Tungsten oxide‐anchored Ru clusters with electron‐rich and anti‐corrosive microenvironments for efficient and robust seawater splitting

Ruthenium (Ru) has been recognized as a prospective candidate to substitute platinum catalysts in water‐splitting‐based hydrogen production. However, minimizing the Ru contents, optimizing the water dissociation energy of Ru sites, and enhancing the long‐term stability are extremely required, but still face a great challenge. Here, we report on creating tungsten oxide‐anchored Ru clusters (Ru–WOx) with electron‐rich and anti‐corrosive microenvironments for efficient and robust seawater splitting. Benefiting from the abundant oxygen vacancy structure in tungsten oxide support, the Ru–WOx exhibits strong Ru–O and Ru–W bonds at the interface. Our study elucidates that the strong Ru–O bonds in Ru–WOx may accelerate the water dissociation kinetics, and the Ru–W bonds will lead to the strong metal–support interaction and electrons transfer from W to Ru. The optimal Ru–WOx catalysts exhibit a low overpotential of 29 and 218 mV at the current density of 10 mA cm−2 in alkaline and seawater media, respectively. The outstanding long‐term stability discloses that the Ru–WOx catalysts own efficient corrosion resistance in seawater electrolysis. We believe that this work offers new insights into the essential roles of electron‐rich and anti‐corrosive microenvironments in Ru‐based catalysts and provide a new pathway to design efficient and robust cathodes for seawater splitting.


INTRODUCTION
5][6][7][8] The exploitation of effective and stable catalysts for seawater electrolysis is of great importance for the green hydrogen economy.
0][11][12][13] Among these Pt-group metals, Ru, with only approximately 4% cost of Pt and very high corrosion resistance, has been considered one of the most promising alternatives for the high-cost Pt, owing to its lower costs and hydrolysis energy.However, the greater adsorption capacity for two intermediates (H* and OH*) in Ru atoms resulted in slow hydrogen production kinetics in both acidic and alkaline conditions. 14][17][18][19] Hence, it is extremely desirable, but it remains a challenge to optimize adsorption/desorption energies for the reaction intermediates to increase the activities of Ru catalysts in seawater splitting.1][22] For instance, Ru atoms in coordination with N and S atoms on a titanium carbide MXene support were reported to show enhanced HER activity; the overpotential needed to get 10 mA cm −2 is 76 mV in an acidic medium. 23The hcp-RuNi alloy was reported with reduced Ru-H binding effect and resulted in excellent HER activities in both acidic and neutral electrolytes. 24][34] However, most of these metal oxide supports exhibit poor corrosion resistance under a high saline environment, and the long-term stability of such metal oxide-supported catalysts under seawater condition are also unclear.It is highly desired to design and synthesize an anti-corrosive Ru-based catalyst with minimizing the Ru contents and optimizing the water dissociation energy on Ru sites for long-term seawater electrolysis.
Here, we report on creating tungsten oxide-anchored Ru clusters (Ru-WO x ) with electron-rich and anti-corrosive microenvironments for efficient and robust seawater splitting.The motivation to choose WO x as support is that 4d metal oxide usually shows high saline corrosion resistance, especially in chloride-rich seawater conditions.Benefiting from the abundant oxygen vacancy structure in WO x support, the experimental studies reveal that the Ru sites are coordinated with WO x through strong Ru-O bonds and Ru-W metal bonds.Our study elucidates that the strong Ru-O bonds in Ru-WO x may accelerate the water dissociation kinetics, and the Ru-W bonds will lead to the strong metal-support interaction and electrons transfer from W to Ru.Furthermore, the nanorods assembled microstructures will benefit the gas release process on the surface of the catalyst and further promote the HER process.Due to these distinctive features, the optimal Ru-WO x catalysts exhibit a low overpotential of 29 and 218 mV at 10 mA cm −2 in alkaline and seawater media, respectively.The outstanding long-term stability discloses that the Ru-WO x catalysts own efficient corrosion resistance in seawater electrolysis.We believe that this study will offer essential guidance for developing highly efficient and robust electrocatalysts for seawater splitting.

