Inspired by the promise of the rational design presented above, MIL-53(Fe)-2OH and MOF-74-Fe catalysts were synthesized by using 2,5-dihydroxyterephthalic acid (H4DOBDC) as ligand and iron as the metal component. Synthesis of MOFs involves solid-to-solid rearrangement and intermediate formation.15-17 As shown in Figure 1a and Figure S1, H4DOBDC have two carboxylic groups rigidly located at an angle of 180° (1,4-benzenedicarboxylic acid functionalities) and two phenolic functional groups.18,19 Deprotonation of H4DOBDC can form two kinds of ligand anions: 2,5-dihydroxyterephthalato anion ([H2DOBDC]2−) with the pKa value of 7.31 and 2,5-dioxidoterephthalato tetra-anion ([DOBDC]4−) with the pKa value of 26.67.20,21 Alkaline environment can be formed spontaneously through the solvothermal decomposition, even if base is absent from the starting reagents.15 The quadridentate ligand [DOBDC]4− can be easily formed, which tends to be incorporated into MOF-74-Fe (Figure S2).15 By controlling the solvent ratio and concentration, H4DOBDC can be selectively deprotonated at only the two carboxylic groups, and consequently form the bidentate anion [H2DOBDC]2−, which leads to the formation of MIL-53(Fe)-2OH (Figure S3).22 Meanwhile, DFT calculations have also revealed the distinctly different electronic structures of these two MOFs. From the electronic distributions near the Fermi level (EF), it is evident that the bonding orbitals in MOF-74-Fe are dominated by the carbon chains with only limited contributions from the Fe sites, leading to the relatively weak electroactivity of Fe (Figure 1b). In MIL-53(Fe)-2OH, however, Fe sites display the highly electron-rich feature, indicating high electroactivity (Figure 1c). Meanwhile, O sites of the phenolic groups and the connection points also facilitate the electron transfer within MIL-53(Fe)-2OH for improved OER performance. The highly distinct electronic structures will lead to the different OER performances in these two MOFs. The projected partial density of states (PDOSs) in Figure 1d further reveals the distinct electronic structures. Notably, Fe-3d orbitals show a large eg-t2g splitting of 3.43 eV, leading to the large barriers for electron transfer in MOF-74-Fe. OH-s,p orbitals locate in the deep position, which mainly serves as the electron reservoir, and the resulting limited overlap between Fe-3d and OH-s,p orbitals increases the difficulties of site-to-site electron transfer. In contrast, the eg-t2g splitting of Fe-3d orbitals has been significantly alleviated to 1.19 eV in MIL-53(Fe)-2OH, which supports much more efficient electron transfer from the MOF to the intermediates (Figure 1e). C-s,p orbitals cover a broad range in both MOFs to facilitate electron depletion. More importantly, two different types of O sites have synergistically improved the OER performance. The O-s,p orbitals from Fe-O-Fe chains have shown a close position towards the EF with good overlap with both Fe-3d and OH-s,p orbitals, indicating that the phenolic groups in MIL-53(Fe)-2OH are able to ensure the fast site-to-site electron transfer.
