Here transition metal-LDHs catalysts are firstly taken as research models to interpret the great important role of CL on promoting OER catalytic performance (The surface morphologies of various transition metal-LDHs are shown in Figure S1). The OER catalytic activity of various transition metal-LDHs catalysts were tested with and without CLs as shown in Fig. 1a-h. Obviously, with CLs on the surface, CoV-LDHs, CoFe-LDHs, NiFeMn-LDHs, NiCu-LDHs, NiCo-LDHs, NiFe-LDHs, NiMn-LDHs and NiV- LDHs all show remarkably enhanced OER catalytic activity, and the overpotentials of above transition metal-LDHs without CLs drop evidently. The above result implies that the promotion effect of CLs for OER catalytic activity is universal for bi/trimetallic LDHs catalysts. In order to avoid decentralization, the analysis and discussion in this paper mainly focuses on NiFe-LDHs. Here the OER reaction kinetics were investigated by electrochemical impedance spectroscopy (EIS) and Tafel slop as shown in Fig. 2, which exhibits the obviously promoted OER kinetics by CLs. Specifically, compared with that of NiFe- LDHs (247.2 Ω of Rct), the NiFe-LDHs/CLs own a smaller charge transfer resistance (199.9 Ω of Rct) (Figue 2a), indicating that CLs can markedly facilitate the charge transfer during OER, further evidenced by the Tafel slope values (Figue 2b). Notably, when the overpotential is low, a small Tafel slope of 89 mV dec− 1 is obtained for NiFe-LDHs/CLs, and it is smaller than that of pure NiFe-LDHs (98 mV dec− 1). However, when the overpotential is high, NiFe-LDHs/CLs displays a much smaller Tafel slope (158 mV dec− 1) and much faster reaction dynamics than NiFe-LDHs (223 mV dec− 1), indicating that the CLs as a transfer mediator can facilitate the transfer rate of protons. This is consistent with previously reported works.17–18,23−24 This is also in line with the OER catalytic activity displayed in the polarization curve. The enhanced intrinsic activity of catalyst is also important for excellent OER performance. As shown in Fig. 2c, the anodic current density of NiFe-LDHs increased dramatically after extra addition of CLs in electrolyte, and NiFe-LDHs/CLs exhibited a low overpotential of 337 mV at a current density of 10 mA cm− 2, which is lower than that of NiFe-LDHs (357 mV), suggesting the remarkable catalytic activity promoted by CLs. Additionally, we explored the remarkable effect of soaking time in CLs-containing electrolyte on catalytic performance, that is, the influence of CLs loading on the catalyst surface on the enhanced OER catalytic performance (Fig. 2d). Obviously, the coordinated CLs can tailor the OER activity through different CL loadings adsorbed on catalysts. Also, the effects of different concentrations of CLs in electrolyte were investigated as shown in Fig. 2e. When the concentration of CLs is 0.10 M for 15 min, the best catalytic performance is achieved. In addition, the coordinated CLs on the surface of catalyst have been detected by Fourier transform infrared (FT-IR) spectroscopy (Fig. 2f). The peak at 1088 cm− 1 can be attributed to γ(OH) bending vibrations of −OH–OOC–Ar (Ar: benzene ring), indicating a proton transfer mediator via the form of hydrogen oxygen bond.25 Two peaks at 1576 and 1498 cm− 1 are clearly seen and are ascribed to νas(–COO– ) and νs(–COO–) of carboxylate,26 respectively. Notably, the peak at 1668 cm− 1 corresponds to the bending vibrations of –COO–,27 demonstrating the existence of CLs. Compared with the peak of metal-oxygen bonds of NiFe-LDHs, a strong and shifted peak appears at ~ 500 cm− 1 for NiFe-LDHs/CLs, indicating a coordination between Ni/Fe metal sites and CLs.28
The above results lead us to think about why and how do CLs promote OER electrocatalytic activity of NiFe-LDHs or other catalysts. Herein, a series of DFT calculations were conducted to further answer this question. Considering the adsorbed form of CLs on the interfaces of catalyst, there are three theoretical models were studied including physical level adsorption (PLA) model (denoted as NiFe-LDHs/CLs-PLA, Figure S2), physical vertical adsorption (PVA) model (denoted as NiFe-LDHs/CLs-PVA, Figure S3), and chemical vertical adsorption (CVA) model (denoted as NiFe-LDHs/CLs-CVA, Figure S4). In the first two models, the CLs are not directly bonded with the interfacial metal atoms and their states are similar, while in the third model, the CLs coordinate with metal atoms. For the convenience of discussion, here we only choose NiFe-LDHs/CLs-PVA and NiFe-LDHs/CLs-CVA models for the comparative studies. Firstly, the interfacial charge distribution in hybrid system was elucidated in above two