RESULTS AND DISCUSSION
The synthetic process is schematically illustrated in Figure 1A that dopamine (DA) molecules assemble with sodium tungstate dihydrate (Na 2 WO 4 ⋅2H 2 O) to form a hybrid metal-organic precursor (DA-W).Then the WO x nanorods on spherical carbon substrate were obtained by a carbonization process of the precursor under 700 • C in Argon atmosphere.The obtained WO x nanorods were then impregnated with different amounts of RuCl 3 in solution and followed with a secondary thermal treatment to achieve the growth of Ru sites on the nanorods (Ru 1 -WO x ).Meanwhile, the Ru-WO x with different Ru contents were synthesized by adjusting the amount of impregnated RuCl 3 , resulting in the catalysts, namely Ru 3 -WO x , Ru 1 -WO x , and Ru 0.5 -WO x (detail in Supporting Information).As the optimized Ru content is obtained at Ru 1 -WO x according to the electrochemical performance, we choose the Ru 1 -WO x as the repetitive sample for the main discussion.The morphologies of the as-synthesized WO x , Ru 3 -WO x , Ru 1 -WO x , and Ru 0.5 -WO x were first revealed by scanning electron microscopy (SEM).The SEM image of WO x exhibits that the diameter of these nanorods is around 100 nm and aggregated into a microspherical structure (Figure S1).The nanorod morphology and the microspherical structure of WO x were well maintained after introducing Ru species (Figure 1B and Figure S1).Transition electron microscopy (TEM) was further applied to confirm the crystal structure and elemental distributions.As shown in Figure 1C, the nanorod structures aggregated with carbon substrate can be observed, and the uniform distribution of Ru sites on WO x nanorods was proven by the energy-dispersive x-ray spectroscopy (EDS) mapping.Meanwhile, the more concentrated C species in the core reveal the aggregated microspherical structure of carbon, which will benefit the conductivity of the Ru-WO x catalysts.Figure 1D shows the HR-TEM image of Ru 1 -WO x , where the WO x substrate exhibits a well-ordered crystal structure, and Ru clusters are well-distributed on the WO x crystals.Meanwhile, the selected area electron diffraction (SAED) pattern on the nanorod clearly shows diffraction rings indexed to (010), (020), and (311) facet of W 10 O 29 .It was demonstrated that the size of the Ru sites becomes larger with increasing the Ru addition from Ru 0.5 -WO x to Ru 3 -WO x , but the morphology of the WO x nanorods does not change (Figures S2-S4).To illustrate the crystal structure of the Ru 1 -WO x materials, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) has been carried out.As shown in Figure 1E-G and Figures S5-S8, there are two types of crystalline structures, W 10 O 29 and distributed Ru nanocluster.The above results all demonstrate that Ru sites coordinated with WO x nanorods were obtained.The supposed material structure is illustrated in Figure 1J, where Ru nanoclusters are bonded on the WO x substrate through oxygen and tungsten.