The crystal structures of the MIL-53(Fe)-2OH and MOF-74-Fe catalysts were investigated by X-ray powder diffraction (XRD). The experimental XRD pattern of the MIL-53(Fe)-2OH sample matches well with the MIL-53(Fe) structure (CCDC No. 734218) in the Imma space group with the unite cell parameters of a = 17.84 Å, b = 6.87 Å, c = 11.84 Å and good agreement factors of weight-profile R-factor Rwp = 7.3% and underweighted R-factor Rp = 5.6% (Figure 2a). Moreover, the XRD pattern of the MOF-74-Fe sample conforms to the MOF-74 structure (CCDC No. 1494751, space group of R-3) with the unite cell parameters of a = 26.13 Å, b = 26.13 Å, c = 6.72 Å (Figure 2b). Simulated XRD patterns are consistent with the experimental profiles of MIL-53(Fe)-2OH and MOF-74-Fe, suggesting the successful formation of the MIL-53(Fe) and MOF-74 structure.23 The Raman spectra reveal the skeleton phonon modes of the carboxylate, phenolate, and benzene rings, and Fe–O in the obtained MOFs (Figure S4).24 The four vibrational peaks appeared at ca. 1606, 1397, 1535, and 1302 cm−1 for the MIL-53(Fe)-2OH are ascribed to the in- and out-of-phase stretching region of the carboxylate group, ν(COO−) vibration, and v(C−O) vibration, respectively.24-26 The peaks at 600 and 404 cm−1 are assigned to C–H stretching region of the benzene ring and Fe–O bonds vibration, respectively.24,26 By Raman spectroscopy analysis, although MIL-53(Fe)-2OH and MOF-74-Fe have similar coordinated structures, the vibration of hydroxyl group, C−O stretching vibration of the hydroxyl groups, carboxyl groups and Fe−O bond vibration are obviously different, indicating that the coordination environments are distinct in these two MOFs. Furthermore, Fourier transform infrared (FT-IR) spectroscopy was also employed to confirm the formation of MOFs (Figure S5). The 1211 cm−1 peak of MIL-53(Fe)-2OH is assigned to the C−O stretching vibration of the hydroxyl groups.27,28 The coordinated hydroxyl groups in MOF-74-Fe are difficult to form hydrogen bonds, resulting in the absorption peak moving to lower frequency.29 Therefore, the C−O stretching vibration peak of MOF-74-Fe shifts to 1192 cm−1. Importantly, the absorption peak at 3404 cm−1 for MIL-53(Fe)-2OH, which is absent for MOF-74-Fe, is related to the variation of phenolic hydroxyl group (νOH), suggesting the successful fabrication of the two target MOF structures.30,31 The XPS survey spectrum for MIL-53(Fe)-2OH reveals the presence of C (62.4 at.%), O (34.3 at.%), Fe (3.3 at.%) (Figure S6 and Table S1). High-resolution C 1s XPS spectra confirm the composition of the C=C/C−C in aromatic rings (284.7 eV), the C−O (286.4 eV), and the carboxylate group (O=C−O, 288.8 eV) for both the MIL-53(Fe)-2OH and MOF-74-Fe samples (Figure S7).8,32 The O 1s XPS spectrum of MOF-74-Fe can be deconvoluted into three energy peaks at 530.8 eV, 531.9 eV, and 533.7 eV, which are ascribed to the Fe−O bonds, the carboxylate of the organic ligands, and absorbed water, respectively (Figure S8).32 The O 1s XPS peak at 532.8 eV for MIL-53(Fe)-2OH is assigned to C−OH bonds. The fitting data showed that the content ratio of the Fe−O bond in MOF-74-Fe was slightly higher than that of MIL-53(Fe)-2OH, presumably related to the different coordination environments of the Fe ions with phenolic hydroxyl and carboxyl groups. The high-resolution XPS spectrum of MIL-53(Fe)-2OH at Fe 2p can be deconvoluted into four characteristic peaks shown in Figure S9, which are assigned to Fe3+ (712.4 eV and 725.6 eV) and associated shakeup satellites (716.6 eV and 730.3 eV), respectively.33,34
X-ray absorption spectroscopy (XAS) was performed to further elucidate the fine structure of the synthesized MOFs. As shown in the X-ray absorption near-edge spectra (XANES) (Figure 2c), the Fe K-edge position of MIL-53(Fe)-2OH and MOF-74-Fe are close to that of Fe2O3. According to the linear fitting of the K-edge energy position of XANES spectra, the oxidation states of Fe in MIL-53(Fe)-2OH and MOF-74-Fe are +3.3 and +2.8, respectively (Figure S10), reflecting the effect of phenolic coordination. The fitting results of the Fe K-edge EXAFS spectra (Table S3) exhibit a higher Fe–O coordination number of MIL-53(Fe)-2OH (6.67) than that of MOF-74-Fe (6.28) in the first shell, which is perhaps related to the higher oxidation state of Fe in MIL-53(Fe)-2OH. The lowest intensity of the pre-edge peak at ≈ 7114 eV appears in MIL-53(Fe)-2OH, indicating the excellent symmetric octahedra [FeO6] structure in MIL-53(Fe)-2OH, which is in good agreement with the simulated