models. From the differential charge density, Fig. 3a exhibits no interaction among electrons at the interfaces in NiFe-LDH/CL-PVA model, while NiFe-LDH/CLs-CVA model shows electron transfer from CLs to Ni atoms at near interface (Fig. 3b), and the adjacent Ni atom shows partial electron donation. This result points out that the CLs can modify the catalyst only in the case of bond formation. Impressively, the d band center model would be a good descriptor of the adsorbate–metal interaction. As shown in Fig. 3c and 3d, the d-band center (Ed) of NiFe-LDHs/CLs-CVA model is far away from the Fermi level from − 2.73 to -3.42 eV after the chemisorption of CLs, and the down-shift of d band center will increase the bonding states and decrease the anti-bonding states, and thus will weaken bonding strength of intermediates,29 and accordingly will enhance the catalytic performance of OER. But the Ed of NiFe-LDHs/CLs-PVA displays almost no little change. Furthermore, the Gibbs free energy (ΔG) of intermediate species, as a key descriptor of OER activity, has been demonstrated to be obviously decreased (Fig. 3e and Figure S5), including adsorption (step i), dissociation (steps ii and iii) and desorption (step iv). Interestingly, all the Gibbs free energy of OER intermediates for NiFe-LDHs/CLs are obviously lower than those of NiFe-LDHs and NiFe-LDHs/ CLs-PVA model. Specifically, lower ΔG value of step i of NiFe-LDHs/CLs-CVA model implies stronger adsorption of OH− compared with those of NiFe-LDHs and NiFe-LDHs/CLs-PVA model, benefiting for the initiation of OER. Importantly, the theoretical overpotentials of NiFe-LDHs/CLs-CVA model is only 370 mV, which is much lower than those of NiFe-LDHs (590 mV) and NiFe-LDHs/CLs-PVA model (600 mV). The above results demonstrated that the CLs can optimize adsorption strength of OER intermediates, and especially can lower reaction energy barriers from optimal Ed to promote the catalytic activity.
To gain in-depth insight into how the adsorbed CLs promote catalytic activity, the interfacial electronic behavior of NiFe-LDHs/CLs was also investigated via concrete experimental characterizations. X-ray photoelectron spectroscopy (XPS) and X-ray absorption spectroscopy (XAS) can be adopted to explore the synergistic electronic effect between CLs and in situ-formed metal (oxy)hydroxides on the surface of NiFe-LDHs. As shown in Fig. 4a, compared with that of the initial NiFe-LDHs, XPS peak of Ni 2p3/2 at 857.6 eV for post-OER NiFe-LDHs displays the oxidation of partial Ni2+ into Ni3+. More importantly, the peak of Ni 2p3/2 of post-OER NiFe-LDHs/CLs shifts to positive binding energy of 857.9 eV, suggesting that the partial electrons are transferred from a neighboring Ni to O/Fe, consistent with differential charge data analysis. In addition, the ratio of Ni3+ 2p3/2 /Ni2+ 2p3/2 of post-OER NiFe-LDHs/CLs jumped to a high value (0.79) compared with that of post-OER NiFe-LDHs (0.64), suggesting the enriched superficial oxyhydroxides transformation of metal sites during the electrochemical activation. The above results reveal the fact that the enriched Ni3+ on the surface of post-OER NiFe-LDHs/CLs was responsible for OER activity enhancement, consistent with previously reported works.30 In contrast, the fitted peak of Fe3+ 2p3/2 of post-OER NiFe-LDHs/CLs shifts to low binding energy of 712.7 eV compared with that of post-OER NiFe-LDHs (712.9 eV) and NiFe-LDHs (713.2 eV) (Fig. 4b). The above results mainly originate from carboxyl oxygen coordinated with Ni/Fe sites that tailors interfacial electron redistribution and charge state and thus induces downshift of metal d-band to penetrate p-band of oxygen intermediates (Figure S6),31 which is in agreement well with the result of DFT calculations. This phenomenon can be well elucidated in term of coordinated oxygen from ligand strengthening electron interactions between Ni and Fe (Fig. 4c). The t2g d-orbitals of Ni2+ in the low-spin state are fully occupied, resulting in e− − e− repulsion between the bridging O2− and Ni2+.32 For Fe3+, 3d5 valence electron configuration with unpaired electrons in the π-symmetry (t2g) d-orbitals implies a weak interaction via π-donation.33 Thus, to keep the balance, the electron interactions between Fe3+ and Ni2+ can well be strengthened via bridging O2− after more oxygen coordination. As such, the partial charge transfer from Ni2+ to Fe3+ will be easily occurred, and this has been proven by the shift of binding energy of Ni2+ 2p3/2 of post-OER NiFe-LDHs/ CLs. As a consequence, there are more empty orbitals at Ni 3d eg energy level and more electrons at Fe 3d t2g energy level after the introduction