Powder X-ray diffraction (XRD) was conducted to explore further the structures and compositions of WO x , Ru 0.5 -WO x , Ru 1 -WO x , and Ru 3 -WO x .Characteristic peaks of each sample are shown to be indexable to W 10 O 29 (PDF # 05-0286), and no apparent diffraction peaks for Ru crystals could be observed owing to the ultra-small size and low contents (Figure 2A).Notably, the crystallinity of W 10 O 29 increased along with the increasing amount of Ru, and a moderate crystallinity was obtained for Ru 1 -WO x .The typical peaks of O-W-O at 673 and 810 cm −1 in the Raman spectrum further demonstrate the presence of WO x substrates (Figure S9).To further understand the electronic structures and element ratios for these samples, X-ray photoelectron spectroscopy (XPS) was performed (Figure 2B).The total element contents were determined from XPS in Figure 2C and Table S1; the Ru contents increased gradually from 5.68 to 7.85 and 16.6 wt.% for Ru 0.5 -WO x , Ru 1 -WO x , and Ru 3 -WO x , respectively.In addi-tion, the chemical state of the O, Ru, and W in these catalysts was further analyzed by high-resolution XPS region scan (Figure 2D-H and Figures S10-S13).The deconvoluted W4f peaks correspond to the binding energy of three W oxidation states, W(VI) (35.51 eV), W(V) (35.14 eV), W(IV) (34.17 eV), and loss feature peak (Figure 2D).It is worth noting that the content of W(VI) increases from 40% to 48-49% with the addition of Ru (Figure 2G).In Ru 3p (Figure 2E), the primary peaks belong to Ru with Ru 0.5 -WO x , indicating that more charges are present in Ru sites of Ru 1 -WO x .For O 1s, two deconvoluted peaks can be detected at 530.38 and 532.07 eV, respectively, corresponding to lattice oxygen and adsorbed water molecules, which may be due to the presence of oxygen vacancies (V O ) (Figure 2F,H). 35,36The doping of Ru sites reduces the proportion of lattice oxygen, indicating more oxygen vacancies.The presence of oxygen vacancies in Ru 1 -WO x is further confirmed by the electron paramagnetic resonance (EPR) measurement (Figure 2I).
To explore the coordination structures of Ru sites in the Ru 1 -WO x catalyst, X-ray absorption spectroscopy (XAS) was also employed.As shown in Figure 3A, the Ru K-edges X-ray absorption near-edge structure (XANES) spectra show that the energy absorption threshold value of Ru 1 -WO x is slightly higher than that of Ru foil and much lower than RuO 2 , demonstrating that the valence state of Ru sites coordinated with WO x are closer to metallic Ru 0 .Further structural information of Ru sites in Ru foil, RuO 2 , and Ru 1 -WO x have been provided by the extended X-ray absorption fine structure (EXAFS) spectra and quantitatively by EXAFS simulations (Figure 3B,C, Figure S14, and Table S2). 37,38The main peak at about 2.36 Å refers to Ru−Ru coordination, confirming the formation of Ru nanoclusters.The peak at about 1.5 and 3.9 Å has, respectively, ascribed to Ru-O and Ru-W coordination, which demonstrates that Ru sites are coordinated with WO x substrates.
The wavelet transform (WT) analysis also proved consistent conclusions (Figure 3D-F).Thus, together with the XRD pattern, XPS spectra, and XAS on the microstructure of Ru 1 -WO x , the existing coordination structures between Ru sites and WO x substrate were confirmed.
The catalytic performances of the synthesized catalysts were investigated in a typical three-electrode system with rotating disk electrodes (RDE) in both alkaline (1.0 M KOH) and seawater electrolyte.The polarization curves of Ru 3 -WO x , Ru 1 -WO x , Ru 0.5 -WO x , WO x , and commercial Ru-C in alkaline conditions are presented in Figure 4A and Figures S15-S17.It can be noticed that the WO x without the coordinated Ru sites show barely visible currents for HER, which demonstrates the Ru atoms are the main active sites.It can be noticed that the Ru 1 -WO x exhibits an overpotential of 29 mV to reach the current density of 10 mA cm −2 , which is much lower than that of Ru 0.5 -WO x (56 mV).The decreased performance of Ru 3 -WO x (40 mV) with higher Ru content might be attributed to the larger particle size of Ru clusters on WO x .The Tafel slope of Ru 1 -WO x is 28 mV dec −1 , lower than that of Ru 3 -WO x (48 mV dec −1 ), Ru 0.5 -WO x (65 mV dec −1 ), and Ru-C (35 mV dec −1 ), indicating that the Ru 1 -WO x possesses the most efficient HER kinetics (Figure 4B,C).Moreover, exchange current density (j 0 ) has been calculated to investigate intrinsic reaction kinetics (Figure S18).Notably, the Ru 1 -WO x (4.47 mA cm −2 ) shows a much higher j 0 value than the Ru-C (2.24 mA cm −2 ), Ru 3 -WO x (1.78 mA cm −2 ), and Ru 0.5 -WO x (1.25 mA cm −2 ), which demonstrates its better intrinsic HER activity in alkaline media.
