experimental results. ELF analysis and electronic structure information of other models are shown in Supplementary Fig. 28-31.
Oxygen transport enhancement in hierarchically porous nanofiber electrode
Although our catalyst exhibits a relatively stable structure, its durability was relatively weak while assembled into fuel cell devices by the conventional spraying method. Therefore, we obtained an integrated membrane electrode by electrospinning and heating esterification, which maximized the exposure of the active sites, improved the resistance during gas transmission, and greatly improved the stability of the PEMFC. To further characterize the distribution of O2 in the MEA, the conventional spraying and cross-linked nanofiber electrode were sliced by a focused ion beam field emission scanning electron microscope (FIB-FESEM), and finally the intrinsic internal structure of the sample was obtained by 3D reconstruction (Supplementary Fig. 33-35, Supplementary Video 1 and 2). Most of the methods to explore the influence of different catalytic layer structures on fuel cells are used to establish a theoretical model in order to simulate the reaction process. This model is generally an ideal homogenization, but the actual catalytic layer structure is often more complex and diverse. The most realistic structure of the catalytic layer can be obtained by FIB-FESEM slice and 3D reconstruction technology, as shown in this work (Fig. 5a, d). The reconstructed 3D structure from FIB-FESEM was imported by using the ImportGeo module based on GeoDict2022. The ProcessGeo module was applied to output the geometric structure file format after processing the 3D image, which was used for further computational fluid dynamics (CFD) analysis. A comprehensive single phase diffusion model considering the transport process and concentration distribution of O2 in the membrane electrode was established by solving the Brinkman equations, which describes the fluid flow in porous media. The diffusion of oxygen in conventional spraying and cross-linked nanofiber structure were both simulated, and the corresponding oxygen concentration distribution obtained is shown in Fig. 5c, f. Generally, the membrane electrode structure assembled by the nanofiber catalytic layer can increase the specific surface area, making the contact area of the reaction larger so that the reaction is more thorough.
Additionally, it can also improve the porosity, which is critical for high performance mass transfer. In the simulation result diagram, the membrane electrode structure with similar morphology to the porous nanofibers therefore exhibits lower oxygen diffusion resistance and has a higher evenly distributed oxygen concentration in the entire three-phase reaction interface. However, the conventional spraying electrode tends to obstruct oxygen diffusion into the catalyst and has relatively lower oxygen concentration. The diffusion coefficient was the proportionality constant between the molar flux caused by molecular diffusion and the concentration gradient (or diffusion driving force) of the species, with its magnitude depending on the properties of the diffused material. According to Fick's first law34, the relative diffusion coefficients of conventional spraying and cross-linked nanofiber electrodes were measured to be 13.4% and 46.9%, respectively. The high relative diffusion coefficient can provide ideal channels for reactants and electrolytes to enter the active center. The porous structure can discharge the produced water in time, improving the power density and stability of PEMFCs.
Moreover, the distribution of relaxation time (DRT) analysis was performed on the obtained EIS data to distinguish the corresponding electrochemical reaction processes at characteristic frequencies. For all selected EIS data, the Kramers-Kronig validity test showed less than 1% error, meeting the requirements for further DRT impedance analysis. As shown in the DRT spectrum (Fig. 5g), the integration results of the characteristic peaks show that the ORR charge transfer resistance, oxygen diffusion resistance, and proton conduction resistance of the cross-linked nanofibers were all lower than those of the conventional spraying electrode, which further indicates that the transmission performance of the overall electrode can be greatly improved by preparing a specific electrode structure. Additionally, the oxygen adsorption capacity of different MEA electrodes was tested by chronoamperometry. Different working electrodes were first immersed in an O2 saturated solution to ensure the exposed active site was filled with adsorbed O2. Then, the working electrode was transferred to an electrolytic cell filled with N2, and chronoamperometry was performed at 0.7 V to consume the O2 adsorbed in the working electrode. As shown in Fig. 5h, the cross-linked nanofiber electrode adsorbed 7.79×10-8 mol O2, much higher than the powder Co-COP/AFGC (3.87×10-8 mol), indicating that the cross-linked nanofiber electrode exposed much more active sites. The results show that mesopores and macropores enhance the mass transfer of reactants and products, thus increasing the density of accessible active sites. Therefore, we can conclude that the nanofiber electrode can improve oxygen utilization and thus boost the performance of fuel cells.
