Synergistic Mechanism of Sub‐Nanometric Ru Clusters Anchored on Tungsten Oxide Nanowires for High‐Efficient Bifunctional Hydrogen Electrocatalysis

Abstract The construction of strong interactions and synergistic effects between small metal clusters and supports offers a great opportunity to achieve high‐performance and cost‐effective heterogeneous catalysis, however, studies on its applications in electrocatalysis are still insufficient. Herein, it is reported that W18O49 nanowires supported sub‐nanometric Ru clusters (denoted as Ru SNC/W18O49 NWs) constitute an efficient bifunctional electrocatalyst for hydrogen evolution/oxidation reactions (HER and HOR) under acidic condition. Microstructural analyses, X‐ray absorption spectroscopy, and density functional theory (DFT) calculations reveal that the Ru SNCs with an average Ru—Ru coordination number of 4.9 are anchored to the W18O49 NWs via Ru—O—W bonds at the interface. The strong metal‐support interaction leads to the electron‐deficient state of Ru SNCs, which enables a modulated Ru—H strength. Furthermore, the unique proton transport capability of the W18O49 also provides a potential migration channel for the reaction intermediates. These components collectively enable the remarkable performance of Ru SNC/W18O49 NWs for hydrogen electrocatalysis with 2.5 times of exchange current density than that of carbon‐supported Ru nanoparticles, and even rival the state‐of‐the‐art Pt catalyst. This work provides a new prospect for the development of supported sub‐nanometric metal clusters for efficient electrocatalysis.


Introduction
Dispersion of active metal nanoparticles (NPs) on supports is an effective strategy to improve the utilization of precious metals and thus reduce catalyst costs, which has been widely used in and the supported NPs often aggregate or fall off from the carbon substrates and thus cause catalyst failure. [15][16][17][18] Moreover, the inherent inertness of carbon usually makes it difficult to participate directly in the electrocatalytic reactions and thus limits their ability to play a more important synergistic role during electrocatalysis. [19,20] In contrast, metal oxides are a promising class of supports in which metal NPs can be anchored on the supports via metal-oxygen bonds, forming strong electronic metalsupport interaction (EMSI). [21][22][23] Particularly for the reducible metal oxides (e.g., TiO 2-, CeO 2-, and WO 3-), the presence of coordinatively unsaturated cationic sites in these oxides not only facilitate electron transport in the oxide network, but also provide a promising way to promote the unique synergistic effect during the catalytic process. [24][25][26] For example, the migration of hydrogen atoms from metal NPs to reducible metal oxide supports (hydrogen spillover) has been widely reported to play a key facilitating role in many heterogeneous catalysis applications. However, despite significant efforts have been made to develop NPs/oxide hybrid materials over recent years, there is still a lack of facile synthetic methods to construct the complexes of metal SNCs and oxide supports. In addition, compared with the extensive attention to the mechanism of metal-support interactions in thermal catalysis, more efforts are also necessary to provide deeper investigation regarding the synergistic mechanism between metal SNCs and oxide supports under specific electrocatalytic reactions. [27,28] Herein, we report a facile one-pot hydrothermal method to prepare the tungsten oxide nanowires supported Ru sub-nanometric clusters (Ru SNC/W 18 O 49 NWs), which exhibit apparent electrocatalytic performance for both hydrogen evolution and oxidation reactions (HER and HOR). Particularly, the Ru SNC/W 18 O 49 NWs only require an overpotential of 21 mV to drive HER current densities of 10 mA cm −2 , with a Tafel slope of 35 mV dec −1 and exchange current density of 2.5 mA cm −2 , outperforming the comparative Ru NPs and most of the other reported catalysts (

Results and Discussion
