Recently, the research group of Professor Wu Gang of the State University of New York at Buffalo published an academic paper entitled “Atomically-dispersed iron sites with a nitrogen-carbon coating as highly active and durable oxygen reduction catalysts for fuel cells” in the journal Nature Energy. The study is the first to report an iron-based catalyst comparable to the precious metal platinum, and the activity and stability of the catalyst have been systematically studied and confirmed in real fuel cells.
This is the result of close cooperation between several scientific institutions. Among them, Professor Xie Jian of Indiana University and Purdue University Indianapolis Joint Branch (IUPUI) and Dr. Hui Xu of Güner Company jointly tested and evaluated the performance of catalysts in fuel cells; The team of Dr. David A. Cullen at Oak Ridge National Laboratory provided high-quality electron microscopy characterization; Dr. Deborah Myers of Argonne National Laboratory used the 57Fe Mössbauer spectroscopy to illustrate the key transformation of the Fe-N-C catalyst at the active site during chemical vapor deposition. Professor Wang Guofeng’s research group at the University of Pittsburgh revealed the source of fe-N-C catalyst activity and stability from the theoretical level through systematic simulation calculations. Professor Feng Zhenxing’s research group of Oregon State University analyzed the structure of catalyst atomic dispersion with the help of synchrotron radiation. Dr. Shawn Litster’s research group at Carnegie Mellon University used X-ray imaging to characterize the structural evolution of the catalytic layer in detail.
The development of non-platinum group metal (PGM-free) catalysts with high activity and stability is key to significantly reducing the cost of proton exchange membrane fuel cells (PEMFCs). In a harsh acidic electrolyte environment, a single-atom dispersed FeN4 locus catalyst (i.e., Fe-N-C) exhibits very good oxygen reduction reaction (ORR) activity. Since the discovery of the metal macrocyclic compound in 1964, the activity of the Fe-N-C catalyst is steadily approaching commercial Pt/C through the continuous efforts of researchers, but stability is still a key obstacle to its application. Due to the difficulty of combining high activity and high stability, it is difficult to further improve the catalytic performance of Fe-N-C.
Carbon planes with local microscopic defects help break oxygen-oxygen bonds. However, this defect-rich carbon structure is prone to oxidation, which in turn triggers the catalyst FeN4 site to dissolve the metal during the ORR process. In contrast, the graphitized carbon structure has higher corrosion resistance/oxidation resistance and can stabilize the FeN4 active site. However, this highly graphitized carbon structure lacks abundant defects and doped nitrogen to carry sufficient active site density. Therefore, the coordination environment and local carbon structure of FeN4 are critical to its intrinsic activity and stability.
Typically, Fe-N-C is synthesized by high-temperature pyrolysis, which results in a less controlled FeN4 coordination environment and a complex variety of active species. For a long time, there has been uncertainty about the origin of OrR activity produced by Fe-N-C, so it is difficult to correlate the catalyst micro-coordination environment with ORR performance. Recently, Jaouen et al. defined the two FeN4 site configurations of the Fe-N-C catalyst by in situ online Mössbauer spectroscopy. One is the FeN4C12 locus (denoted as the S1 locus) with four pyrrole-type N ligands, such as Fe (Pc); The second is the FeN4C10 site (denoted as the S2 site) with four pyridine-type N ligands, whose structure is consistent with the molecular catalyst synthesized by Surendranath et al. The S1 site has higher ORR intrinsic activity, but is easily irreversibly converted to inactive ferric trioxide during the ORR process. The S2 site has higher resistance to the removal process of metals and higher stability, but its ORR activity is relatively low. Therefore, in catalyst design, it is crucial to reasonably increase the density of the S2 site.
To this end, Professor Wu Gang’s team developed a high stability and high activity Fe-N-C preparation method, which realizes the excellent stability of the synthesized catalyst while ensuring high activity by modulating the carbon structure and coordination environment of the FeN4 site. The study first used high-temperature treatment of ZIF-8 precursors containing Fe2O3 nanoparticles, followed by critical high-temperature NH4Cl treatment to provide a wealth of defects. The prepared Fe-N-C catalyst (Fe-AC) has a large number of highly active S1 sites, under acidic conditions, the half-wave potential (E1/2) is as high as 0.915 VRHE, which is significantly higher than that of commercial Pt/C catalysts; Under fuel cell test conditions, the area specific activity was as high as 41.1 mA cm−2 (@0.9 ViR-free, H2-O2, 150 kPaabs, 80 °C), close to the target set by the U.S. Department of Energy (DOE). However, due to the low stability of the S1 site, the ORR activity of Fe-AC is difficult to sustain. On this basis, the researchers used chemical vapor deposition (CVD) technology to further deposit nitrogen-doped carbon films on the Fe-AC catalyst to prepare the Fe-AC-CVD catalyst. This CVD process achieves a transition from a highly active S1 site to a highly stable S2 site. Thanks to the formation of a large number of S2 sites, the stability of Fe-AC-CVD has been significantly improved. The results of the accelerated attenuation test conducted in the membrane electrode assembly (MEA) show that the stability of Fe-AC-CVD is significantly better than that of commercial Pt/C, and the performance after the attenuation test is the same as that of Pt/C. Fe-AC-CVD also showed good stability during the 320-hour fuel cell stability test, with a current attenuation rate of only 0.132 mA cm−2 h−1 at 0.67 V. In addition, the source of Fe-N-C activity and stability was elaborated by means of STEM at atomic-level resolution, X-ray absorption spectroscopy, 57Fe Mössbauer spectroscopy and DFT calculation, laying a foundation for the further design and preparation of highly active and stable Fe-N-C catalysts.
