Spillover of active oxygen intermediates of binary RuO₂/Nb₂O₅ nanowires for highly active and robust acidic oxygen evolution

Abstract

Over-oxidation of surface ruthenium active sites of RuO_x-based electrocatalysts leads to the formation of soluble high-valent Ru species and subsequent structural collapse of electrocatalysts, which results in their low stability for the acidic oxygen evolution reaction (OER). Herein, a binary RuO₂/Nb₂O₅ electrocatalyst with abundant and intimate interfaces has been rationally designed and synthesized to enhance its OER activity in acidic electrolyte, delivering a low overpotential of 179 mV at 10 mA cm⁻², a small Tafel slope of 73 mV dec⁻¹, and a stabilized catalytic durability over a period of 750 h. Extensive experiments have demonstrated that the spillover of active oxygen intermediates from RuO₂ to Nb₂O₅ and the subsequent participation of lattice oxygen of Nb₂O₅ instead of RuO₂ for the acidic OER suppressed the over-oxidation of surface ruthenium species and thereby improved the catalytic stability of the binary electrocatalysts.

---

New concepts

The lattice oxygen mechanism (LOM) can break the scale relationship of Ru-based electrocatalysts for the acidic oxygen evolution reaction (OER) through the involvement of lattice oxygen, which can greatly enhance the activity. However, the generation of surface high-valent ruthenium species accompanied by the formation of oxygen vacancies leads to dissolution of surface Ru species and then decay of catalytic activity. Herein, we designed and synthesized a binary RuO₂/Nb₂O₅ nanowire catalyst with intimate interfaces, which enabled the spillover of the reactive intermediate *O from RuO₂ to Nb₂O₅ and thereby delivered significantly improved activity and durability for the acidic OER. Combination of various in situ and ex situ techniques demonstrated the occurrence of the spillover of active oxygen intermediates from RuO₂ to Nb₂O₅ and the subsequent participation of lattice oxygen of Nb₂O₅ instead of RuO₂ for the acidic OER, thereby suppressing the over-oxidation of surface ruthenium species. We anticipate that this work can be beneficial in stimulating the study of oxygen spillover at the interface of binary oxide catalysts for designing high performance OER electrocatalysts.

Introduction

The rational design and synthesis of Ru-based acidic OER electrocatalysts with low overpotential and high durability are essential for the development of electrocatalytic hydrogen generation from water electrolysis in acidic media.^1,2 Previous investigations have suggested that the surface Ru species of RuO₂ tend to be over-oxidized during the acidic OER and generate the soluble high valent Ru species under external potentials of over 1.4 V.^1,3,4 Especially, defective RuO_x has been reported to deliver high catalytic activity for the acidic OER, which has been recognized to undergo a lattice oxygen mediated (LOM) pathway.^3,5–7 Participation of lattice oxygen leads to the formation of oxygen vacancies and subsequent production of high valent Ru species, thereby unavoidably inducing the serious dissolution of surface electroactive high-valent Ru species and consequent collapse of the crystal structure of the highly defective RuO_x-based electrocatalysts during the long-term operation. Thus, it's highly expected that new strategies to realize high performance Ru-based electrocatalysts for acidic OER can be developed.

To address this challenge, various interfacial and structural engineering strategies have been developed to modulate the valence state of Ru and suppress the over-oxidation of surface Ru species, including the hybridization of Ru-based catalysts with other components, chemical doping, formation of solid solution, etc.^8–11 Moreover, alternating between the LOM pathway of the defective RuO_x-based electrocatalysts and adsorbate evolution mechanism (AEM) pathway or oxide path mechanism (OPM) also showed the potential to enhance the catalytic activity and durability of Ru-based catalysts for the acidic OER.^3,8,12,13 However, the catalytic performances were still far below the expectations for the practical applications, and part of the above-mentioned methodologies improved the catalytic stability of Ru-based electrocatalysts at the expense of activity.^14–16 Recently, the spillover of active oxygen intermediates has been proposed as a promising strategy to simultaneously enhance the activity and stability of IrO_x-based OER electrocatalysts in acidic electrolyte.¹⁷ We recently reported the spillover of active oxygen intermediates from RuO₂ to MoO₃ in the binary RuO₂/MoO₃ electrocatalysts for the enhanced OER activity and stability in acidic electrolyte.¹⁸ In this approach, MoO₃ with a low solubility in acidic media was integrated with RuO₂, in which the active oxygen species was initially generated on the surface active sites of RuO₂ and then spilled over towards MoO₃ for the release of oxygen. The spillover of the active oxygen species from RuO₂ to MoO₃ for the subsequent release of oxygen can avoid the generation of oxygen vacancies on RuO₂ and consequently suppresses the over-oxidation of surface Ru species, which theoretically improves the catalytic activity and stability. To date, only a few electrocatalysts have been reported to realize the simultaneously improved activity and stability for the acidic OER through the oxygen spillover strategy. To investigate the universal applicability of the oxygen spillover strategy for enhancing the acidic OER performance of RuO₂, the exploration of new binary electrocatalysts is anticipated.

