Electronic optimization and modification of efficient Ir clusters embedded onto Ni–Mo–P for electrocatalytic oxygen evolution reaction

The implementation of low-cost and efficient electrocatalysts for water oxidation is crucial for the development of industrial water electrolysis; however, they often suffer from inferior activity or poor stability. Herein, we demonstrated 0.7-nm iridium clusters embedded onto Ni–Mo–P (Ir-NMP) that exhibited an ultralow overpotential of 290 mV vs reversible hydrogen electrode (RHE) to reach the current density of 50 mA cm−2 in 1 M KOH, together with low Tafel slope of 63.9 mV dec−1, high mass activity of 2604 A gIr−1, and excellent catalytic stability with almost complete retention of activity within more than 30 h. According to characterizations and analyses, the interlaced crystalline and amorphous structure of Ni − Mo − P made homogeneous embedment of ultrathin Ir clusters onto NMP substrate, and the introduction of Ir clusters enabled the optimization of electronic structure for Ni and Mo species in NMP, which were available to the highly efficient and durable Ir-NMP electrocatalyst for OER.


Introduction
Electrochemical water splitting has been identified as one of the most important and promising approaches for hydrogen production, which includes hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) [1][2][3][4]. Regarding the fact of 4e − process, OER possesses sluggish reaction kinetics and thus becomes the main difficulty in electrochemical water splitting [5][6][7][8]. Due to their outstanding performances and active catalytic mechanisms of OER, Ru-and Ir-based materials are highly desirable electrocatalysts [9,10]; however, low abundance and high cost significantly hinder their applications [11][12][13]. As attractive alternatives to precious metal-based materials, transition metal compounds (TMCs) have attracted tremendous interest [14][15][16][17][18], and have been considered as catalysts for electrochemical oxygen evolution owing to their active sites of metal-O (M-O), M-S, and M-P bonds [19][20][21][22]. For example, iron-encapsulated CNTs on carbon fiber showed boosted oxygen evolution reaction performances [23]. However, OER over TMCs suffers from inferior electrochemical performances and poor stability, attributed to unstable structure and inactive intrinsic activity. Although exciting progress in the development of TMCs has been made, it is still challenging to achieve high activity and excellent durability, and determine the structure-activity relationship due to extremely easy structure reconstruction during oxygen evolution reaction [24,25]. Accordingly, an efficient and stable electrocatalyst that can simultaneously preserve low cost and high mass activity for OER is highly desired and significant for further exploration of electrochemical water splitting and practical applications of hydrogen energy.
Toward the goal of an efficient and low-cost electrocatalyst, a paramount issue is to enhance the utilized efficiency of metal atoms in precious metal-based materials [26][27][28]. Reducing the size of metal nanocrystals to sub-nanometer metal clusters or even to isolated metal atoms is expected to be efficient strategy [29,30], which synthesizes desired catalysts to achieve the maximum atomic efficiency for electrolysis [13,31,32]. Furthermore, the low coordination and unsaturated configuration in sub-nanometer metal clusters and single-atom catalysts result in increased number of active sites for catalysis compared with nanocrystals [33][34][35]. Highly dispersed isolated metal atoms and sub-nanometer metals clusters have attracted tremendous research interest in the field of catalysis [36,37], since they demonstrate untraditional activity and different selectivity in enormous catalytic reactions [38][39][40][41] due to their unique atomic structure and electronic property in comparison with bulk or nanosized catalysts. For instance, anchored Ir single atoms onto oxygen vacancies on a defective CoOOH surface (Ir 1 /VO-CoOOH) were beneficial to stabilize intermediates and lower the energy barrier of OER owing to hydrogen bonding [42]. The introduction of strong Brønsted acid sites (e.g., tungsten oxides, WO x ) with an isolated atom structure into RuO 2 was available for accelerating the deprotonation of OER intermediates [43]. However, a rational design of sub-nanometer metal clusters and single-atom catalysts for electrochemical oxygen evolution together with efficient property and excellent durability has been often ignored. More importantly, dispersing precious metal atoms into subnanometer size and even into isolated single atoms embedded in supports with well-designed atomical and electronic structure can serve as great models to study complex oxygen evolution reaction pathways and design economical and environmentally friendly electrocatalysts in sustainable hydrogen production [41,44].