MIL-53(Fe) structure (CCDC: 734218) shown in Figure 2a and Figure S3. The pre-edge peak intensity of MOF-74-Fe is slightly higher, indicating the slightly lowered symmetry of the coordination sites. The coordination environment of Fe in MIL-53(Fe)-2OH is actually quite different from that of MOF-74-Fe.35,36 This effectively suggests that the different oxygenic coordination environment alters the local electron density of the metal sites and thus impact the catalytic activity.37 The Fourier transform of the extended XAFS (EXAFS) spectra at the Fe K-edge and the corresponding fitting results are summarized in Figure 2d and Figure S11-12. The main peak of Fourier transforms (FTs) curves for MIL-53(Fe)-2OH (1.59 Å) and MOF-74-Fe (1.64 Å) correspond to the nearest Fe−O coordination, with no indication of the formation of Fe−Fe bond.38 The signal of wavelet transforms (WT) for the k3-weighted Fe K-edge EXAFS curves related to Fe−Fe bonds has been detected in the Fe foil but not in MIL-53(Fe)-2OH and MOF-74-Fe, giving a strong indication that there are only atomically dispersed Fe sites in the MOF samples (Figure 2e). The maximum intensity signal at ~5.1 Å−1 for MIL-53(Fe)-2OH is related to the Fe–O bond, which exhibits a positive shift in contrast to that of MOF-74-Fe (~4.75 Å−1), indicating the different electronic structure between them. The Fe K-edge EXAFS data reveal a Fe–O bond length of 1.976 ± 0.04 Å and the Fe–O–C bond length of 3.179 ± 0.008 Å for MIL-53(Fe)-2OH, which show great consistency with the simulated structure based on the XRD analysis (Table S3 and Figure S13). Of note, the Fe–O bond length of MOF-74-Fe is slightly longer (1.996 ± 0.045 Å) than that of MIL-53(Fe)-2OH, and tallies with its lower oxidation state, which is also very much consonant with the XRD analysis (Figure S14). The different electronic structures of MIL-53(Fe)-2OH from MOF-74-Fe optimize the adsorption/desorption ability of oxygenic intermediates (OH*, OOH*, O*) and thus improves the catalytic OER activity.24,39
N2 adsorption-desorption isotherms reveal that the Brunauer-Emmett-Teller (BET) surface of MIL-53(Fe)-2OH is 7.3 m2 g−1, which is lower than that of the MOF-74-Fe (10.4 m2 g−1) (Figure S15). These two Fe-based MOFs catalysts exhibit an obvious adsorption hysteresis loop with typical IV-type isotherms, implying the mesoporous character. The measured pore sizes of MIL-53(Fe)-2OH and MOF-74-Fe are mainly concentrated at 1.48 nm and 1.59 nm, respectively, which are in good agreement with their crystal structure (Figure 3a,b). The morphologies of MIL-53(Fe)-2OH and MOF-74-Fe samples were examined by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As shown in Figure 3c and Figure S16-17, MIL-53(Fe)-2OH exhibits a regular octahedral structure. SEM image and corresponding energy dispersion spectrum (EDS) elemental mappings of MIL-53(Fe)-2OH are shown in Figure 3d. Apparently, the Fe, C, and O elements were distributed over the whole surface. The elemental mapping demonstrates the uniform distribution of Fe, C, and O elements in MIL-53(Fe)-2OH. In comparison, MOF-74-Fe exhibits polyhedral prism morphology with a width of 0.5-1 µm and a length of 10-25 µm (Figure 3e and Figure S18). The MOF-74-Fe prism self-assembles into micro-flowers. TEM images reveal that MOF-74-Fe is indeed composed of nanorods (Figure S19). As examined by SEM-EDS, the elemental mapping demonstrates the uniform distribution of Fe, C, and O elements in MOF-74-Fe (Figure 3f and Table S2). The electrocatalytic OER performances of MIL-53(Fe)-2OH and MOF-74-Fe samples were studied in a three-electrode system in the O2-saturated 1.0 M KOH solution at room temperature. According to the polarization curves shown in Figure 4a, the MIL-53(Fe)-2OH catalyst exhibits the lowest overpotential of 215 mV at a current density of 10 mA cm−2, compared to that of the MOF-74-Fe (242 mV), and IrO2 (335 mV) catalysts. The Tafel slope of the MIL-53(Fe)-2OH catalyst (45.4 mV dec–1) is also lower than that of MOF-74-Fe (49.5 mV dec–1), IrO2 (99.7 mV dec–1), and NF (108.2 mV dec–1) (Figure 4b, Figure S20 and Table S4). The electrochemical active surface area (ECSAs) of the electrocatalyst is proportional to the electrochemical double-layer capacitance (Cdl), which can be determined by measuring the scanning rate-dependent cyclic voltammetry (CV) of the non-Faraday region (Figures S21). The measured Cdl value of MIL-53(Fe)-2OH is 4.2 mF cm−2, which is closed to that of the MOF-74-Fe (4.1 mF cm−2) and larger than that of IrO2 (3.0 mF cm−2) (Figure S22).