Herein, a binary electrocatalyst of ruthenium oxide and niobium oxide nanowire (RuO₂/Nb₂O₅) was synthesized via a facile electrospinning/thermal treatment approach. RuO₂/Nb₂O₅ delivered a remarkable low overpotential of 176 mV at a current density of 10 mA cm⁻², a Tafel slope of 73 mV dec⁻¹, and a mass activity of 296 A g⁻¹, suggesting substantially superior catalytic performance in comparison to commercial RuO₂. Furthermore, RuO₂/Nb₂O₅ demonstrated an exceptional stability, enduring for at least 750 h in 0.5 M H₂SO₄. The S-numbers of RuO₂/Nb₂O₅ (3.01 × 10⁵) and C-RuO₂ (1.07 × 10³) indicated a significantly reduced dissolution rate of ruthenium in RuO₂/Nb₂O₅. Extensive experiments demonstrated that the integration of Nb₂O₅ and RuO₂ effectively inhibited the over-oxidation of surface Ru species and avoided the dissolution of Ru of the binary electrocatalysts for the acidic OER. Catalytic mechanism studies suggested the occurrence of spillover of the active oxygen species from RuO₂ to Nb₂O₅ during the acidic OER. Such a binary electrocatalyst broke the dilemma between the activity and stability of Ru-based electrocatalysts by preserving lower oxidation states of surface Ru species for a long period in the acidic OER.

Synthesis and characterization of RuO₂/Nb₂O₅

The synthetic process of the RuO₂/Nb₂O₅ binary nanowire electrocatalysis involves electron-spinning and subsequent thermal treatment at various temperatures. The details can be found in the ESI.† Typically, the nanowire-like precursors were prepared by electro-spinning of a polyvinyl pyrrolidone (PVP) solution containing RuCl₃ and NbCl₅ in the desired molar ratios. Then, calcination under a flow of air removed the polymers and transformed the wire-like precursors into binary oxide catalysts.¹⁹

Three RuO₂/Nb₂O₅ electrocatalysts were synthesized with the Ru/(Ru + Nb) molar ratios of 0.7, 0.6, and 0.5, which were determined using inductively coupled plasma emission spectrometry (ICP-ES, Table S1, ESI†). The catalysts were named RuO₂/Nb₂O₅-x, where x represented the molar ratio of Ru/(Ru + Nb). The transmission electron microscopy (TEM) image demonstrated the nanowire-like morphology of the binary RuO₂/Nb₂O₅-0.6 electrocatalysts (treated at 400 °C) with an average diameter of ∼100 nm (Fig. 1a). The electrocatalysts were composed of small nanoparticles, indicating the highly accessible surface and largely exposed active sites for electrocatalysis. High-resolution TEM (HRTEM) images of RuO₂/Nb₂O₅-0.6 revealed a crystal structure with a lattice fringe spacing of 0.32 nm, which corresponded to the (110) lattice plane of rutile phase RuO₂ (Fig. 1b),²⁰ while no lattice fringe information of Nb₂O₅ was found, which could be attributed to the poor crystalline structure and small size of Nb₂O₅ in the catalysts. Fig. 1c showed the X-ray diffraction (XRD) patterns of RuO₂/Nb₂O₅-0.6 and RuO₂. RuO₂ was synthesized through the identical approach of RuO₂/Nb₂O₅ in the absence of the Nb precursor (Fig. S1, ESI†). The characteristic peaks of RuO₂ and RuO₂/Nb₂O₅ electrocatalysts matched well with the database profile of RuO₂ (rutile phase, PDF#: 97-064-7373), suggesting the formation of rutile phase RuO₂ in the catalysts.²¹ The absence of characteristic XRD peaks of Nb₂O₅ further indicated the poorly crystallized and/or small Nb₂O₅ particles in the electrocatalysts. Energy-dispersive X-ray spectroscopy (EDS) elemental mapping images suggested the uniform distribution of Ru, Nb and O in a single nanowire of RuO₂/Nb₂O₅-0.6, demonstrating the presence of Nb in the electrocatalyst (Fig. 1d). As shown in Table S2 (ESI†), the atomic and weight ratios of Ru/(Ru + Nb) in RuO₂/Nb₂O₅-0.6 derived from EDS elemental mapping images were 0.63 and 0.65, respectively, which were close to the results of ICP (0.59 and 0.62). Similar EDS and ICP results further indicated that the distribution of elements within the catalyst was uniform. To further determine the formation of a binary structure of the electrocatalyst, Raman spectra of RuO₂, Nb₂O₅ and RuO₂/Nb₂O₅-0.6 were recorded and analyzed. In the Raman spectrum of RuO₂/Nb₂O₅-0.6, the characteristic peaks at 517 and 624 cm⁻¹ corresponded to the E_g and A_1g vibrational modes of RuO₂, respectively (Fig. 1e).²² The broad peak surrounding at 748 cm⁻¹ was attributed to the Nb–O vibrational peak of Nb₂O₅. In comparison with the characteristic peak of highly crystalline commercial Nb₂O₅ appearing at 693 cm⁻¹, the blueshift and broadening of the characteristic peak of the RuO₂/Nb₂O₅-0.6 catalysts by nearly 50 cm⁻¹ suggested the poor crystalline and small size of Nb₂O₅ in the binary electrocatalysts.²³

Fig. 1 Characterization studies of RuO₂/Nb₂O₅-0.6. (a) TEM image of RuO₂/Nb₂O₅-0.6. (b) HRTEM image of RuO₂/Nb₂O₅-0.6. (c) XRD patterns of RuO₂/Nb₂O₅-0.6 and RuO₂. (d) Elemental X-ray mappings of RuO₂/Nb₂O₅-0.6. (e) Raman spectra of RuO₂/Nb₂O₅-0.6, RuO₂ and Nb₂O₅.