Inspired by the electrochemical stability of precious metal catalysts and the good activity of transition metal compounds, herein, we reported that Ir clusters onto NMP were realized by successfully combining active Ir species and non-precious transition-metal phosphides. A highly efficient and durable Ir-NMP electrocatalyst was synthesized through a simple combination of hydrothermal method and two-step annealing process, where Ir clusters with the average size of 0.7 nm were homogeneously embedded onto the surface of Ni-Mo-P substrates with one-dimensional nanorod structure composed of two-dimensional nanosheet subunits. The as-synthesized Ir-NMP electrocatalyst delivered an ultralow overpotential of 340 mV vs RHE at 100 mA cm −2 , a low Tafel slope of 63.9 mV dec −1 , a high mass activity of 2604 A g Ir −1 , and large anodic currents in alkaline media, together with excellent catalytic stability with almost complete retention of activity within more than 30 h. Impressively, the interlaced crystalline and amorphous structure of Ni-Mo-P made the Ir clusters homogeneously embedded onto NMP substrate, and the introduction of Ir clusters enabled the optimization of electronic structure and chemical states for Ni and Mo species in NMP, which greatly accelerated the four-electron reaction kinetics toward efficient activity and excellent durability for alkaline OER. This work not only fabricates a promising cost-effective OER electrocatalyst with superior OER performances but also provides an efficient strategy for syntheses of sub-nanometer cluster materials for industrial applications.

Syntheses of catalysts
Briefly, an optimized nickel-molybdenum precursor on the surface of carbon paper was firstly realized via a hydrothermal method at 120 ℃ for 6 h with the presence of nickel chloride, ammonium molybdate, and urea. Firstly, Ni-Mo precursors were synthesized through the hydrothermal method. In the typical procedure, 5 mmol nickel chloride (NiCl 2 ·6H 2 O) and 1 mmol ammonium molybdate ((NH 4 ) 6 Mo 7 O 24 ·4H 2 O) were firstly completely dissolved in 60 mL deionized water, and then 22.5 mmol urea was added into the mixed solution. After continuously stirring for 1 h, a uniform green solution was obtained. Simultaneously, a piece of carbon paper (CP, 2 × 2 cm 2 ) was immersed in a 0.1-mol L −1 diluted hydrochloric acid (HCl) solution for 24 h and cleaned with deionized water and ethanol to completely remove residual HCl. The pre-treated carbon paper and the above uniform green solution were put into a 100-mL Teflon-lined stainless-steel autoclave and annealed at 120 ℃ for 6 h in a constant-temperature oven. After finishing the hydrothermal method, Ni-Mo precursors supported on the surface of carbon paper (Ni-Mo/CP) were cleaned using deionized water and ethanol and dried overnight in the furnace at 60 ℃. The obtained samples were Ni-Mo/ CP. Secondly, the Ni-Mo-P supported on the surface of carbon paper (NMP/CP) was prepared through the simple annealing process. The above synthesized Ni-Mo/CP and 1 g sodium hypophosphite (NaH 2 PO 2 ) powder were separately added into two crucibles. In order to obtain effective phosphating treatment, the crucible with NaH 2 PO 2 was positioned in the upstream side, and the distance between two crucibles was 5 cm. Subsequently, the phosphorization reaction was performed at 550 ℃ for 1 h under N 2 flow, and then the Ni-Mo-P supported on the surface of carbon paper was successfully synthesized. Finally, the Ir clusters embedded on Ni-Mo-P were synthesized through further annealing treatment with the presence iridium precursors. The above resultant Ni-Mo-P was immersed into 5 mmol L −1 (mM) hexachloroiridium acid hydrate (H 2 IrCl 6 ·6H 2 O) for 30 min, and then the obtained samples were taken out and dried, followed by the annealing process at 300 ℃ for 1 h in N 2 atmosphere. After annealing treatment, the Ir-NMP supported on the surface of carbon paper was cleaned using deionized water and ethanol and dried overnight in the furnace at 60 ℃; thus, Ir clusters embedded onto Ni-Mo-P were synthesized, which was also denoted as Ir-NMP or Ir-NMP-5 mM. For comparison, different concentrations of Ir precursors (1 and 10 mmol L −1 ) were investigated with other reaction process; thus, we synthesized the Ir-NMP-1 mM and Ir-NMP-10 mM catalysts.