The SEM images in Supplementary Fig. 36 illustrate the structure and morphology of the membrane electrode after the fuel cell test. Firstly, there were many trenches formed on the traditional membrane spraying electrode due to the erosion of water after the test, making the distribution of catalyst gradually uneven. The structure and morphology of the PAA nanofiber electrode clearly collapsed after the test, and the three-dimensional porous frame structure was destroyed. In contrast, the morphology of the cross-linked nanofiber electrode still maintained its original state, suggesting that the cross-linked nanofiber electrode obtained by esterification of two polymers containing carboxyl group and hydroxyl group has excellent water stability. While after the durability test, the XANES and EXAFS spectra of the Co-COP/AFGC cathode catalyst recovered, confirming the stability of the structure (Supplementary Fig. 37). The XPS peaks associated with Co2+ and Co3+ were retained for the cross-linked nanofiber catalyst after the stability test, and the categories of C have not been changed, both of which further indicate that the cross-linked nanofibers were relatively stable during the ORR process (Fig. 5i).
In this work, a highly stable cross-linked nanofiber electrode was directly prepared by electrospinning technology based on a liquid processability Co-COP obtained using a pyrolysis-free strategy. The liquid processability Co-COP contains robust conjugated systems and abundant ordered N-coordinated Co single-atom centers. The hierarchical porous structure obtained by electrospinning technology fully exposes the Co-N4 active sites. Compared with the conventional spraying electrode morphology, the macropores of single nanofibers in the electrospinning film are beneficial for the diffusion of gas to the active sites. The secondary micron-scale macropores between the filament skeletons make a large amount of gas diffuse into the catalyst layer through the low-pressure drop, forming a high-speed electron transfer network. The lower oxygen diffusion resistance was further confirmed by FIB-FESEM and CFD experiments. The relative diffusion coefficient was 46.9% according to Fick's first law, which can provide a good channel for reactants to enter the active sites and remove the produced water so as to not inhibit the reaction. The results show that the power density of the integrated nanofiber electrode is 1.72 times higher than that of the conventional spraying electrode. Significantly, the cross-linked nanofiber electrode could still maintain 85% of the original voltage after 150 hours of durability testing. Moreover, this nanofabrication technique has good scalability. For example, the performance of the obtained 200 cm2 film could maintain good consistency at a randomly selected position on that film. By designing a specific electrode structure to maximize the overall electrode transport efficiency and improve the availability of active sites, we have overcome the transport limitation of conventional spraying electrodes and provided a new scheme for membrane electrode processing of PEMFCs in the future.
Materials Synthesis
Synthesis of Co-COP/AF-GC. A liquid processability Co-COP with a Co-N4-C structure was synthesized by pyrolysis-free method. XC-72 was carbonized in a high-temperature programmed furnace to obtain a highly graphitized carbon material, and then reacted with concentrated nitric acid for 12 h. Then, the Co-COP/AFGC catalyst was prepared by combining the Co-COP and acid functionalized graphite carbon (AFGC).