Catalytic performance

Initially, the catalytic performance of a series of RuO₂/Nb₂O₅ electrocatalysts was evaluated in a three-electrode electrolytic cell with 0.5 M H₂SO₄ as the electrolyte, in which graphite rod and Hg/Hg₂Cl₂/KCl electrode were employed as the counter and reference electrodes, respectively. All electrode potentials were calibrated and normalized to the reversible hydrogen electrode (RHE). After calcination under different temperatures, various RuO₂/Nb₂O₅ electrocatalysts were synthesized, exhibiting similar morphological features and the same crystal phase (Fig. 1 and Fig. S2–S5, ESI†). Afterwards, the catalytic performance of the prepared catalysts was analyzed to screen out the optimal composition and treatment temperature. As shown in Fig. S6 and S7 (ESI†), the RuO₂/Nb₂O₅-0.6 electrocatalyst treated at 400 °C delivered the best OER activity with the lowest overpotential of 179 mV at 10 mA cm⁻² and the smallest Tafel slope of 73 mV dec⁻¹, which was identified as the optimal electrocatalyst in this work. Then, RuO₂, the commercially available Nb₂O₅ and RuO₂ (C-RuO₂) catalysts, as well as the electron-spun RuO₂ electrocatalysts synthesized through the identical process of RuO₂/Nb₂O₅-0.6 were selected as the control catalysts to highlight the importance and roles of Nb₂O₅ in improving catalytic performance of RuO₂/Nb₂O₅-0.6. As shown in their linear voltammetry scanning (LSV) curves (Fig. 2a), the bare OER activity of Nb₂O₅ was observed, indicating the catalytic inertness of Nb₂O₅. Comparatively, the RuO₂/Nb₂O₅-0.6 catalysts delivered the lowest overpotential of 179 mV at a current density of 10 mA cm⁻², which was much lower than those of RuO₂ (215 mV) and C-RuO₂ (256 mV). The catalytic current density of RuO₂/Nb₂O₅-0.6 at 1.5 V was 75.2 mA cm⁻², which was 2.7 times higher than that of the electron-spun RuO₂ electrocatalysts (27.9 mA cm⁻²) and 6.1 times higher than that of the commercial C-RuO₂ electrocatalysts (12.3 mA cm⁻²). The reaction kinetics were further evaluated by the derived Tafel slopes. The Tafel slope of RuO₂/Nb₂O₅-0.6 (73 mV dec⁻¹) was significantly lower than that of RuO₂ (107 mV dec⁻¹) and C-RuO₂ (164 mV dec⁻¹), indicating that the combination of ruthenium oxide and niobium oxide significantly enhanced the OER kinetics of the binary electrocatalysts (Fig. 2b).

Fig. 2 Catalytic performance of various electrocatalysts in 0.5 M H₂SO₄ electrolyte. (a) LSV curves of RuO₂/Nb₂O₅-0.6, RuO₂, C-RuO₂, and Nb₂O₅. (b) Tafel slopes of RuO₂/Nb₂O₅-0.6, RuO₂, C-RuO₂, and Nb₂O₅. (c) Chronopotentiometry tests of RuO₂/Nb₂O₅-0.6 and C-RuO₂.

In order to explore the intrinsic activity of the catalysts, the electrochemically active surface area (ECSA) was calculated from the electrochemical double layer capacitance (C_DL) by cyclic voltammetry (CV) tests (Fig. S8, ESI†).²⁴ After analyzing and fitting the data, the C_DL values of various electrocatalysts were calculated to be 2.73, 5.18 and 11.89 mF for the commercial C-RuO₂, RuO₂ and RuO₂/Nb₂O₅-0.6, respectively. C_S was the capacitance of an atomically smooth planar surface, which was 0.06 mF cm⁻² in acidic media.²⁵ According to ECSA = C_DL/C_S, the RuO₂/Nb₂O₅-0.6 electrocatalysts possessed the highest ECSA of 198.2 cm², which was significantly higher than that of RuO₂ (86.4 cm²) and commercial RuO₂ (45.5 cm²).²⁴ By normalizing with the respective ECSA, the RuO₂/Nb₂O₅-0.6 electrocatalysts delivered the highest intrinsic OER activity of 200 mV at 0.1 mA cm_ECSA⁻² under the acidic electrolyte, in comparison with that of RuO₂ (210 mV at 0.1 mA cm_ECSA⁻²) and commercial C-RuO₂ (217 mV at 0.1 mA cm_ECSA⁻²).

In addition, the turnover frequency (TOF) served as a pivotal metric for assessing catalyst activity. At an applied potential of 1.5 V, the TOF value of RuO₂/Nb₂O₅-0.6 reached 7.7 s⁻¹, representing an almost sevenfold increase compared to RuO₂ (1.7 s⁻¹) and tenfold enhancement in comparison to C-RuO₂ (0.7 s⁻¹), as illustrated in Fig. S9a (ESI†). Moreover, the mass activity of RuO₂/Nb₂O₅-0.6 was 296 A g⁻¹ at 1.5 V, surpassing that of RuO₂ (73 A g⁻¹) and C-RuO₂ (30 A g⁻¹), as depicted in Fig. S9b (ESI†). These significantly elevated TOF values and mass activities illustrated the enhanced intrinsic activity of RuO₂/Nb₂O₅-0.6 through the incorporation of Nb₂O₅.

The d-band of various catalysts, which was directly related to the strength of interaction between the Ru and guest molecules, was also determined by high-resolution valence-band (VB) XPS spectra. As shown in Fig. S10 (ESI†), our findings revealed that the d-band center of RuO₂/Nb₂O₅-0.6 was located at −1.53 eV, positioning it midway between C-RuO₂ (−1.55 eV) and RuO₂ (−1.48 eV). It was noteworthy that the closer proximity of the d-band center to the Fermi level implied a stronger binding strength between the catalyst and the adsorbed intermediates, which could be inconducive to their desorption.²⁶ Therefore, RuO₂/Nb₂O₅-0.6 showed a balanced adsorption capacity for these intermediates, ultimately contributing to its high activity.