of local ligand environment.34 Furthermore, the normalized X-ray absorption spectra was utilized to study the fine structure of NiFe-LDHs, post-OER NiFe-LDHs and post-OER NiFe-LDHs/CLs as shown Fig. 4d-g. Specifically, an obvious shift of Ni K-edge XANES adsorption-edge to a higher energy is seen for post-OER NiFe-LDHs/CLs compared with those of NiFe- LDHs and post-OER NiFe-LDHs, indicative of increased charge state of Ni (Fig. 4d), consistent with XPS analysis results. In addition, the mild decline of Fe valence state after CLs adsorption can also be verified by the shift of Fe adsorption-edge (Fig. 4e). Moreover, the Fourier transform (FT) of extended X-ray absorption fine structure (EXAFS) spectra at Ni K-edge implies the dominant peak at 1.6 Å is ascribed to Ni/Fe–O bond, and the relatively weak peaks at 2.7 Å is associated with Ni–Ni(Fe) bonds (Fig. 4f). Interestingly, we note that the Ni–O bond length slightly becomes shorter by 0.2 Å, and simultaneously, the Ni–Ni/Fe bond length becomes longer, suggesting the presence of lattice strain under CLs coordination.14,34 The result further demonstrates that the CLs can regulate the interatomic distances and induce the partial distorted structure, and meanwhile alter the electronic structures of active sites that could help to optimize the intrinsic activity.35–36 Similarly, in Fig. 4g, FT-EXAFS curves at Fe K-edge shows the similar shifts of Fe-O and Fe-Ni/Fe bond lengths. Such above peak shifts could be further understood by a slight local distortion of the octahedral coordination of Fe/Ni,37 which is closely related to eg/tg orbits of the active sites that strongly affect the adsorption strength of intermediates.38
Inspired by the high catalytic performance promoted by the additional CLs on the surface of catalysts, we presume that if the catalyst itself contains CLs, what will the promoting effect be like? Typically, metal-organic framework material with CLs as OER catalyst were systemically explored, because most of studies mainly focused on the synergistic effect between metals or lower coordinated metals and so on, always ignoring the effect of local ligand environments on the catalytic performance improvement. With this thinking, NiFeMn trimetal-organic frameworks (NiFeMn-MOFs) was chosen as our research model, because CLs in NiFeMn-MOFs are relatively stable, which is critical key for unveiling the roles of CLs at the atomic level on OER activity. Corresponding evidences for confirming stability of CLs in hybrid system, including the crystal structure, morphology, surface electronic feature, and functional groups of NiFeMn-MOFs before and after cycling test have been analyzed as shown in Figure S7-12. The OER activity of NiFeMn-MOFs was also evaluated in a conventional three-electrode cell containing 1.0 M KOH solution at a scan rate of 5 mV s− 1. For comparison study, the different contents of CLs in post-OER NiFeMn-MOFs-x min toward OER activity were investigated (x = 5, 10 and 15, x represents the soaking time for post-OER NiFeMn-MOFs in 6.0 M KOH). The concentration of CL increases in electrolyte with soaking time increasing, evidenced by the UV-vis spectroscopy in Fig. 5a. This result means that the residual CLs in the frameworks will decrease through metal cations transforming into (oxy) hydroxides (Figure S10f).39 Strikingly, NiFeMn-MOFs exhibit a high catalytic activity with a low overpotential of 253 mV at the current density of 10 mA cm− 2 (Fig. 5b), which is 9, 17 and 18 mV less than those of post-OER NiFeMn-MOFs-5 min, post-OER NiFeMn-MOFs-10 min, post-OER NiFeMn-MOFs-15 min, respectively, suggesting that the CLs can modulate catalytic activity of OER. Impressively, NiFeMn-MOFs demonstrated a remarkably fast reaction kinetics with a low Tafel slop value of 41 mV dec− 1 (Fig. 5c), which is 2.7, 5.0 and 5.2 times lower than those of post-OER NiFeMn-MOFs-5 min, post-OER NiFeMn- MOFs-10 min, post-OER NiFeMn-MOFs-15 min, respectively. In addition, the electrochemical impedance spectroscopy (EIS) was performed to explore interfacial reactions and OER kinetics (Figure S13). Notably, an ultrasmall semicircle at low-frequency region associated with the mass diffusion process for NiFeMn- MOF compared with other samples can be observed, demonstrating a stronger adsorption/desorption of reaction intermediates (such as OH−) on the surface of catalyst.40 Moreover, the value of charge transfer resistance (Rct) for NiFeMn-MOFs (0.76 Ω) is smaller than those of post-OER NiFeMn-MOFs-5/10/15 min samples (0.81, 1.03 and 1.06 Ω, respectively), indicating fast electron transport migration between the solution and electrode interface.