Characterizations
The morphologies of the above samples were assessed using a scanning electron microscope (SEM, TESCAN), high-resolution transmission electron microscopy (TEM, Talos F200X), scanning transmission electron microscopy (STEM, F200X), and aberration-corrected annular dark field scanning transmission microscopy (AC-HAADF-STEM, Themis Z). High-angle annular dark field-STEM images were recorded using a convergence semi angle of 11 mrad and inner and outer collection angles of 59 and 200 mrad, respectively. Energy dispersive X-ray spectroscopy (EDS) was performed using 4 in-column Super-X detectors. The crystalline structures of these samples were examined using X-ray diffractometer (XRD, D/max 2500 V), together with Cu-Kα radiation and operations of 40 kV and 150 mA. The Ir loading of the Ir-NMP catalyst was identified using an inductively coupled plasma optical emission spectrometer (ICP-OES, ELAN DRC-e). The electronic structure and chemical states of the Ir-NMP catalyst were collected using X-ray photoelectron spectrometer (XPS, Escalab 250Xi) equipped with a monochromatic Al Kα X-ray source. The obtained XPS spectra were fitted using the XPSPEAK41 software and corrected using the C 1 s peak (284.8 eV).

Electrochemical test
The electrocatalytic performances of the samples were performed in 1 M KOH aqueous solution using the electrochemical station (CHI760E) with a three-electrode system where our synthesized samples were directly utilized as working electrode, while the carbon rod and the HgO/Hg electrode were employed as counter electrode and reference electrode, respectively. Linear sweep voltammetry (LSV) curves were collected at 5 mVs −1 scan rate and then were corrected with 85% resistance compensation owing to iR drop. Electrochemical impedance spectroscopy (EIS) was performed in the frequency range of 10 kHz to 0.1 Hz at the open circuit voltage. Cyclic voltammetry (CV) sweeps were also performed at the voltage from 1.35 to 1.65 vs RHE with a scan rate of 200 mV s −1 to evaluate the CV durability of the Ir-NMP catalyst and commercial IrO 2 . The long-term stability of the Ir-NMP catalyst was also measured through continuous chronoamperometry operation at a potential of 1.4 V vs RHE, and the commercial IrO 2 was operated a the potential of 1.73 V vs RHE to detect its OER stability.

Morphological and electronic characterizations
Iridium clusters embedded on Ni-Mo-P were successfully synthesized through a simple combination of hydrothermal method and two-step annealing process. Briefly, an optimized nickel-molybdenum precursor on the surface of carbon paper was firstly realized via a hydrothermal method at 120 ℃ for 6 h with the presence of nickel chloride, ammonium molybdate, and urea. The scanning electron microscope (SEM) images showed that the nickel-molybdenum precursors on the surface of carbon paper were demonstrated as one-dimensional nanorod structure composed of two-dimensional nanosheets (Fig. S1). After annealing treatment with the presence of P precursor at 550 ℃, the NMP substrate was formed. The SEM images of pure NMP revealed a similar nanosheet-nanorod structure in comparison with that of nickel-molybdenum precursors (Fig. S2). The as-prepared NMP was regarded as the substrate and immersed in a hexachloroirdiium acid hydrate solution, followed by dried process and annealing treatment in nitrogen atmosphere at 300 ℃ to obtain the Ir-NMP catalyst. As shown in Fig. 1, the morphology of the Ir-NMP catalyst retained the same structure of NMP, together with the existence of hierarchical nanorod structure with nanosheet subunits.