Stability was another key parameter of electrocatalysts. The chronopotentiometry (CP) test was employed to assess the OER stability of RuO₂/Nb₂O₅-0.6, which exhibited high catalytic durability during a period of 750 h at a constant current density of 10 mA cm⁻² (Fig. 2c). The performance of RuO₂/Nb₂O₅-0.6 was superior to that of the majority of recently reported Ru-based electrocatalysts in acidic electrolytes (Table S3, ESI†). Negligible decay was observed for RuO₂/Nb₂O₅-0.6. Based on the results of the ICP-ES tests, the S-numbers (S = n_O2/n_Ru) of various catalysts at 10 mA cm⁻² were calculated according to a previous report.²⁷ Compared to that of C-RuO₂ (1.07 × 10³), the S-number of RuO₂/Nb₂O₅-0.6 was as high as 3.01 × 10⁵, indicating that the dissolution of Ru was significantly suppressed during the long-term operation. Characterization studies on the spent catalysts indicated the preserved structural features of RuO₂/Nb₂O₅-0.6 during the OER (Fig. S11–S13, ESI†).

These results demonstrated the high chemical and structural stability and suppressed Ru dissolution of RuO₂/Nb₂O₅-0.6 for the acidic OER.

Chemical state analysis

To understand the influence of Nb on the electronic structures of RuO₂/Nb₂O₅-0.6 and their roles in the catalytic activity, chemical states of Ru were analyzed by using X-ray photoelectron spectroscopy (XPS, Fig. 3). For RuO₂/Nb₂O₅-0.6, the peak with a binding energy of 462.6 eV could be categorized as the 3p XPS characteristic peak of Ru⁴⁺, which was shifted to the lower binding energy by 0.2 eV compared with those of RuO₂ and C-RuO₂ (at 462.8 eV).^28,29 Niobium 3d XPS spectroscopy showed that the RuO₂/Nb₂O₅-0.6 shifted to the lower binding energy by 0.5 eV (from 207.1 to 206.6 eV) compared to Nb₂O₅. The fitted deconvolution spectra of Nb contained 3d_5/2 and 3d_3/2 peaks in Nb₂O₅, which represented the Nb⁵⁺ states at 207.1 and 209.9 eV, respectively.³⁰ The Nb⁴⁺ peaks were also observed for RuO₂/Nb₂O₅-0.6 with a derived Nb⁴⁺/Nb⁵⁺ ratio of 0.92 for RuO₂/Nb₂O₅-0.6, indicating the presence of abundant oxygen vacancy in the as-synthesized RuO₂/Nb₂O₅-0.6 electrocatalysts. In addition, the O 1s XPS spectra of various electrocatalysts were deconvoluted into lattice oxygen (O_L), oxygen atoms adjacent to defects (O_V) and hydroxyls (O_H).³¹ The binding energy for lattice oxygen of the RuO₂/Nb₂O₅-0.6 was located at 530 eV, which was shifted towards higher binding energy by 0.5 eV compared to the synthetic RuO₂ and commercial C-RuO₂ (529.5 eV). The positive shift of the lattice oxygen indicated the transfer of electrons from lattice oxygen to metal sites, which reduced the oxidation state of Ru and Nb. The characteristic peaks located at 530.7 eV and 532.1 eV were attributed to O_V and O_H, respectively. The proportion of O_H of RuO₂/Nb₂O₅-0.6 (O_H/O_V = 1.49) was considerably higher than that of commercial C-RuO₂ (O_H/O_V = 0.94) and RuO₂ (O_H/O_V = 1.02).³² Furthermore, the O_H/O_V value of RuO₂/Nb₂O₅-0.6 was higher than that of RuO₂/Nb₂O₅-0.5 (0.68) and RuO₂/Nb₂O₅-0.7 (1.08). This finding illustrated that the binary electrocatalysts exhibited higher catalytic activity, accompanied by a larger O_H/O_V ratio, among catalysts with varying amounts of Nb₂O₅ introduced (Fig. S14 and Table S4, ESI†). The enhanced hydroxyl adsorption with defects at the active sites modulated the adsorption energy of the oxygen intermediates and enhanced the catalytic activity and stability of RuO₂/Nb₂O₅-0.6.

Fig. 3 XPS spectroscopies of catalysts. (a) Ru 3p XPS spectra of RuO₂/Nb₂O₅-0.6, RuO₂, C-RuO₂. (b) Nb 3d XPS spectra of RuO₂/Nb₂O₅-0.6, Nb₂O₅. (c) O 1s XPS spectra of RuO₂/Nb₂O₅-0.6, RuO₂, C-RuO₂.