To insight into how CLs promote reaction kinetics, OH activation/dissociation that strongly determine fast kinetics were investigated via in situ Raman, FT-IR spectroscope, pH-dependent and deuterium kinetic isotope effects (KIEs). It is well known that Lewis bases (SO42−/PO43−/COO−/CO32−/SeO42−) in solution are a proton acceptor,41 which can be utilized as a transfer station for protons to facilitate the proton transmission rate, evenly perform the functions of OH activation and dissociation. To verify above hypothesis and explain why NiFeMn-MOF possesses a remarkably superior reaction kinetics, in situ Raman measurement was firstly carried out. As shown in Fig. 6a, potential-dependent in situ Raman spectroscopy displays the structure evolution during OER. With potential increasing, the strong signals of C-COOH (308 cm− 1),42 COO− (1402 cm− 1),43 COOH (1683 cm− 1),23, 42 and free OH− (924 cm− 1)44 can be detected, nevertheless in ex situ spectrum, all above signals of various groups disappear (Figure S14), indicating the real process of proton transfer mediator via the form of −COO−HO−, and subsequent adsorption/activation/dissociation of OH−, consistent with previously reported works.23 Specifically, as schematically illustrated in Fig. 6b, the free OH− from bulk electrolyte firstly react with interfacial coordinatively unsaturated Ar-COO− in form of Ar-COO–HO−, accelerating the rate of proton transfer (Step 1: proton transfer station). Secondly, the O atom from Ar-COO–HO− transfers to metal site to form Ar-COO–HO* (* denotes as the active site), and followed by the activation and dissociation (Step 2), namely, another OH− from electrolyte will react with H of Ar-COO–HO* to produce adsorbed O* intermediate. Meanwhile the proton transfer continues to form Ar-COO–HO− (Step 2). Thirdly, another OH− from Ar-COO–HO− will bond with O* intermediate to form Ar-COO–HOO* under applying potential (Step 3). Similarly, under the role of activation and dissociation of carboxylic group, the Ar-COO–HOO* will be deprotonated via OH−/COO coupled reaction to form Ar-COO–H(OO*)−OH, and subsequently produce oxygen molecular and water (Steps 4 and 5). The above process of transfer/ activation/dissociation of hydroxyls realizes remarkable catalytic activity of NiFeMn-MOF. Additionally, the step of proton transfer was further confirmed by in situ FT-IR spectroscope (Fig. 6c). Compared with that of initial NiFeMn-MOFs, two peaks at 1642 and 1057 cm− 1 in Fig. 6c can be attributed to carboxyl group23,27,43,45,47 and hydroxyl group25, respectively, appeared during OER, and become stronger with potential increasing, indicating the uncoordinated carboxylate provided the extra hydrogen site as a proton transfer bridge. Furthermore, the study of pH dependence can provide useful insight on the kinetics and intermediates of the reaction. As shown in inset of Fig. 6d and 6e, the pH-dependent value range is 0.6 for NiFeMn-MOFs, and the slop of change is ~ 2.5 times lower than that of NiFeMn-LDHs, suggesting negligible pH-dependent OER kinetics for NiFeMn-MOFs and higher intrinsic activity with help proton transfer mediator. Moreover, it is known that deuterium KIEs can reflect proton transfer and dissociation, and thus will help to verify above described catalytic processes. As shown in Fig. 6f and 6g, the existence of KIEs demonstrated proton transfer involved in OH−/OD− dissociation, which affects the OER reaction rate. Obviously, the LSV curve of NiFeMn-MOFs in a 1.0 M NaOH/H2O solution exhibits much higher current density compared with NiFeMn-MOFs in 1.0 M NaOD/H2O solution by a factor of 6 at the initial stage, but quickly decreases with the potential increasing, suggesting fast cleavage of O − H bonds. However, the KIEs value of NiFeMn-LDHs shows a larger primary isotope effects in the absence of CLs, strongly proving that superior kinetics derived from the CLs that can help handling the proton transfer and dissociation on the surface of catalyst.