Synthesis of nanofibers electrode. Firstly, the ink was prepared according to the following scheme: 200 mg Co-COP/AF-GC was used as catalyst, 2.08 g 5% Nafion solution was added as ionomer, and ultrasonicated for 30 min. Subsequently, 15 mg of polyacrylic acid (PAA) and 15 mg of polyvinyl alcohol (PVA) were added and stirred at room temperature for 24 hours to obtain the final ink. (Because PVA is difficult to dissolve, so PVA solution was first prepared. Specifically, 0.5 g PVA was dissolved in 20 mL water and stirred at 80 ℃ for 12 h to obtain 5% PVA solution.) The ink was sucked into a 5 mL syringe and electrospun using a 19-gauge stainless steel needle. A piece of aluminum foil was wrapped on the collector, the rotation speed was set at 150 rpm, the distance between the collector and the needle was 10 cm, the injection speed was 0.6 mL h− 1, and the voltage was 8 kV. After spinning, the membrane was separated from the collector, and then placed in an oven at 140°C for cross-linking and esterification for about 3 h. The PAA nanofiber electrode was prepared using the same method except that the polymer PAA/PVA was replaced by PAA and without heat treatment.
Fuel cell performance testing
The fuel cell test station (Scribner 850E) was used for testing in current-scanning mode. High purity H2 and air/oxygen were introduced into the anode and cathode respectively, and the flow rates were set at 0.4 L min− 1 and 0.8 L min− 1, respectively. The gas back pressure on the two electrodes was set to 1 bar, and the relative humidity (% RH) was controlled by adjusting the anode and cathode gas temperature and the fuel cell temperature. The polarization curves and power densities at different voltages were obtained by scanning current method. The EIS for H2 and O2/air at different current densities was recorded in the frequency range of 100 kHz to 0.1 Hz. Using Gamry software to calculate and simulate the system, and the impedance equivalent circuit diagram was obtained. To investigate the electrochemically active surface area (ECSA), cyclic voltammetry (CV) tests were performed by using the CHI 630D (CH Instruments, USA) module. With H2 as the anode and N2 as the cathode, CV curves were obtained in the potential range of 0.05 V to 1.1 V at a scan rate of 20 mV s− 1.
FIB-FESEM experimental process
Firstly, Small block samples were selected and glued to the nail sample table. After cutting, grinding and polishing by German Leica EM TXP precision all-in-one machine, ion polishing was carried out. Secondly, the German Leica RES 102 ion polishing instrument was used for surface argon ion polishing: two ion guns, the ion gun voltage was 5 kV, the current was 2.2 mA, the ion gun inclination angle was 2º during polishing, and the polishing time was 4 hours, and the polishing area was about 1 cm2. Thirdly, the carbon conductive layer was coated on the surface, and the British EMITECH K950X coating instrument was used to coat the carbon conductive layer on the surface of the sample, and the tip of the carbon rod used for evaporation was about 2 mm.
Data availability
Data generated or analyzed during this study are included in this article and/or its supplementary information files. Raw data are available from the authors on request.
Acknowledgments
This work was supported by the National Key Research and Development Program of China (2017YFA0206500); the NSF of China (21922802); the Beijing Natural Science Foundation (JQ19007); Talent cultivation of State Key Laboratory of Organic-Inorganic Composites; “Double-First-Class” construction projects (XK180301, XK1804-02); Distinguished Scientist Program at BUCT (buctylkxj02). We thank the 1W1B station for XAFS measurements in Beijing Synchrotron Radiation Facility (BSRF).
Author Contributions
B. Y. performed the synthesis, structural characterizations and electrochemical tests. Q. H. performed CFD simulations. L. H. conducted DFT calculation. Y. L. helped in the electrospinning experiment. B. Y., T. O., X. Y., G. W. and Z. X. analyzed the results and wrote this paper. Z. X. supervised and led this project. All authors commented on the paper.
Competing interests
The authors declare that they have no competing interests.
Additional information
Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/.
- Setzler, B. P., Zhuang, Z., Wittkopf, J. A. & Yan, Y. Activity targets for nanostructured platinum-group-metal-free catalysts in hydroxide exchange membrane fuel cells. Nat. Nanotechnol. 11, 1020–1025 (2016).