Catalytic mechanism

To investigate the reaction mechanism of the binary RuO₂/Nb₂O₅ electrocatalysts, the pH-dependent tests of various catalysts were performed and compared (Fig. 4a). In contrast to the coupled proton–electron transfer step in the AEM pathway of bare RuO₂ with good crystallinity, the LOM pathway breaks this scaling relationship, where the proton transfer step becomes the potential-determining step (PDS).³³ Therefore, the pH-dependent kinetics of RuO₂/Nb₂O₅-0.6 sufficiently reflected the participation of lattice oxygen in the OER procedure, indicating the LOM pathway for the OER. When the relationship was further compared at 1.5 V, a linear relationship was fitted by a linear fit between the logarithm of the current density and the pH values. The proton reaction orders derived from the slope (ρ = (∂(log i)/∂pH)_E) were 0.8 and 0.4 for RuO₂/Nb₂O₅ and C-RuO₂, respectively, confirming the dominant LOM pathway of RuO₂/Nb₂O₅-0.6 for the acidic OER.³⁴ In order to probe the PDS of the binary electrocatalysts, isotope experiments with 0.5 M H₂SO₄ in the H₂O and D₂O as solvents were performed. The Tafel slope of RuO₂/Nb₂O₅ significantly increased from 84 mV dec⁻¹ in H₂O to 142 mV dec⁻¹ in D₂O, suggesting that the deprotonation process was PDS of the binary catalysts,³⁵ while the Tafel slopes of the commercial RuO₂ showed a minor change for the isotope experiments, indicating that the AEM pathway was dominant (Fig. 4b).

Fig. 4 Electrochemical mechanism tests of catalysts. (a) Current densities of RuO₂/Nb₂O₅-0.6 and C-RuO₂ at 1.5 V versus RHE as a function of the pH values of electrolytes. (b) Comparison of RuO₂/Nb₂O₅-0.6 and C-RuO₂ in 0.5 M H₂SO₄, where H₂O or D₂O was used as the solvent. (c) Operando FTIR spectra recorded in the potential range of 0.9–1.6 V vs. RHE for C-RuO₂. (d) Operando FTIR spectra recorded in the potential range of 0.9–1.6 V vs. RHE for RuO₂/Nb₂O₅-0.6.

To further reveal the OER reaction mechanism of the binary catalysts, operando Fourier transform infrared (FTIR) spectroscopy was employed to monitor the reaction intermediates. When the applied potential exceeded the reaction onset potential of the C-RuO₂ electrocatalysts (>1.3 V), a pair of the distinguishable absorption peaks located at 1050 cm⁻¹ and 1150 cm⁻¹ were observed (Fig. 4c). The pair absorption peaks were identified as OO*, which was an intermediate prior to the release of oxygen in the LOM approach. Generally, the absorption peak at 948 cm⁻¹ was attributed to OOH* as a reaction intermediate in the AEM.⁷ The dynamically enhanced peak intensity was observed, suggesting that the increase in the applied potential accelerated the generation of the reaction intermediates. The presence of the absorption peaks of OOH* and OO* in the C-RuO₂ catalysts was attributed to the coexistence of the LOM pathway and the AEM pathway in the acidic OER (Fig. 4c). However, only OO* adsorption peaks located at 1050 cm⁻¹ and 1180 cm⁻¹ were detected in the FTIR spectra of RuO₂/Nb₂O₅-0.6, suggesting that the catalytic pathway of the binary electrocatalysts was dominated by the LOM pathway for the acidic OER (Fig. 4d). The observation of OO* peaks at a lower potential of 0.9 V suggested high catalytic OER performance of the binary electrocatalysts with more reactive intermediates generated on the surface of RuO₂/Nb₂O₅-0.6. It has been widely recognized that surface Ru species of RuO₂ are converted to the higher oxidation state through the LOM pathway for the acidic OER, therefore leading to the formation of the soluble ruthenium species and subsequent decay of catalytic activity.

Next, the charge transfer that occurred between catalyst surfaces and electrolyte was investigated by electrochemical impedance spectroscopy (EIS) to understand the roles of Nb incorporated with RuO₂ for enhanced catalytic stability as well as activity. The Bode point plots of RuO₂/Nb₂O₅-0.6 and commercial C-RuO₂ with potentials from 1.1 V to 1.6 V are shown in Fig. S15 (ESI†), respectively. The characteristic peaks in the high frequency region (10¹–10³ Hz) were attributed to the surface oxidation of the catalyst. The characteristic peaks in the low-frequency region (10⁻¹–10¹ Hz) reflected the adsorption of reaction intermediates on the surface of the catalyst.³⁶ For the C-RuO₂ catalysts, the phase angle decreased slowly in the low-frequency range with increasing potential, indicating that C-RuO₂ had insufficient capacity to adsorb intermediates^37,38 whereas the significantly decreased phase angle of RuO₂/Nb₂O₅-0.6 in the low frequency region suggested that the binary electrocatalysts showed less resistance to adsorb intermediates at the same voltage. The peak shape of C-RuO₂ significantly changed over 1.4 V, suggesting the unstable surface of C-RuO₂ with the increased potentials. In contrast, the barely changed characteristic peaks of RuO₂/Nb₂O₅-0.6 in the high-frequency region with the increased potentials revealed more stable surface oxidation states of Ru at higher potentials. The comparative results demonstrated high catalytic activity and stability of the binary RuO₂/Nb₂O₅-0.6 electrocatalysts for the acidic OER.