The mass activities and turnover frequency (TOF) of all samples are further calculated based on the Ru amount from XPS results, assuming that total Ru atoms were engaged in the catalytic process.Notably, the Ru 1 -WO x shows a TOF value of 3.66 H 2 s −1 per Ru atom at the overpotential of 100 mV, which is more than two times the values for commercial Ru-C (Figure 4D and Figure S19).At the same time, the mass current density of Ru 1 -WO x is 5 A mg −1 , which is 2.2, 3.3, and 2.2 times greater than that of Ru 0.5 -WO x (2.3A mg −1 ), Ru 3 -WO x (1.5 A mg −1 ), and Ru-C (2.3A mg −1 ), respectively (Figure S20).In addition, electrochemical impedance spectroscopy (EIS) was carried out to investigate the kinetics of electron transfer in the reaction (Figure 4E).The semicircle in the high-frequency region is related to electron transfer resistance within electrodes, which indicates Ru 1 -WO x with the fastest electron transfer in HER. 39The Ru 1 -WO x also exhibits much higher electrochemical double-layer capacitance (C dl ) (14.9 mF cm −2 ) compared with Ru 0.5 -WO x (8.9 mF cm −2 ), Ru 3 -WO x (7.1 mF cm −2 ), and WO x (5.7 mF cm −2 ) (Figure 4F and Figure S21).Moreover, the Ru 1 -WO x displays excellent long-term durability without obvious decay after running for 80 000 s, while the decay of Ru-C is even more severe to 80 mV after running for approximately 40 000 s (Figure 4G).Based on the above analyses, an overall property evaluation indicates that Ru 1 -WO x with an appropriate amount of Ru loading exhibits the lowest overpotential, the smallest Tafel slope, and the largest ECSA among the contrast catalysts leading to the best catalytic activity (Figure 4H).
Benefiting from the inexhaustible resources of seawater, seawater splitting is of growing attention from researchers.For practical application, we examined the HER performance of the catalysts in simulated seawater with pH = 8.27.We measured the hydrogen production capability of as-synthesized catalysts in simulated seawater.Compared to their corresponding LSV curves in alkaline media, all as-synthesized samples showed a decrease in activity in simulated seawater, which may be due to the decrease of conductivity of seawater electrolytes.As expected, the best-performing Ru 1 -WO x catalyst only needs a 218 mV overpotential to drive a current density of 10 mA cm −2 in seawater, which is much lower than that of Ru-C (350 mV) (Figure 5A).The corresponding Tafel slopes show that all the WO x -supported Ru catalysts (187, 158, and 159 mV dec −1 for Ru 0.5 -WO x , Ru 1 -WO x , and Ru 3 -WO x ) are showing much higher kinetics when compared with Ru-C catalyst (254 mV dec −1 ) (Figure 5B,C).The calculated j 0 for Ru 1 -WO x (0.55 mA cm −2 ) is higher than Ru 3 -WO x (0.52 mA cm −2 ), Ru 0.5 -WO x (0.50 mA cm −2 ), and Ru-C (0.45 mA cm −2 ), indicating the similar order of reaction kinetics for Ru-WO x and Ru-C (Figure S22).To further study the inherent hydrogen-evolving activity of Ru atoms, the TOF value and mass activity were calculated.As shown in Figure 5D, Ru 1 -WO x shows TOF values of 0.215 s −1 at 100 mV overpotential, which is higher than the Ru 0.5 -WO x (0.186 s −1 ), Ru 3 -WO x (0.11 s −1 ), and Ru-C (0.036 s −1 ).At the same time, the mass current density of Ru 1 -WO x is 0.13 A mg −1 , similar to Ru 0.5 -WO x (0.12 A mg −1 ), which is two and four times greater than that of Ru 3 -WO x (0.06 A mg −1 ) and Ru-C (0.03 A mg −1 ), respectively (Figure S23).Meanwhile, we compared the overpotentials of previously reported catalysts at 10 mA cm −2 , still confirming the excellent performance of Ru 1 -WO x in seawater (Figure 5E).More significantly, stable hydrogen production in seawater is allowed for more than 60 h at 10 mA cm −2 with Ru 1 -WO x catalyst, which is much better than the commercial Ru-C (Figure 5F).We believe that the excellent stability in seawater can be attributed to the essential roles of electron-rich and anti-corrosive microenvironments bring by the strong metal-support interaction and electrons transfer from W to Ru in the WO x -anchored Ru catalysts (Figure 5G).