The Ru SNC/W 18 O 49 NWs were synthesized via a facile onepot hydrothermal method using tungsten (VI) chloride (WCl 6 ), ruthenium (III) chloride (RuCl 3 ), and ethanol as metal precursor and solvent, respectively (Figure 1a). Transmission electron microscopy (TEM) images show that the obtained Ru SNC/W 18 (Figure 1c). The magnified AC-HAADF-STEM image and corresponding 3D HAADF-STEM surface plot image clearly show that the Ru SNC with a size of ≈1 nm is anchored on the surface of the NWs (Figures 1d,e). Moreover, as presented in high-resolution TEM (HRTEM) images (Figure 1f), the direction consistent lattice fringes with a spacing of 0.38 nm can be assigned to the (010) planes of monoclinic W 18 (Figure 1g). According to the EDS analysis results, the atomic ratio of Ru to W is ≈4.4:95.6 ( Figure S3, Supporting Information). X-ray diffraction (XRD) patterns of Ru SNC/W 18  X-ray absorption near edge spectroscopy (XANES) and extended X-ray absorption fine structure (EXAFS) were measured to further identify the electronic state and local coordination environment of different samples. Figure 2c depicts the normalized Ru K-edge XANES spectra of Ru SNC/W 18 O 49 NWs, Ru NP/W 18 O 49 NWs, and Ru/C, with the Ru foil and RuO 2 as references. The absorption edge of Ru/C almost overlaps with the Ru foil, which suggests the metallic state of Ru in it. In sharp   (Figure 2d). This conclusion is also confirmed by the W L 3 -edge XANES spectra analysis ( Figure S6   The average coordination number of the Ru-Ru bond in Ru SNC/W 18 O 49 NWs is only 4.9, which is obviously lower than that of Ru NP/W 18 O 49 NWs (CN = 9.9) and Ru foil (CN = 12), corresponding to the limited atom number of SNC. [23] In addition, the average coordination number of the Ru-O bond is 2.6 in Ru SNC/W 18 O 49 NWs, with a longer bond length (1.983 Å) than that of RuO 2 (1.970 Å), indicating its different Ru-O coordination environment compared to the latter. These results suggest that Ru SNCs are anchored to the W 18 O 49 support through oxygen atoms at the interface (inset of Figure 2g).
Upon confirming the structural features of Ru SNC/W 18 O 49 NWs, we subsequently evaluated its efficacy in electrocatalysis. As a proof of concept, we demonstrate Ru SNC/W 18 O 49 NWs as an advanced catalyst for HER and HOR, which are two crucial reactions for realizing the future hydrogen economy. Currently, Pt is considered to be the state-of-the-art electrocatalyst for HER and HOR in acidic condition, but they suffer from high cost. Therefore, the development of efficient and low-cost electrocatalysts for HER and HOR has received extensive attention, and many advanced non-Pt electrocatalysts have been reported in recent years. [29][30][31][32] Due to the attractive performance and desirable cost, Ru has become a highly promising candidate to replace Pt for HER and HOR, while the hydrogen binding at Ru sites (Ru-H) is an important factor limiting its activity. [33,34] Considering the unique proton conductivity of tungsten oxides, which is expected to emerge unique catalytic synergies with the Ru SNCs in hydrogen electrocatalysis. [35,36] We first evaluated the electrocatalytic HER performance of the obtained Ru SNC/W 18   measured under the same conditions. Figure 3a shows the HER polarization curves of Ru SNC/W 18  NWs show an overpotential of 21 mV to achieve the current density of 10 mA cm −2 , only 2 mV more than Pt/C (19 mV), while that of Ru/C and Ru NP/W 18 O 49 NWs is 89 and 118 mV, respectively. Tafel slopes were evaluated to gain insight into the kinetics of the HER process. As shown in Figure 3b, a Tafel slope of 35 mV dec −1 was measured for Ru SA/W 18 O 49 NWs, close to the level of Pt/C (21 mV dec −1 ), while the much larger Tafel slopes were detected for Ru/C (118 mV dec −1 ) and Ru NP/W 18 O 49 NWs (101 mV dec −1 ). With a low Tafel slope, the HER rate of Ru SNC/W 18 O 49 NWs will increase rapidly with increasing overpotential, leading to a competitive advantage for practical applications. By extrapolating the Tafel plots, the exchange current density of Ru SNC/W 18 O 49 NWs was obtained (2.5 mA cm −2 ), which is higher than the most of reported Rubased catalysts and even Pt/C (1.4 mA cm −2 ) (Figure 3c;  Figure 3f, the polarization curve of Ru SNC/W 18 O 49 NWs shows negligible changes after 5000 potential cycles. The good durability of the Ru SNC/W 18 O 49 NWs was also demonstrated by the chronopotentiometry (CP) test, in which the overpotential changed slightly during the 20 h continuously electrolyzed at a current of 10 mA cm −2 . The HER activity of Ru SNC/W 18 O 49 NWs in 1 M KOH was also tested, which is obviously lower than that of in 0.5 m H 2 SO 4 , probably due to the different electrocatalytic mechanism between alkaline and acidic conditions ( Figure S15  method using