The morphology of Ir-NMP was further confirmed by transmission electron microscopy images and high-angle annular dark field-scanning transmission electron microscopy images in Fig. 2a and b, which also showed nanosheet structures of the Ir-NMP catalyst. Figure 2a also reveals the selected area electron diffraction (SAED) patterns of the aspapered Ir-NMP catalyst. Diffraction spots that corresponded to Ni 2 P and NiMoP 2 were observed, further demonstrating the missing nanocrystals of Ir species in the Ir-NMP catalyst. Moreover, high-resolution (HR) TEM images revealed that nanoparticles with an average size of 10 nm were shown on the surface of nanosheet subunits (Fig. 2c), and the clear lattice spacing on ultrathin nanosheets was 0.200 and 0.221 nm and assigned to the (104) crystal plane of NiMoP 2 and the (111) plane of Ni 2 P, respectively (Fig. 2d), consistent with SAED patterns. The lattice fringes assigned to Ir or IrO 2 nanocrystals were also absent in HRTEM images of the Ir-NMP catalyst. Energy-dispersive spectroscopy (EDS) mapping images showed the existence of Ir, Ni, Mo, P, and O elements throughout the whole Ir-NMP catalyst (Fig. S3). The uniform and homogeneous distribution of Ir and P elements overall Ir-NMP verified the successful phosphorization and introduction of Ir species. The ratios of above Ni, Mo, and P elements were 15.5, 6.2, and 31.6 atom%, further indicating the successful synthesis of Ni-Mo-P (Fig. S4). The content of Ir was revealed to be 1.93 wt.%, determined by ICP measurements and agreement with EDS results. Considering HRTEM images and EDS results, it was found that amorphous IrMo x P y was observed in the Ni-Mo-P nanosheets, corresponding to the side of crystalline region. Hence, it was concluded that Ir species were embedded on the crystalline/ amorphous Ni-Mo-P with hierarchical nanorod structure composed of nanosheets subunit, which was successfully In order to distinguish the Ir species in the Ir-NMP catalyst, aberration-corrected transmission electron microscopy was performed. As shown in Figs. 3a to c, highly dispersed iridium species in the Ir-NMP catalyst could be directly observed by atomic-resolution HADDF-STEM images, which was composed of a mixture of major Ir clusters and little atomically dispersed Ir, without the existence of iridium nanoparticles. In Fig. 3d, iridium clusters are clearly observed as bright aggregates labeled with blue circles, accompanied with a size distribution of approximately 0.7 nm for these clusters (Fig. S5). In addition to the observation of Ir clusters, the homogeneous dispersion of Ir, Ni, Mo, P, and O elements in the Ir-NMP catalyst was also confirmed by the atomic resolution EDS elements mapping images (Fig. 4), further unambiguously demonstrating the uniformly introduction of Ir clusters into Ni-Mo-P substrates with nanosheet structure. The uniformly anchored Ir clusters and isolated Ir atoms onto NMP could simultaneously enhance the metal efficiency and increase active sites that are favorable for electrocatalytic efficient OER in a cost-effective route [45][46][47].
To further characterize the crystalline phase and electronic structure of the Ir-NMP catalyst, X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) were conducted. As exhibited in Fig. S6, the obvious diffraction peaks in pure NMP and the Ir-NMP catalyst were corresponded to the crystalline carbon paper (PDF#26-1080), owing to the strong intensity of crystalline CP and a small content of NMP or the Ir-NMP catalyst grown on the surface of the carbon paper. The Ir-NMP powders were synthesized through the same combination of hydrothermal method and two-step annealing process without the utilization of carbon paper, which was in order to further verify the crystalline structure of the Ir-NMP catalyst. Impressively, in Fig. 5a, there are weak diffraction peaks appeared at 40.6 and 45.2°, which could be indexed to the characteristic (111) planes of Ni 2 P (PDF#03-0953) and (104) planes of NiMoP 2 (PDF#30-0862). Moreover, it was mentioned that no additional diffraction signals of Ir-based species were existing in the Ir-NMP catalyst. These results were consistent with the above STEM images, further confirming the embedment of Ir clusters on the surface of crystalline/amorphous Ni-Mo-P. In the Ir 4f XPS spectra (Fig. 5b), the Ir-NMP catalyst showed two sets of doublets where the fitted pair appeared at 62.6 and 65.9 eV assigned to Ir 4+ 4f 7/2 and Ir 4+ 4f 5/2 [31], while another fitted couple centered at 67.1 and 68.7 eV was attributed to Ni 3p [31,32]. Accordingly, the majority of Ir clusters embedded onto the Ni-Mo-P substrate was approximately + 4 oxidation state, indicating that Ir clusters were highly oxidated in the Ir-NMP catalyst and those were available for the enhancement of electrocatalytic water oxidation [48,49].