As calculated from geometric current densities and ECSAs, the highest specific current density (jECSA) is achieved over MIL-53(Fe)-2OH (3.24 mA cm−2), which is much higher than that of MOF-74-Fe (1.42 mA cm−2) and IrO2 (0.08 mA cm−2) catalysts, indicating its superior intrinsic activity for OER (Figures S23). To further illustrate the electrode reaction kinetics during the catalytic OER process, electrochemical impedance spectroscopy (EIS) was performed. As shown in Figure 4c, the Nyquist plots reveal that MIL-53(Fe)-2OH has an ultra-low charge transfer resistance (Rct) of about 0.65 Ω at the overpotential of 300 mV (at 1.53 V vs. RHE) in alkaline condition, which is lower than that of the MOF-74-Fe (0.80 Ω) and IrO2 (18.32 Ω) catalysts, indicating a much faster charge transfer on the MIL-53(Fe)-2OH surface in the electrochemical reaction process. These results firmly demonstrated the excellent activity of MIL-53(Fe)-2OH for OER. In addition, the turnover frequency (TOF) and the mass activity (MA) of the catalysts were also determined to illustrate the intrinsic activity at the constant overpotential of 300 mV (Figure 4d).24,40 More details of the kinetic model have been presented in section 1.4 and 1.5 of the Supplementary Information. The high TOF of 1.44 s−1 is obtained for the MIL-53(Fe)-2OH catalyst, prominently exceeding that of MOF-74-Fe (0.59 s−1) and IrO2 (0.0177 s−1) (Figure S24). The MA value of MIL-53(Fe)-2OH could reach as high as 357.90 A g−1, significantly larger than those of MOF-74-Fe (153.60 A g−1) and IrO2 (62.34 A g−1) (Figure S25 and Table S5). Taken together, these results clearly indicate that the MIL-53(Fe)-2OH exhibits exceptional catalytic OER activities in terms of the lower Tafel slope and lower overpotential, which exceeds most of the recently reported MOFs OER catalysts, as shown in Figure 4e and Table S6.4,14,25,34,41-52
The OER performances of MOF-74-Fe and MIL-53(Fe)-2OH have further been elucidated by DFT calculations from both electronic structures and reaction energy. In MOF-74-Fe, the size of eg-t2g splitting increases with the coordination numbers (Figure 5a). The existence of OH vacancy also cannot evidently lower the eg-t2g splitting, demonstrating that electron transfer in Fe sites of MOF-74-Fe is lowered due to the large barriers. For different types of C sites within MOF-74-Fe, most carbon sites displayed low p-band centers, which demonstrated their limited contributions to the electron transfer during OER (Figure S26). For MIL-53(Fe)-2OH, the carbon sites in the benzene ring are more electroactive than that in O-C-O and the formation of OH vacancy further improves the electron transfer (Figure 5b). This indicates that the existence of the electroactive phenolic groups is able to not only facilitate the electron transfer but also improve the overall electroactivity. Since the electronic structures are highly correlated with the OER performances, the PDOSs of key intermediates on both MOFs are demonstrated (Figure 5c-d). For MOF-74-Fe, it is noted that the upshifting trend of the σ orbitals from O-species has shown an evident deviation at OOH*, which potentially increases the barriers for the conversion from O* to OOH*. Meanwhile, such a deviation is absent in MIL-53(Fe)-2OH, which achieves the efficient conversions of intermediates with low energy barriers. These results further confirm that the superior OER performances of MIL-53(Fe)-2OH originate from the optimal electronic structures modulated by the phenolic groups. The loss of OH groups in both MOFs shows high energy costs of 3.34 eV and 2.65 eV in MOF-74-Fe and MIL-53(Fe)-2OH, respectively (Figure S27). This illustrates that the involvement of OH groups from MOF to promote the OER is challenging. More importantly, this further reveals that the more electroactive Fe and O sites in MIL-53(Fe)-2OH are the key factors for the highly efficient OER process. The reaction energy change of OER has been compared between the two MOFs. With