Figure 1: The formation of fe-N-C active sites was explored with the help of in situ transmission electron microscopy and EELS with atomic-level resolution. First, NH4Cl was used to prepare a high-performance single-atom dispersed Fe-N-C catalyst, and then the stability of the catalyst was further improved by CVD technology.
Figure 2: The structure of the catalyst was studied with the help of HR-TEM, XANES, EXAFS, and EELS with atomic-level resolution. The XANES and EXAFS results show that the Fe atoms in the catalyst are mainly present in a form similar to FeN4 in the standard sample ironphthalocyanine. The RRDE test showed that the Fe-N-C catalyst (Fe-AC) prepared by NH4Cl treatment exhibited ORR activity far exceeding that of commercial Pt/C under acidic conditions, but its stability was not good. After the CVD treatment, the ORR activity of Fe-AC (Fe-AC-CVD) was slightly attenuated, but its stability was significantly improved.
Figure 3: Fuel cell testing studies the activity and stability of the Fe-N-C catalyst. The results showed that the area specific activity of Fe-AC in MEA was as high as 41.1 mA cm−2 (@0.9 ViR-free, H2-O2, 150 kPaabs, 80°C), which was close to the target set by DOE. Fe-AC-CVD after CVD treatment remained highly active in MEA. Consistent with the results of the RRD study, Fe-AC exhibited excellent initial performance in MEA, but was poorly stable; Compared to Fe-AC, Fe-AC-CVD performance was only slightly attenuated after 30,000 laps of fuel cell durability testing. In addition, Fe-AC exhibits excellent initial performance at 0.67 V, but the current decays rapidly; Fe-AC-CVD exhibits good stability, with a current decay rate of only 0.132 mA cm−2 h−1 at 0.67 V, which is even better than commercial Pt/C catalysts.
Figure 4: The structural evolution of the Fe-N-C cathode catalytic layer is explored in depth by nano-X-ray imaging technology and electrochemical impedance spectroscopy. Nano-X-ray imaging technology showed that the Fe-N-C catalyst did not undergo significant changes in particle size and porosity during the accelerated attenuation test. Electrochemical impedance spectroscopy tests show that after fe-AC accelerated attenuation at 30000 laps, the charge transfer resistance of 0.8 A cm−2 increased from 0.2510 to 0.3420 Ω cm2, while the charge transfer resistance corresponding to Fe-AC-CVD increased from only 0.2695 to 0.2755 Ω cm2。
Figure 5: The transformation of carbon structure during CVD was studied using in situ scanning transmission microscopy (IL-STEM) and the K-side EELS orientation distribution of carbon elements. Studies have shown that after CVD treatment, the surface of the Fe-AC catalyst is covered with a highly graphitized carbon layer of about 1 nm. STEM-EDS analysis showed that the nitrogen content of the carbon layer was higher than that of the carbon structure of the inner layer, and the IL-STEM test showed that the carbon layer contained single atom Fe, which showed that the CVD treatment did not completely mask the active site of the catalyst.
Figure 6: The transformation of Fe-AC active sites during CVD was studied with the help of 57Fe Mössbauer spectroscopy and DFT calculations. The 57Fe Mössbauer test showed that the Fe-AC catalyst contained two active sites of ORR: the S1 site with high activity but poor stability (FeN4C12), and the S2 site with excellent stability but weak activity (FeN4C10). Prior to CVD treatment, the relative ratios of S1 and S2 loci in Fe-AC were 65% and 29%, respectively, which explains the excellent initial ORR activity of Fe-AC, but this high activity is difficult to maintain in a fuel cell environment. After the CVD treatment, the relative proportions of S1 and S2 loci in Fe-AC-CVD were 53% and 42%, respectively, indicating that Fe-AC-CVD achieved a good balance in activity and stability. The study found that the S2 locus has a more negative forming energy than the S1 locus, indicating that the S2 locus is more likely to form thermodynamically, which lays the foundation for the transition from the S1 locus to the S2 locus. In addition, DFT calculations show that the five-membered carbon ring with pyrrole-type nitrogen can open the ring at high temperature conditions and capture carbon elements from the environment, and then transform into a six-membered carbon ring with pyridine-type nitrogen, thereby realizing the transition from the S1 site to the S2 site.
Figure 7: The DFT calculation illustrates the difference in ORR activity between the two active sites of Fe-AC (Figures 7a and 7b correspond to the S1 site, and Figures 7c and 7d correspond to the S2 sites). Charge distribution calculations show that Fe in the S2 locus has more empty d orbitals, which have stronger adsorption of the O2 molecule, resulting in difficult dissociation of the O-O bond, hence the 4e at the S2 locus− ORR activity is lower than the S1 site.
Figure 8: The DFT calculation illustrates the difference in stability between the two active sites of Fe-AC-CVD (S1 and S2). To this end, the authors propose a mechanism for inactivation of the FeN4 active site, which consists of three primitive steps: 1. Protonation of two adjacent nitrogen atoms in the FeN4 site; 2. Fe center combines with O2 to form O2-FeN2 center; Fe-O2 is removed from the N4 coordination environment, eventually forming an N4 site with holes. DFT calculations show that primitive step 1 has the highest energy barrier and is therefore a rapid step deactivated by the FeN4 active site. The study further shows that for the S2 locus (Figure 8b), the energy barrier of the primitive step 1 is 2.08 eV, which is much higher than the corresponding energy barrier of the S1 locus. Therefore, the S2 locus has higher eigen-stability than S1.
(Source: Science Network)
Related paper information:https://doi.org/10.1038/s41560-022-01062-1