Furthermore, the implementation of Ir clusters onto the Ni-Mo-P surface could offer a significant influence of the electronic structure of the Ni-Mo-P matrix, and the optimized NMP could probably accelerate OER process. As revealed in Fig. 6a, the XPS spectra of Ni 2p in the NMP were identified by the two characteristic fitted peaks at 856.5 and 861.3 eV, which were assigned to Ni 2+ 2p 3/2 and corresponding satellite signal, respectively [50,51]. When we introduced Ir clusters into pure NMP, the obtained Ir-NMP catalyst also demonstrated two fitted  peaks at 857.1 and 861.3 eV, whose binding energy assigned to Ni 2+ 2p 3/2 exhibited a positive shift of approximately 0.6 eV compared with that of pure NMP (Fig. 6b). Such a positive shift in the Ir-NMP catalyst indicated an increased valence state of Ni species on the surface of the Ir-NMP catalyst, which probably enhanced electrocatalytic water oxidation [52]. Furthermore, the three fitted XPS peaks positioned at 230.7, 232.7, and 235.8 eV in the spectra of Mo 3d for pure NMP could be corresponding to Mo δ+ (0 < δ < 4), Mo 4+ , and Mo 6+ for Mo 3d 3/2 , respectively (Fig. 6c). Considering the binding energies of Mo species, Mo δ+ was related with the formation of Mo-P bond, and Mo 4+ and Mo 6+ were attributed to the existing Mo-O bonds on the surface of Ni-Mo-P [53,54]. In comparison with pure NMP, the XPS spectra of Mo 3d for the Ir-NMP catalyst showed a distinct negative shift in binding energy (Fig. 6d) and increased content of Mo δ+ from 5.8 to 7.9%, demonstrating the Mo species in the Ir-NMP catalyst were obviously modified compared with those of pure Ni-Mo-P. Combining with the above Ni 2p XPS spectra, it indicated that the electrons of Mo species were transferred to the neighboring Ni elements owing to the presence of Ir clusters, which resulted in modifications of the electronic structure for Ni 2p and Mo 3d in the Ir-NMP catalyst. The electron acceptor of Ni centers was favorable to weaken the adsorption barrier of OH*; meanwhile, the electron-deficient Mo centers were beneficial to improve the desorption of OOH* and O 2 , which were indicative of the enhancement of OER performances over the Ir-NMP catalyst [53,55]. As shown in Fig. 6e, the deconvoluted P 2p spectra of pure NMP revealed the two fitted peaks resulted from two distinct phosphorus environments of 2p 1/2 and 2p 3/2 [55]. For the Ir-NMP catalyst in Fig. 6f, the P 2p spectra at 134.6 and 133.7 eV showed a positive shift of binding energies compared with those of pure NMP, which indicated that the electronic interactions between the P and metal species may lead to charge redistribution in the Ir-NMP catalyst [55]. Overall, it was concluded that the electronic structure and chemical states of NMP were efficiently optimized when we introduced Ir clusters into NMP, and the obtained Ir-NMP catalyst demonstrated the generation of Ir 4+ , increased valence state of Ni, and electron-deficient Mo/P species, which were possibly conducive to improvement of electrocatalytic water oxidation.