a standard three-electrode system. Figure 4a shows the HOR polarization curves of different catalysts in H 2 -saturated 0.5 M H 2 SO 4 . The anode current density of Ru SNC/W 18 O 49 NWs increases sharply with increasing potential, which is even higher than that of the state-of-the-art Pt/C, and commercial Ru/C catalyst, demonstrating its excellent catalytic performance toward HOR ( Figure S16 and Table S5, Supporting Information). In sharp contrast, Ru NP/W 18 O 49 NWs, W 18 O 49 NWs, and Ru/C show much poor HOR activity ( Figure S17, Supporting Information). We also tested the HOR polarization curves of Ru SNC/W 18 O 49 NWs at different rotating speeds (Figure 4b). The limiting current density increases along with the elevation of the rotating speed, demonstrating the H 2 mass-transport controlled process. The Koutecky-Levich plots are shown in the inset of Figure 4b. Note that the LSV obtained in N 2 -saturated electrolyte was used as the background to reduce the influence of non-Faraday current ( Figure S18, Supporting Information). An average slope of 11.1 mA −1 cm 2 rpm 1/2 is obtained, which is close to the theoretical number and indicates the current is derived from the twoelectron transfer HOR process ( Figure S19, Supporting Information). Parasitic oxygen reduction reaction (ORR) at the anode due to accidental air leakage into the anode flow field, followed by transient potential jumps and severe corrosion flow field at the cathode, is currently one of the important challenges facing Ptbased HOR catalysts. [37] As shown in Figure 4c, Ru SNC/W 18 O 49 NWs exhibit poor response to O 2 compared to Pt/C, suggesting its better transient stability in practical applications. Moreover, the anode Pt catalyst of PEMFC is readily poisoned by impurity gas such as CO that existed in hydrogen fuel. [38,39] Such poisoning is caused by the preferential CO binding on metal sites, which consequently blocks the sites for hydrogen adsorption and dissociation. We, therefore, examined the HOR activities of the Ru SNC/W 18 O 49 NWs catalyst in the presence of CO. Interestingly, the Ru SNC/W 18 O 49 NWs showed a small decrease of HOR current even in the presence of 1000 ppm CO, indicating excellent selectivity for the HOR against CO (Figure 4d). In contrast, significant declining HOR activity on the Pt/C catalyst was detected at the same CO concentration, suggesting serious poisoning of the active sites for H 2 oxidation by CO binding. Further chronoamperometry test revealed that Ru SNC/W 18 O 49 NWs possess excellent stability during HOR catalysis and better tolerance to CO poisoning than that of Pt/C catalyst (Figure 4e). In addition, TEM and XPS were carried out to further characterize the Ru SNC/W 18 O 49 NWs after the electrocatalytic test. As shown in Figure S20 (Supporting Information), Ru SNC/W 18 (Figure 5a). Moreover, the charge density difference analysis reveals the electron transfer from the Ru cluster to W 18 O 49 , which is consistent with the results of XPS and XAS (inset of Figure 5a). As proposed by Nørskov et al., the adsorption energy of hydrogen has been widely employed as a descriptor for predicting the HER/HOR performance of catalysts, which should be an optimal value for neither too strong nor too weak binding. [40] The active sites of the catalyst capture the hydrogen atom by interacting with the unsaturated electron on the H 1s orbital, thus the changed electron density of Ru is expected to affect its adsorption with the hydrogen atom. Therefore, we then calculated the adsorption energy of hydrogen on Ru/W 18 O 49 , pure W 18 O 49, and Ru for comparison. As shown in Figure 5b, the calculated hydrogen adsorption strength of Ru/W 18

Conclusion
In summary, we reported a facile one-pot hydrothermal method for the synthesis of sub-nanometric Ru clusters loaded W 18  NWs exhibit a significantly enhanced HER/HOR activity than that of compared Ru NPs, and it can even rival the activities of the state-of the-art Pt/C catalyst. Moreover, the excellent durability and great tolerance to CO and O 2 impurities, coupled with the relatively low cost of Ru, makes Ru SNC/W 18 O 49 NWs an attractive candidate to replace Pt for PEM-based hydrogen electrocatalysis. This work demonstrates the great potential of constructing oxides-supported sub-nanometric clusters as high-performance electrocatalysts.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.