To further elucidate the enhanced catalyst stability, CV curves were recorded to examine the oxidation states of the surface Ru species during the acidic OER (Fig. 5a). Two sets of redox peaks located around 0.6 V and 1.25 V were observed for the C-RuO₂ electrocatalysts, which represented Ru^3+/4+ and Ru^4+/6+ redox pairs, respectively.³⁹ Thus, the surface Ru species of C-RuO₂ could be oxidized into high valent Ru species, leading to the serious over-oxidation of catalysts and formation of soluble Ru species during the acidic OER.⁴⁰ Thus, it's not surprising to observe the poor catalytic stability of C-RuO₂. Comparatively, only the redox peaks of Ru^3+/4+ was observed for RuO₂/Nb₂O₅-0.6, indicating that the introduction of Nb effectively inhibited the over-oxidation of Ru in the catalyst and stabilized the binary electrocatalysts for the acidic OER. Afterwards, the chemical states of two catalysts before and after the acidic OER were analyzed by XPS. For the RuO₂/Nb₂O₅-0.6 catalyst, the binding energy of Ru 3p barely changed after 10 h of reaction at a current density of 10 mA cm⁻². In contrast, the Ru 3p binding energy of C-RuO₂ increased from 462.8 eV of the fresh electrocatalysts to 463 eV of the spent ones, indicating the over-oxidation of Ru species of C-RuO₂ (Fig. S16, ESI†). As shown in the XPS spectra of Nb 3d of RuO₂/Nb₂O₅-0.6 before and after the acidic OER, the peaks of Nb were negatively shifted by 0.2 eV, suggesting the formation of oxygen vacancies at the Nb site (Fig. 5b).⁴¹ The electron transfer pathway in the LOM was based on a redox-catalyzed reaction of oxygen, in which lattice oxygen was electrochemically activated in the form of released oxygen intermediates, following the pathway OH⁻–(OO)²⁻–O₂, essentially a hybridization of neighbouring non-bonded oxygen atoms (O⁻), with no transfer of electrons to the external circuit.⁴² Thus, it's logical to predict the occurrence of spillover of O* intermediates from RuO₂ to Nb₂O₅ and the participation of lattice oxygen in Nb₂O₅ for the acidic OER, like in the previous reports.¹⁸ In this way, the OO* intermediate was formed on Nb₂O₅ instead of RuO₂ for the steps of the release of oxygen, generation of oxygen vacancies and rehabilitation of lattice oxygen. Therefore, the over-oxidation and dissolution of surface Ru species of the binary RuO₂/Nb₂O₅ electrocatalysts could be suppressed, which was consistent with the EIS profile (Fig. S15, ESI†) and CV results (Fig. 5a).

Fig. 5 Electrochemical stability analysis and evidence of oxygen spillover phenomena. (a) CV curves of RuO₂/Nb₂O₅-0.6, Nb₂O₅ and C-RuO₂ measured from 0.1 to 1.45 V vs. RHE. (b) Nb 3d XPS spectra of RuO₂/Nb₂O₅-0.6 before and after the stability test. Operando Raman spectra recorded in the potential range of 0.9–1.4 V vs. RHE for (c) C-RuO₂ and (d) RuO₂/Nb₂O₅-0.6.

To investigate the potential of oxygen spillover in the binary electrocatalysts, oxygen temperature-programmed desorption (O₂-TPD) profiles were recorded for Nb₂O₅ and RuO₂ (Fig. S17, ESI†).⁴³ The desorption peak of oxygen from RuO₂ was observed at a significantly higher temperature of 431 °C compared to 303 °C for Nb₂O₅, suggesting that RuO₂ requires more energy for O₂ desorption than Nb₂O₅. The O₂-TPD results indicated that the formation of oxygen vacancies on Nb₂O₅ was more facile than RuO₂. Based on these findings, the thermodynamic feasibility of oxygen spillover involving the transfer of intermediate O* species from RuO₂ to Nb₂O₅ and the participation of lattice oxygen in Nb₂O₅ for the acidic OER, was established.

To experimentally confirm the occurrence of oxygen spillover, in situ Raman spectroscopy was used to dynamically monitor the surface M–O species of the RuO₂/Nb₂O₅-0.6 and C-RuO₂ catalysts at different external potentials. The characteristic peaks of Ru–O in C-RuO₂ were located at 520 cm⁻¹, 640 cm⁻¹ and 705 cm⁻¹, which represented the E_g, A_1g and B_2g vibrational modes of Ru–O, respectively (Fig. 5c).^22,44 Since the characteristic peak of E_g was more obvious and the peak position was easier to be distinguished, the characteristic E_g peak of Ru–O of C-RuO₂ was mainly analyzed. As shown in Fig. 5c, the E_g peak of C-RuO₂ gradually shifted to the higher wavenumber by 6 cm⁻¹ (from 520 cm⁻¹ to 526 cm⁻¹) with the increased external applied voltage from the open-circuit potential to 1.6 V, indicating that the surface Ru–O bonds in C-RuO₂ were gradually compressed and the valence state of Ru was elevated progressively with the increased potentials.⁴⁵ As for RuO₂/Nb₂O₅-0.6, the characteristic peak of Ru–O was barely shifted with the increased potential, whereas the characteristic peak of Nb–O showed a redshift of 15 cm⁻¹ (from 766 cm⁻¹ to 751 cm⁻¹), indicating the significantly stretched Nb–O of the binary electrocatalysts (Fig. 5d).¹⁷ The reduced oxidation state of Nb combined with the LSV and CV curves showed that Nb suppressed the elevation of the oxidation state of Ru during the oxygen evolution reaction through the modification of bonding with oxygen. Therefore, compared to C-RuO₂, the over-oxidation of Ru in RuO₂/Nb₂O₅-0.6 was noticeably suppressed, leading to the stabilized surface Ru species. Significantly, as the circuit was disconnected after applying 1.6 V for a period of time, the characteristic E_g peak of Ru–O in RuO₂/Nb₂O₅-0.6 gradually recovered to the initial state (522 cm⁻¹). Compared with the pristine binary electrocatalyst, the characteristic Raman peak of Nb–O also gradually restored to the same position before the reaction (763 cm⁻¹, Fig. S18, ESI†). These shifts further illustrated the structural stability of RuO₂/Nb₂O₅-0.6 in the acidic OER. Combined with the results of CV curves, XPS and in situ FT-IR, in situ Raman spectra strongly suggested the reactive oxygen species spilled over from RuO₂ to Nb₂O₅, consistent with the preserved E_g of Ru–O and shifted 15 cm⁻¹ of Nb–O in the binary RuO₂/Nb₂O₅ electrocatalysts.