U = 0 V, both MOFs display the largest energy cost at the conversion from O* to OOH* as the rate-determining step (Figure S28). The energy barrier in MIL-53(Fe)-2OH is smaller than that of the MOF-74-Fe, which is consistent with the results of electronic structures. After introducing the equilibrium potential, we notice that the improved OER performances of MIL-53(Fe)-2OH are attributed to suitable binding of OH*, which results in a reduced energy barrier of 0.23 eV than that of the 0.32 eV in MOF-74-Fe (Figure 5e and Figure S29). On the other side, the over-binding of OH* leads to a continuous uphill trend in MOF-74-Fe, which has affected the efficiency of the OER process. The reaction mechanism of OER has been
demonstrated through the structural configurations (Figure S30-31). Notably, the frameworks of MOF are able to remain relatively stable during the adsorption of the intermediates. Moreover, the selectivity between 2e− and 4e− OER have been compared based on the conversion reaction of O*. For both MOF-74-Fe and MIL-53(Fe)-2OH, the 4e− OER is more preferred due to the much smaller energy barrier of O* → OOH* than O* → ½ O2, indicating the high selectivity towards 4e− OER.
The chronopotentiometry responses (E-t) curve in Figure 6a demonstrates that the as-synthesized MIL-53(Fe)-2OH and MOF-74-Fe catalysts possess good durability after the OER test for more than 100 hours at 100 mA cm−2. The polarization curves of the MIL-53(Fe)-2OH and MOF-74-Fe catalysts after the stability test have no obvious changes of the overpotential before and after the chronopotentiometry test at the high current density of 100 mA cm−2 (Figure S32). The overpotential of MIL-53(Fe)-2OH after 100 h of the stability test increases only 2 mV at the current density of 100 mA cm−2, which suggests that the MIL-53(Fe)-2OH exhibits excellent long-term stability (Figure 6b). The overpotential after the stability test increases 10 mV at the current density of 100 mA cm−2 for MOF-74-Fe catalyst (Figure S33-34). Cdl values of the MIL-53(Fe)-2OH (3.8 mF cm−2), MOF-74-Fe (3.9 mF cm−2) and IrO2 (2.9 mF cm−2) catalysts exhibit no obvious changes comparing to their initial ones (Figure S35-37). The overpotential at the current density of 100 mA cm−2 for MIL-53(Fe)-2OH after 100 hours of stability test (266.8 mV) only increases 0.6 mV compared to the initial value (266.2 mV), which is much lower than that of MOF-74-Fe (5.9 mV) and IrO2 (32.4 mV). Compared with MOF-74-Fe and IrO2, the Rct value of MIL-53(Fe)-2OH also show the minimum increase after the stability test (0.26 Ω) compared to MOF-74-Fe (0.80 Ω) and IrO2 (21.1 Ω) at the overpotential of 300 mV, confirming the superior electrochemical durability of the MIL-53(Fe)-2OH catalyst (Figure 6c and Figure S38-40)
Furthermore, XRD, XPS, TEM and SEM were performed to probe the essential difference in the catalysts’ structure during the OER process. As shown in Figure S41-42, the XRD patterns of MIL-53(Fe)-2OH and MOF-74-Fe after the OER stability test agree well with that of FeO(OH) (ICSD No. 94874), which demonstrates the formation of iron hydroxides.53 The structure and electronic properties of MIL-53(Fe)-2OH after stability test was further analyzed by XPS. The XPS analysis discloses the presence of Fe (4.3 at.%), C (36.5 at.%) and O (59.2 at.%) elements in MIL-53(Fe)-2OH catalysts after OER for 100 h, in which O content was significantly higher than that of the initial MIL-53(Fe)-2OH (Table S1). As shown in Figure S43a, the high-resolution XPS spectrum of Fe 2p shows deconvoluted peaks for Fe3+ (711.9 eV, 725.8 eV).27,54,55 The O 1s spectrum can be split into three peaks at 530.7, 532.8, and 534.1 eV, which are ascribed to the Fe−O, C−OH, and absorbed water, respectively (Figure S43b).56,57 These results suggest a structural transformation of the MIL-53(Fe)-2OH catalyst.26,27,58 MIL-53(Fe)-2OH catalyst (Figure S46). The HRTEM image of MIL-53(Fe)-2OH displays the lattice fringes with the interlayer distance