- Xie, X., Shang, L., Xiong, X., Shi, R. & Zhang, T. Fe Single-Atom Catalysts on MOF-5 Derived Carbon for Efficient Oxygen Reduction Reaction in Proton Exchange Membrane Fuel Cells. Adv. Energy Mater. 12, 2102688 (2022).
- He, Y. et al. Single Cobalt Sites Dispersed in Hierarchically Porous Nanofiber Networks for Durable and High-Power PGM-Free Cathodes in Fuel Cells. Adv. Mater. 32, 2003577 (2020).
- Tong, M. et al. Operando Cooperated Catalytic Mechanism of Atomically Dispersed Cu – N4 and Zn – N4 for Promoting Oxygen Reduction Reaction. Angew. Chemie - Int. Ed. 60, 14005–14012 (2021).
- Wang, J. et al. Synergistic effect of well-defined dual sites boosting the oxygen reduction reaction. Energy Environ. Sci. 11, 3375–3379 (2018).
- Chen, Y. et al. Isolated Single Iron Atoms Anchored on N-Doped Porous Carbon as an Efficient Electrocatalyst for the Oxygen Reduction Reaction. Angew. Chemie - Int. Ed. 56, 6937–6941 (2017).
- Wang, G. et al. Highly active atomically dispersed CoN4 fuel cell cathode catalysts derived from surfactant-assisted MOFs: Carbon-shell confinement strategy. Energy Environ. Sci. 12, 250–260 (2019).
- Liu, J. et al. Honeycomb-like Self-Supported Co-N-C Catalysts with an Ultrastable Structure: Highly Efficient Electrocatalysts toward Oxygen Reduction Reaction in Alkaline and Acidic Solutions. ACS Appl. Energy Mater. 4, 2522–2530 (2021).
- Wang, Z. et al. Understanding the Synergistic Effects of Cobalt Single Atoms and Small Nanoparticles: Enhancing Oxygen Reduction Reaction Catalytic Activity and Stability for Zinc-Air Batteries. Adv. Funct. Mater. 31, 2104735 (2021).
- Ye, H. et al. Sustained-Release Method for the Directed Synthesis of ZIF-Derived Ultrafine Co-N-C ORR Catalysts with Embedded Co Quantum Dots. ACS Appl. Mater. Interfaces 12, 57847–57858 (2020).
- Shi, C. et al. Hierarchical N-doped carbon spheres anchored with cobalt nanocrystals and single atoms for oxygen reduction reaction. Nano Energy 87, 106153 (2021).
- Das, G. et al. Multifunctional redox-tuned viologen-based covalent organic polymers. J. Mater. Chem. A 4, 15361–15369 (2016).
- Skorjanc, T., Shetty, D. & Trabolsi, A. Pollutant removal with organic macrocycle-based covalent organic polymers and frameworks. Chem 7, 882–918 (2021).
- Miao, Z. et al. A Novel Strategy for the Construction of Covalent Organic Frameworks from Nonporous Covalent Organic Polymers. Angew. Chemie - Int. Ed. 58, 4906–4910 (2019).
- Guo, J., Lin, C. Y., Xia, Z. & Xiang, Z. A Pyrolysis-Free Covalent Organic Polymer for Oxygen Reduction. Angew. Chemie - Int. Ed. 57, 12567–12572 (2018).
- Hijazi, I. et al. Carbon nanotube-templated synthesis of covalent porphyrin network for oxygen reduction reaction. J. Am. Chem. Soc. 136, 6348–6354 (2014).
- Ma, R. et al. Pyrolysis-free synthesis of single-atom cobalt catalysts for efficient oxygen reduction. J. Mater. Chem. A 10, 5918–5924 (2022).
- Peng, P. et al. In Situ Charge Exfoliated Liquid processability Covalent Organic Framework Directly Used for Zn-Air Flow Battery. ACS Nano 13, 878–884 (2019).