CONCLUSION
Our study demonstrates the tungsten oxide-anchored Ru catalysts with electron-rich and anti-corrosive microenvironments for efficient and robust seawater splitting.The strong Ru-O bond at the interface provides accelerated water dissociation kinetics, while the Ru-W bond with strong metal-support interaction and electrons transfer from W to Ru resulted in the anti-corrosive microenvironments.Therefore, the resulting Ru 1 -WO x catalysts exhibit a low overpotential of 29 and 218 mV at 10 mA cm −2 in alkaline and seawater media, respectively.Moreover, the Ru 1 -WO x catalyst owns efficient corrosion resistance in seawater electrolysis with continuous hydrogen production at 10 mA cm −2 for more than 60 h.It is believed that this work will not only offer new insights into the essential roles of electron-rich and anti-corrosive microenvironments, but will also provide new pathways to design efficient and robust catalysts for green hydrogen production.

METHODS
Synthesis of Ru-WO x and WO x : First, 2.5 mmol (0.475 g) dopamine hydrochloride (DA) was dissolved in 25 mL water, resulting in 0.1 M dopamine in water solution.
The pH value of the dopamine solution was adjusted to approximately 2 by adding 2 mL of 1 M HCl.After that, 25 mL sodium tungstate solution (containing 2.5 mmol Na 2 WO 4 ⋅2H 2 O) was dropped slowly (about 10 min) into the DA solution.The reaction happens immediately, then the precipitation happens with further addition of sodium tungstate.Finally, a greenish-yellow precipitate was formed and collected after reaction for 1 h by centrifugation, and was washed with DI water and ethanol three times.The products were dried in an oven at 60 • C overnight to obtain this hybrid metal-organic precursors (DA-W).Then the DA-W precursor was carbonized in an Argon furnace at 700 • C for 2 h, with a ramp of 5 • C/min.This black powder was named as WO x .The WO x (60 mg) was re-dispersed in 5 mL water by ultrasonic treatment for 20 min, then 3/1/0.5 mL metal-phen solution (10 mg/mL, prepared by mixing RuCl 3 with water) was added into the above dispersion and ultra-sonicated for another 20 min.Then, the mixed solutions were placed in a drying oven at 60 • C, and then placed in the crucible and transferred to a tube furnace at 600 • C for 2 h with a ramp rate of 5 • C/min in an Ar atmosphere to obtain WO x -supported Ru catalysts, which are denoted as Ru 3 -WO x , Ru 1 -WO x , and Ru 0.5 -WO x , respectively.For comparison, WO x supports without Ru content also were placed in a tube furnace at 600 • C for 2 h with a ramp rate of 5

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.

F I G U R E 1
Synthesis process and characterization of Ru sites coordinated with WO x. (A) Schematic illustration of the synthetic procedure for Ru-WO x .(B) The scanning electron microscopy (SEM) images.(C) The energy-dispersive x-ray spectroscopy (EDS) mapping.(D) Transition electron microscopy (TEM) and selected area electron diffraction (SAED) pattern images.(E-G) The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images, and (J) the corresponding crystal structure illustration of the as-synthesized Ru 1 -WO x .