Based on all of the above analysis and experimental evidence, a catalytic pathway is proposed for RuO₂/Nb₂O₅-0.6 with high catalytic activity and stability through the LOM pathway (Fig. 6). Initially, water molecules are absorbed and activated on the surface Ru sites of the binary electrocatalysts to generate OH* species. After the deprotonation of OH*, the O* intermediates spill over from Ru sites to Nb sites. With the participation of the lattice oxygen of Nb₂O₅, OO* intermediates are formed on Nb sites, followed by the release of O₂ associated with the generation of oxygen vacancies and rehabilitation of lattice oxygen on Nb₂O₅. Such a catalytic pathway of RuO₂/Nb₂O₅ effectively suppresses the over-oxidation of Ru species and avoids the formation of soluble high valent Ru species, leading to the high catalytic activity and stability for the acidic OER despite undergoing the LOM mechanism.

Fig. 6 Schematic overview of the proposed oxygen spillover for the OER on the surface of RuO₂/Nb₂O₅-0.6 catalysts.

Conclusions

In summary, we presented the highly active and robust binary RuO₂/Nb₂O₅ electrocatalysts for the acidic OER, delivering a low overpotential of 176 mV at 10 mA cm⁻² and a high stability of at least 750 h in 0.5 M H₂SO₄. Incorporation of Nb₂O₅ with RuO₂ enabled the spillover of active O* intermediate from RuO₂ to Nb₂O₅. Thus, the participation of lattice oxygen of Nb₂O₅ instead of RuO₂ gave the OO* intermediate for the release of oxygen, which effectively inhibited the over-oxidation of surface Ru species and thereby significantly improved the catalytic stability of the binary electrocatalysts. Extensive experimental investigations verified the proposed oxygen spillover reaction pathway for the enhanced catalytic performance of the binary electrocatalysts. This work provided seminal insights into solving the stability problem of Ru-based catalysts and offered a methodology for the rational design of high performance electrocatalysts for the acidic OER.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

The authors acknowledge the Basic and Applied Basic Research Foundation of Guangdong Province (2023A1515012288) and the Fundamental Research Funds for the Central Universities (D5000210829). Yongquan Qu acknowledges the Fundamental Research Funds for the Central Universities (D5000210601).