of 0.29 nm, which correspond to the (220) planes of FeO(OH) (Figure S46c). It can be known from Figure S47 that the C, O, and Fe elements are spread over the whole region for the MIL-53(Fe)-2OH catalyst after for 100 h OER stability test at 100 mA cm−2. The morphology of MIL-53(Fe)-2OH and MOF-74-Fe after stability test was observed by SEM and TEM. As shown in Figure S44, the MIL-53(Fe)-2OH is transformed from the original regular octahedral structure to the morphology of bulk stacking. For MOF-74-Fe, the polyhedral prism morphology is transformed into sheets (Figure S45). TEM analysis further confirmed the phase transition of the
The structure-property dependency of the MIL-53(Fe)-2OH catalyst during the OER process was further investigated by the real-time kinetic simulation. Following the procedure of our previous work,59,60 the standard activation free energies for the five elementary reaction steps ( for the oxidative adsorption step, for the first oxidative transition step, for the associative desorption step, for the second oxidative transition step and for the oxidative desorption step) and the standard free energy for intermediate adsorption ( , for OH*, for O* and for OOH*) of OER process were evaluated by building the OER kinetic model and then best-fitting (R2 = 0.999) the kinetic current density of MIL-53(Fe)-2OH (Figure 6d-e). More details of the kinetic model are presented in section 1.7 of the Supplementary Information. It can be seen in Figure S48-49 that the free energy for the associative desorption step ( ) is much higher than that for the second oxidative transition step ( ) and the formation of OOH* ( ), indicating that the 4e− pathway dominates the OER process on the MIL-53(Fe)-2OH catalyst system, in good agreement with the DFT calculations. The free energy diagram shown in Figure 5f indicates the formation of the reaction intermediate OOH* from the adsorbed O* is the rate-determining step on the MIL-53(Fe)-2OH surface, which is in good agreement with the DFT calculations. Figure 6e shows the adsorption isotherms of the reaction intermediates. The fractional coverages of the intermediates, , and , are closely correlated to their standard free energy. The low accelerates the formation and adsorption of O*, leading to the high coverage of O* on the MIL-53(Fe)-2OH surface. In the higher overpotential range, the consumption rate of O* to form OOH* increases, resulting in the decrease of and the increase of the fractional coverage of OOH*. The strong adsorption of the intermediates on the MIL-53(Fe)-2OH surface promotes the generation of the FeO(OH) phase during the OER process. It is worth noting from Figure 6f that the activation energies for the rate-determining step are very close on the initial and post-tested MIL-53(Fe)-2OH catalyst, confirming the robustness of the catalyst with an invariant OER activity even after the long-term OER operational testing.
Although MIL-53(Fe)-2OH catalyst exhibits excellent OER stability within the testing time interval, the chemical structure of the MIL-53(Fe)-2OH catalyst has changed after the OER stability testing, so its kinetic behavior and intermediate coverages may also change after the OER stability test. Indeed, the OER kinetics of the post-tested MIL-53(Fe)-2OH catalyst is dependent on both the formation of OOH* and the adsorption of OH* with an increased energy barrier of 280 meV compared to the initial one (171 meV) (Figure 6f). Because of the higher energy barrier for the formation of the first intermediate (OH*), the fractional coverages of all the intermediates (OH*, O*, and OOH*) on the post-tested MIL-53(Fe)-2OH surface are lower than those of the initial counterpart, confirming a changed kinetic behavior of the MIL-53(Fe)-2OH catalyst during the OER process (Figure S50). Thus the real-time kinetic simulation opens a new direction in exploring the OER mechanism and can help to develop robust OER electrocatalysts with competitive activity and stability.