Electrochemical performances
The electrochemical catalytic OER performances of the Ir-NMP catalyst were evaluated in 1 M KOH aqueous solution using a three-electrode system, together with pure NMP and commercial IrO 2 for comparisons. Figure 7a displays the Fig. 6 Chemical states of the Ir/NMP and pure NMP catalysts. a Ni 2p XPS spectra of pure NMP. b Ni 2p XPS spectra of the Ir-NMP catalyst. c Mo 3d XPS spectra of pure NMP. d Mo 3d XPS spectra of the Ir-NMP catalyst. e P 2p XPS spectra of pure NMP. f P 2p XPS spectra of the Ir-NMP catalyst linear sweep voltammetry (LSV) curves for the Ir-NMP catalyst and reference samples. The Ir-NMP catalyst showed sharp start to catalyze the OER at an overpotential of 147 mV vs RHE and required an ultralow overpotential of 400 mV vs RHE to achieve a current density of 200 mA cm −2 for electrocatalytic water oxidation, along with rapidly increased current density. In contrast, pure NMP and commercial IrO 2 catalysts displayed insufficient OER activity, which could not obtain the current density of 50 mA cm −2 with an applied potential of 1.7 V vs RHE, indicating their inert intrinsic activities for OER. It was obvious that the activity of electrocatalytic water oxidation for NMP was extremely enhanced after Ir incorporation, surpassing those of commercial IrO 2 and pure NMP with respect to both onset potential and current density. Notably, it was observed that an oxidation peak occurred at a potential of around 1.4 V vs RHE, which was owing to the oxidation reaction of Ni species during electrocatalytic OER process and related with the redox potential of Ni 2+ and Ni 3+ [56,57]. Moreover, Fig. 7b shows that the Ir-NMP catalyst delivered a current density of 50 mA cm −2 at an overpotential of 290 mV vs RHE, outperformed than those of pure NMP and commercial IrO 2 (530 mV vs RHE). The required overpotential for commercial IrO 2 at a current density of 100 mA cm −2 was around 1.7-fold of that of the Ir-NMP catalyst, further confirming the superiority of the Ir-NMP catalyst for electrocatalytic OER. Additionally, to further analyze the influence of Ir content in the Ir-NMP catalyst on electrocatalytic water oxidation, different concentrations of Ir precursors were investigated (Fig. S7). Ir-NMP catalysts were synthesized through Ni-Mo-P immersing into hexachloroiridium acid hydrate solution with different concentrations from 1 to 10 mM, which was denoted as Ir-NMP-1 mM, Ir-NMP, and Ir-NMP-10 mM, respectively. It was observed that the Ir-NMP catalyst with 1 mM exhibited poor OER activity, which indicated that 1 mM Ir precursors were not conducive to obtain abundant metal active sites for OER in the Ir-NMP catalyst. With higher concentration up to 5 and 10 mM, the obtained Ir-NMP catalysts exhibited boosted OER activity compared with that of 1 mM, and Ir-NMP-5 mM possessed the best activity for water oxidation. It was probably because more metal active sites were formed when 10-mM precursor was applied; however, it also resulted in agglomeration of Ir clusters.
Tafel plots derived from the corresponding polarization curves were further investigated to determineelectrocatalytic kinetics of OER, where a smaller value of Tafel slope was preferred for OER electrocatalysts [58][59][60]. Figure 7c shows a smaller Tafel slope of the Ir-NMP catalyst (63.9 mV dec −1 ) compared with that of commercial IrO 2 (102.3 mV dec −1 ), indicating faster OER kinetics and electron transfers over the Ir-NMP catalyst. In addition, the OER activities of the catalysts were also examined by mass activities, which were calculated according to Ir loading. At an overpotential of 290 mV vs RHE, the Ir-NMP catalyst exhibited a mass activity of 1304 A g Ir −1 , which was significantly higher than that of commercial IrO 2 . Moreover, the Ir-NMP catalyst displayed ultrahigh mass activity of 1831 and 2604A g Ir −1 at representative overpotentials of 315 and 340 mV vs RHE, respectively, indicative of superior mass activity of the Ir-NMP catalyst and cost-effective OER using this catalyst.