References

1. Q. Wang, Y. Cheng, H. B. Tao, Y. Liu, X. Ma, D. S. Li, H. B. Yang and B. Liu, Angew. Chem., Int. Ed., 2023, 62, e202216645.
2. L. Li, G. Zhang, C. Zhou, F. Lv, Y. Tan, Y. Han, H. Luo, D. Wang, Y. Liu, C. Shang, L. Zeng, Q. Huang, R. Zeng, N. Ye, M. Luo and S. Guo, Nat. Commun., 2024, 15, 4974.
3. X. Rong, J. Parolin and A. M. Kolpak, ACS Catal., 2016, 6, 1153–1158.
4. L. Li, P. Wang, Q. Shao and X. Huang, Adv. Mater., 2021, 33, 2004243.
5. C. Roy, R. R. Rao, K. A. Stoerzinger, J. Hwang, J. Rossmeisl, I. Chorkendorff, Y. Shao-Horn and I. E. L. Stephens, ACS Energy Lett., 2018, 3, 2045–2051.
6. N. Hodnik, P. Jovanovič, A. Pavlišič, B. Jozinović, M. Zorko, M. Bele, V. S. Šelih, M. Šala, S. Hočevar and M. Gaberšček, J. Phys. Chem. C, 2015, 119, 10140–10147.
7. C. Lin, J.-L. Li, X. Li, S. Yang, W. Luo, Y. Zhang, S.-H. Kim, D.-H. Kim, S. S. Shinde, Y.-F. Li, Z.-P. Liu, Z. Jiang and J.-H. Lee, Nat. Catal., 2021, 4, 1012–1023.
8. R. Y. Fan, Y. S. Zhang, J. Y. Lv, G. Q. Han, Y. M. Chai and B. Dong, Small, 2024, 20, e2304636.
9. L. Zhou, Y. Shao, F. Yin, J. Li, F. Kang and R. Lv, Nat. Commun., 2023, 14, 7644.
10. Z. Niu, Z. Lu, Z. Qiao, S. Wang, X. Cao, X. Chen, J. Yun, L. Zheng and D. Cao, Adv. Mater., 2024, 36, 2310690.
11. W. Gou, Y. Wang, M. Zhang, X. Tan, Y. Ma and Y. Qu, Chin. J. Catal., 2024, 60, 68–106.
12. Z. Feng, C. Dai, P. Shi, X. Lei, R. Guo, B. Wang, X. Liu and J. You, Chem. Eng. J., 2024, 485, 149992.
13. Z. Wang, W. A. Goddard and H. Xiao, Nat. Commun., 2023, 14, 4228.
14. Y. Wu, R. Yao, Q. Zhao, J. Li and G. Liu, Chem. Eng. J., 2022, 439, 135699.
15. Y. Xue, J. Fang, X. Wang, Z. Xu, Y. Zhang, Q. Lv, M. Liu, W. Zhu and Z. Zhuang, Adv. Funct. Mater., 2021, 31, 2101405.
16. J. Wang, C. Cheng, Q. Yuan, H. Yang, F. Meng, Q. Zhang, L. Gu, J. Cao, L. Li, S.-C. Haw, Q. Shao, L. Zhang, T. Cheng, F. Jiao and X. Huang, Chem, 2022, 8, 1673–1687.
17. Z. Shi, J. Li, J. Jiang, Y. Wang, X. Wang, Y. Li, L. Yang, Y. Chu, J. Bai, J. Yang, J. Ni, Y. Wang, L. Zhang, Z. Jiang, C. Liu, J. Ge and W. Xing, Angew. Chem., Int. Ed., 2022, 61, e202212341.
18. W. Gou, S. Zhang, Y. Wang, X. Tan, L. Liao, Z. Qi, M. Xie, Y. Ma, Y. Su and Y. Qu, Energy Environ. Sci., 2024, 17, 6755–6765.
19. X. Tan, M. Zhang, D. Chen, W. Li, W. Gou, Y. Qu and Y. Ma, Small, 2023, 19, e2303249.
20. J. Okal and M. Zawadzki, Appl. Catal., B, 2009, 89, 22–32.
21. K. V. K. Rao and L. Iyengar, Acta Crystallogr., Sect. A, 1969, 25, 302–303.
22. A. V. Korotcov, Y. S. Huang, K. K. Tiong and D. S. Tsai, J. Raman Spectrosc., 2007, 38, 737–749.
23. A. L. Paulsen, F. Borup, R. W. Berg and S. Boghosian, J. Phys. Chem. A, 2010, 114, 7485–7493.
24. M.-R. Gao, M. K. Y. Chan and Y. Sun, Nat. Commun., 2015, 6, 7493.
25. J. Chen, G. Liu, Y.-Z. Zhu, M. Su, P. Yin, X.-J. Wu, Q. Lu, C. Tan, M. Zhao, Z. Liu, W. Yang, H. Li, G.-H. Nam, L. Zhang, Z. Chen, X. Huang, P. M. Radjenovic, W. Huang, Z.-Q. Tian, J.-F. Li and H. Zhang, J. Am. Chem. Soc., 2020, 142, 7161–7167.
26. J. He, X. Zhou, P. Xu and J. Sun, Adv. Energy Mater., 2021, 11, 2102883.
27. S. Geiger, O. Kasian, M. Ledendecker, E. Pizzutilo, A. M. Mingers, W. T. Fu, O. Diaz-Morales, Z. Li, T. Oellers, L. Fruchter, A. Ludwig, K. J. J. Mayrhofer, M. T. M. Koper and S. Cherevko, Nat. Catal., 2018, 1, 508–515.
28. P. Singh and M. S. Hegde, Chem. Mater., 2009, 21, 3337–3345.
29. J. Fu, K. Yang, C. Ma, N. Zhang, H. Gai, J. Zheng and B. H. Chen, Appl. Catal., B, 2016, 184, 216–222.
30. K. Islam, R. Sultana, B. Satpati and S. Chakraborty, Vacuum, 2022, 195, 110675.
31. Y. Zhang, Y. Zhang, Z. Zeng and D. Ho, Mater. Horiz., 2023, 10, 2904–2912.
32. H. Liu, Z. Zhang, J. Fang, M. Li, M. G. Sendeku, X. Wang, H. Wu, Y. Li, J. Ge, Z. Zhuang, D. Zhou, Y. Kuang and X. Sun, Joule, 2023, 7, 558–573.
33. N. Zhang and Y. Chai, Energy Environ. Sci., 2021, 14, 4647–4671.
34. D. Antipin and M. Risch, Electrochem. Sci. Adv., 2023, 3, e2100213.
35. W. Li, F. Li, H. Yang, X. Wu, P. Zhang, Y. Shan and L. Sun, Nat. Commun., 2019, 10, 5074.
36. W. Gou, Z. Xia, X. Tan, Q. Xue, F. Ye, S. Dai, M. Zhang, R. Si, Y. Zou, Y. Ma, J. C. Ho and Y. Qu, Nano Energy, 2022, 104, 107960.
37. R. Šimpraga, G. Tremiliosi-Filho, S. Y. Qian and B. E. Conway, J. Electroanal. Chem., 1997, 424, 141–151.
38. B. Zhou, Y. Li, Y. Zou, W. Chen, W. Zhou, M. Song, Y. Wu, Y. Lu, J. Liu, Y. Wang and S. Wang, Angew. Chem., Int. Ed., 2021, 60, 22908–22914.
39. N. Cong, Y. Han, L. Tan, C. Zhai, H. Chen, J. Han, H. Fang, X. Zhou, Y. Zhu and Z. Ren, J. Electroanal. Chem., 2021, 881, 114955.
40. K. Klyukin, A. Zagalskaya and V. Alexandrov, J. Phys. Chem. C, 2019, 123, 22151–22157.
41. K. K. Rahangdale and S. Ganguly, Phys. B, 2022, 626, 413570.
42. X. Wang, H. Zhong, S. Xi, W. S. V. Lee and J. Xue, Adv. Mater., 2022, 34, 2107956.
43. Y. Wang, J. Wu, G. Wang, D. Yang, T. Ishihara and L. Guo, Appl. Catal., B, 2021, 285, 119873.
44. Y. S. Huang and F. H. Pollak, Solid State Commun., 1982, 43, 921–924.
45. A. Rahman, S. Mondal, M. Modak, A. Singh, N. S. Thayat, H. Singh, J. K. Clegg, H. K. Poswal, V. Haridas and S. P. Thomas, Small, 2024, 2402120.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4nh00437j

‡ These authors contributed equally to this study.

---