F
I G U R E 2 Structural comparisons of Ru-WO x .(A) XRD patterns, (B) XPS survey scans of Ru 3 -WO x , Ru 1 -WO x , Ru 0.5 -WO x , and WO x , and (C) the corresponding atom elemental ratios.High-resolution XPS region scan of (D) W 4f, (E) Ru 3p, and (F) O 1s of Ru 3 -WO x , Ru 1 -WO x , Ru 0.5 -WO x , and WO x , and the corresponding (G) W species ratio and (H) percentage values of lattice oxygen and absorbed oxygen.(I) EPR spectra of Ru 1 -WO x and WO x .
3p 3/2 (461.88 eV) and Ru 3p 1/2 (484.14 eV) of Ru(0) and other peaks coming from Ru(IV), which indicates the presence of metallic Ru in these samples.The Ru(0) peak of Ru 1 -WO x shows a 0.15 eV positive shift value compared F I G U R E 3 The coordination environment of Ru sites.(A) Ru K-edge XANES spectra and (B) Fourier transforms of the Ru K-edge EXAFS spectra of Ru foil, RuO 2 , and Ru 1 -WO x .(C) FT-EXAFS fitting curves of Ru in Ru 1 -WO x .(D-F) Wavelet transform (WT) images of Ru foil, RuO 2 , and Ru 1 -WO x , respectively.

F I G U R E 4
Electrocatalytic HER activities and stabilities in alkaline media.(A) Linear sweep voltammetry curves, and (B) the corresponding Tafel plots of Ru 3 -WO x , Ru 1 -WO x , Ru 0.5 -WO x , WO x , and Ru-C in 1.0 M KOH.(C) Overpotentials at 100 mA cm −2 and Tafel slopes, (D) turnover frequency (TOF) values, (E) Nyquist plots, and (F) C dl plots inferred from CV curves of Ru 3 -WO x , Ru 1 -W x , Ru 0.5 -WO x , WO x, and Ru-C.(G) Stability of Ru 1 -WO x and Ru-C working at the current density of 10 mA cm −2 .(H) Overall evaluation of the properties during HER.

F I G U R E 5
Electrocatalytic analysis on HER activities and stabilities in seawater.(A) Linear sweep voltammetry curves.(B) The corresponding Tafel plots of Ru 3 -WO x , Ru 1 -WO x , Ru 0.5 -WO x , WO x , and Ru-C in seawater.(C) Histograms of overpotential at 10 mA cm −2 and Tafel slope.(D) Turnover frequency (TOF).(E) Comparison of overpotentials at 10 mA cm −2 of Ru 1 -WO x with those of reported HER catalysts. 40-49(F) Durability tests of Ru 1 -WO x and Ru-C at 10 mA cm −2 in seawater.(G) Schematic diagram representation of water dissociation formation of H 2 process on Ru 1 -WO x .
• C/min in an Ar atmosphere to obtain the control sample, named as WO x .Yiming Zhang, Shuang Li, Chong Cheng, and Yinghan Wang conceived the idea and designed the project.Yiming Zhang performed the major experiments and analyzed the results.Weiqiong Zheng, Huijuan Wu, and Ran Zhu assisted with figure production.Mao Wang, Tian Ma, and Zhiyuan Zeng helped with the data analysis and experiment design.Yiming Zhang, Zhiyuan Zeng, Chong Cheng, and Shuang Li wrote and edited the manuscript.Chong Cheng and Shuang Li supervised the whole project.All authors discussed the results and commented on the manuscript.This work was financially supported by the National Natural Science Foundation of China (grant number 52273269) and the Sichuan Science and Technology Program (grant numbers 2023YFH0027, 2023YFH0008).We acknowledge the financial support from Fundamental Research Funds for the Central Universities and the State Key Laboratory of Polymer Materials Engineering (grant numbers sklpme2022-3-07 and sklpme2021-4-02).Zhiyuan Zeng thanks GRF (CityU11308923) and the Basic Research Project from Shenzhen Science and Technology Innovation Committee (No. JCYJ20210324134012034).We gratefully acknowledge Dr. Mi Zhou and Dr. Chao He at Sichuan University for their assistance with the experiments.
A U T H O R C O N T R I B U T I O N S A C K N O W L E D G M E N T S