The electrocatalytic OER durability of the Ir-NMP catalyst was assessed by successive cyclic voltammetry (CV) cycle test and chronoamperometry at an initial current density of 40 mA cm −2 (Fig. 8). As shown in Fig. 8a, for long-term durability of OER, the Ir-NMP catalyst showed an increase in current density in the first 5 h (approximately 20 mA cm −2 ), attributed to efficient electrochemical activation during electrochemical OER process. Then, the Ir-NMP catalyst retained almost 100% of its initial current density after more than 30 h of OER operation, with negligible decay in current density. It was probably attributed to the strong interfacial effect between Ir clusters and crystalline/amorphous Ni-Mo-P, which efficiently hindered the agglomeration and shedding of active sites in the Ir-NMP catalyst. It also indicated that excellent stability for electrochemical water oxidation was achieved over the Ir-NMP catalyst. On the contrary, commercial IrO 2 demonstrated obvious attenuation during current-time test, whose current density rapidly decreased from 40 to 7 mA cm −2 within 6 h. Furthermore, the polarization curve of the Ir-NMP catalyst after measured 5000 (5 k) times of CV cycles exhibited approximately 100% of its initial curve. In contrast, after 1000 (1 k) times of CV cycles, commercial IrO 2 catalyst underwent serve degradation of OER activity. These results further indicated outstanding durability of the Ir-NMP catalyst for electrocatalysis in alkaline media. A structural evaluation of the Ir-NMP catalyst after long-term stability test under approximately 60 mA cm −2 for 30 h was performed to investigate electrocatalytic stability. From the TEM images (Figs. 9a-c) and EDS mapping images (Fig. S8), there was no obvious structural collapse and particle aggregations in the Ir-NMP catalyst after stability test, together with the absent nanocrystals of Ir species, proving that the original surface structure of the Ir-NMP catalyst was maintained during long-term OER operation. The XPS results also confirmed that the Ir, Ni, Mo, and P elements existed in the Ir-NMP catalyst after stability test (Fig. 9d-f and Fig. S9). A slightly positive shift of 0.1-0.2 eV for these elements was observed in the Ir-NMP catalyst compared with the fresh Ir-NMP catalyst, indicating the transformation of metal species and phosphorus element into a slightly higher valence attributed to electrocatalytic oxidation reaction process. Moreover, Table S1 shows the comparisons of OER performances of recently reported Ir-and Ru-based catalysts in 1 M KOH, which indicated the superiority of Ir-NMP catalyst in electrocatalytic OER. Hence, the Ir-NMP catalyst with ultra-small Ir clusters and crystalline/amorphous structure of NMP exhibited the robust OER properties of low overpotential, low Tafel slope, high mass activity, and outstanding long-term durability, which was superior to commercial IrO 2 , pure NMP, and recently reported noble metal-based catalysts. The improved OER performances of the Ir-NMP a Long-term stability of the Ir-NMP catalyst and commercial IrO 2 . b CV stability of the Ir-NMP catalyst. c CV stability of commercial IrO 2 catalyst were probably explained by the following characters: (i) the incorporation of Ir clusters and isolated atoms embedded onto Ni-Mo-P could be considered as active sites for electrocatalytic, together with high oxidation state of Ir species (Ir 4+ ), which was favorable to enhance OER activity; (ii) the optimization of electronic structure and chemical states for Ni and Mo species in NMP derived from the introduction of Ir clusters, where the resulting Ni and Mo centers were probably available for electrocatalytic activities; (iii) the formation of one-dimensional nanorod structure composed of two-dimensional nanosheet subunits in the Ir-NMP catalyst, which was beneficial to the mass transfer process of OER.

Conclusion
In summary, the highly efficient and durable Ir-NMP electrocatalyst was synthesized through this simple combination of hydrothermal method and two-step annealing process. The interlaced crystalline and amorphous structure of Ni-Mo-P made 0.7-nm Ir clusters homogeneously embedded onto NMP substrate, and the introduction of Ir clusters enabled the optimization of electronic structure for NMP, which greatly accelerated the four-electron reaction kinetics toward efficient OER activity and superior durability in alkaline electrolytes. The as-synthesized Ir-NMP electrocatalyst delivered an ultralow overpotential of 340 mV vs RHE at 100 mAcm −2 , low Tafel slope of 63.9 mV dec −1 , high mass activity of 2604 A g Ir −1 , and large anodic currents in alkaline media, together with excellent catalytic stability with almost complete retention of activity within 33 h. This work fabricates a promising and low-cost OER electrocatalyst with superior performances and provides a simple pathway for the synthesis of low-dimensional metal catalysts for widespread applications.