- Shui, J., Wang, M., Du, F. & Dai, L. N-doped carbon nanomaterials are durable catalysts for oxygen reduction reaction in acidic fuel cells. Sci. Adv. 1, e1400129 (2015).
- Fu, X. et al. Tailoring FeN4 Sites with Edge Enrichment for Boosted Oxygen Reduction Performance in Proton Exchange Membrane Fuel Cell. Adv. Energy Mater. 9, 1803737 (2019).
- Fu, X. et al. Evolution of atomic-scale dispersion of FeNx in hierarchically porous 3D air electrode to boost the interfacial electrocatalysis of oxygen reduction in PEMFC. Nano Energy 83, 105734 (2021).
- Hwang, W., Kim, S., Ahn, C. Y., Cho, Y. H. & Sung, Y. E. Origami-Based Flexible and Simple Tubular Polymer Electrolyte Membrane Fuel Cell Stack. ACS Energy Lett. 6, 3195–3202 (2021).
- Wang, Z. et al. Fe, Cu-Coordinated ZIF-Derived Carbon Framework for Efficient Oxygen Reduction Reaction and Zinc–Air Batteries. Adv. Funct. Mater. 28, 1802596 (2018).
- Qi, M. et al. A robust esterified nanofibre electrode for proton exchange membrane fuel cells. J. Mater. Chem. A 8, 5298–5307 (2020).
- Shui, J., Chen, C., Grabstanowicz, L., Zhao, D. & Liu, D. J. Highly efficient nonprecious metal catalyst prepared with metal-organic framework in a continuous carbon nanofibrous network. Proc. Natl. Acad. Sci. U. S. A. 112, 10629–10634 (2015).
- Wang, G., Osmieri, L., Star, A. G., Pfeilsticker, J. & Neyerlin, K. C. Elucidating the Role of Ionomer in the Performance of Platinum Group Metal-free Catalyst Layer via in situ Electrochemical Diagnostics. J. Electrochem. Soc. 167, 044519 (2020).
- Wang, J. et al. Synergistic effect of well-defined dual sites boosting the oxygen reduction reaction. Energy Environ. Sci. 11, 3375–3379 (2018).
- Mardle, P., Thirunavukkarasu, G., Guan, S., Chiu, Y. L. & Du, S. Comparative Study of PtNi Nanowire Array Electrodes toward Oxygen Reduction Reaction by Half-Cell Measurement and PEMFC Test. ACS Appl. Mater. Interfaces 12, 42832–42841 (2020).
- Wang, X. et al. Electrospun Thin-Walled CuCo2O4@C Nanotubes as Bifunctional Oxygen Electrocatalysts for Rechargeable Zn-Air Batteries. Nano Lett. 17, 7989–7994 (2017).
- Kabir, S. et al. Toward Optimizing Electrospun Nanofiber Fuel Cell Catalyst Layers: Microstructure and Pt Accessibility. ACS Appl. Energy Mater. 4, 3341–3351 (2021).
- Uddin, A. et al. High Power Density Platinum Group Metal-free Cathodes for Polymer Electrolyte Fuel Cells. ACS Appl. Mater. Interfaces 12, 2216–2224 (2020).
- Xu, X. et al. Fe-N-C with Intensified Exposure of Active Sites for Highly Efficient and Stable Direct Methanol Fuel Cells. ACS Appl. Mater. Interfaces 13, 16279–16288 (2021).
- Wan, X. et al. Fe–N–C electrocatalyst with dense active sites and efficient mass transport for high-performance proton exchange membrane fuel cells. Nat. Catal. 2, 259–268 (2019).
- Wang, B., Zhang, H., Yang, Y. & Xu, M. Diffusion and filamentous bacteria jointly govern the spatiotemporal process of sulfide removal in sediment microbial fuel cells. Chem. Eng. J. 405, 126680 (2021).
There is NO Competing Interest.
- supportinginformation.docx
supporting information