English

• Proton exchange membrane fuel cells (PEMFCs) are regarded as one of the most promising technologies converting the hydrogen to electrical energy in the hydrogen community [1]. Nevertheless, the commercialization process of PEMFCs is severely hindered by the rare and high-cost platinum group metal (PGM) catalysts used for boosting the sluggish oxygen reduction reaction (ORR) in cathode [2,3]. As an alternative technology, anion exchange membrane fuel cells (AEMFCs) have attracted widespread attention because the use of platinum group metal-free (PGM-free) catalysts is permitted in the alkaline electrolyte, which could significantly reduce fuel cell cost to approach DOE target (US$ 30/kW) [4]. Although many non-precious metal catalysts with Pt-like ORR performance have been developed over the past decades, the inherently 2–3 orders of magnitude slower kinetics of anodic hydrogen oxidation reaction (HOR) than that in acidic media bring about new challenge for the development of high-performance PGM-free HOR electrocatalysts [5].

  Currently, most research on non-precious metal HOR catalysts has been focused on nickel-based materials, due to the earth abundance and considerable reactivity to activate hydrogen [6,7]. However, the too strong hydrogen binding energy (HBE) on pristine Ni metal leads to substantial energy barrier of H₂O formation in the Volmer step, thus restricting the practical application of monometallic Ni-based electrocatalysts [7]. To solve this problem, alloying with other metals has been adopted to regulate the electronic structure of nickel, thereby weakening its hydrogen adsorption strength towards enhanced HOR performance in alkaline condition. As expected, the developed NiMo [8], MoNi₄ [9] and WNi₄, NiCu [10], CoNiMo [11], and Ni_5.2WCu_2.2 [12] show improved HOR activity than the pristine Ni. Besides HBE mechanism, Koper and McCrum proposed that hydroxyl species also played a vital role in alkaline HOR process and emphasized the balanced adsorption of H* and OH* on active sites [13]. Thus, some Ni/metal oxide heterostructures with the optimized OH* adsorption capacity, such as Ni/MoO₂ [14], Ni/NiO [15], Ni/V₂O₃ [16] have been reported to exhibit excellent HOR activity. Despite these advances, the intrinsic activity of Ni-based catalysts is much inferior to those of PGM-based catalysts and still needs to be improved. At the same time, the mechanism of basic HOR remains controversial due to lack of direct experimental evidence for detecting the dynamic adsorption of intermediates. It calls for the in situ monitoring the dynamic change process of the reaction intermediates and active sites of the catalyst under working conditions.

  Motivated by this challenge, we herein constructed Ni/NbO_x heterostructure catalyst with controllable oxygen vacancy content as model catalyst. The presence of oxygen vacancy can not only optimize OH* adsorption strength, but also tailor the electronic structure of Nb and Ni towards regulated binding strength of H* intermediates. Such a regulation contributes to accelerated rate-determining Volmer step and thus facilitating the whole HOR process. Combining in-situ Raman spectroscopy and density functional theory (DFT) calculations, we monitored the process that the bias-induced hydroxyl species adsorption on oxygen vacancy. Specifically, the dynamic binding of OH* with Nb⁴⁺ is triggered upon applying a certain potential, which would react with H* adsorbed on Ni surface to complete HOR process. As a result, the optimal Ni/NbO_x catalyst reaches the HOR diffusion limiting current at a small overpotential of only 60 mV and exhibits intrinsic exchange current density of 0.036 mA/cm², which is much higher than pure Ni and exceeds most PGM-free alkaline HOR catalysts reported. This work provides a new perspective on the rational design of alkaline HOR catalysts.

  The urchin-like Ni/NbO_x microsphere catalyst was synthesized via a two-step hydrothermal growth and high temperature pyrolysis process (see more experimental details in Supporting information). As shown in Fig. 1a, we firstly synthesized the NbO_x-NiC₂O₄ heterostructure precursor on nickel foam (NF) through a hydrothermal process. The X-ray diffraction (XRD) spectrum and scanning electron microscopy (SEM) images indicate that the NiC₂O₄NbO_x precursor was successfully prepared (Fig. S1 in Supporting information). Then the precursor was annealed at 500 ℃ to form Ni/NbO_x heterostructure (Fig. S2 in Supporting information). And we also prepared catalysts with different Ni/Nb ratios by adjusting the amount of Nb, with the lower amount of Nb added being named Ni/NbO_x-L and the higher amount of Nb being named Ni/NbO_x-H. In addition, Ni or Nb₂O₅ was synthesized by the same method as that of Ni/NbO_x sample, except that no niobium source or NF was added during the hydrothermal process (Fig. S3 in Supporting information). The XRD spectra prove the formation of Ni/NbO_x heterogeneous structure. In Fig. 1b, all the Ni/NbO_x samples exhibit the typical diffraction peaks at 44.5°, 51.5° and 76.4° corresponding to (111), (200), (220) plane of metallic Ni (PDF #04–0850), as well as 22.6°, 28.6°, 36.7° corresponding to the (001), (100), (101) crystal plane of Nb₂O₅ (PDF #28–0317), respectively. Moreover, from Ni/NbO_x-L to Ni/NbO_x-H, as the relative content of nickel decreases, the diffraction peaks' intensity corresponding to Ni decreases and the diffraction peaks intensity of Nb₂O₅ increase, which is in line with the energy-dispersive X-ray (EDX) data and inductively coupled plasma atomic emission spectroscopy (ICP-AES) measurements (Fig. S4 and Table S1 in Supporting information). All the above experimental results prove the successful preparation of different Ni/Nb ratio Ni/NbO_x heterogeneous structure.

  Figure 1

  

  Figure 1.  (a) Schematic illustration of the synthesis of Ni/NbO_x. (b) XRD spectra of Ni/NbO_x-L, Ni/NbO_x, Ni/NbO_x-H. (c) TEM images of Ni/NbO_x. (d) SAED of Ni/NbO_x. (e) HRTEM image of Ni-NbO_x. (f-i) IFFT images and lattice fringes corresponding to Ni and NbO_x in (e), respectively. (j) EDX elemental mapping of Ni/NbO_x.

  DownLoad: Full-Size Img PowerPoint

  The sea urchin-like structure composed of nanowires could be observed via the transmission electron microscopy (Fig. 1c and Fig. S2). High-resolution transmission electron microscope (HRTEM) and inverse fast Fourier transform (IFFT) images clearly display abundant heterogeneous interface with resolved lattice fringes of 0.382 nm and 0.214 nm, which correspond to Nb₂O₅ (001) and Ni (111) plane, respectively (Figs. 1e-i). In addition, the selected area electron diffraction (SAED) image also displays the different planes ascribed to Ni (111), Ni (220), Ni (222) and Nb₂O₅ (001) for Ni/NbO_x sample (Fig. 1d). Noted that the lattice spacing corresponding to the Ni (111) surface at Ni/NbO_x interface is larger than the lattice spacing 0.203 nm corresponding to the standard Ni (111), which is because the heterogeneous interface is composed of Ni (111) and Nb₂O₅ (001) planes. The standard lattice spacing of Nb₂O₅ (001) is 0.393 nm, which is larger than that of Ni (111), so the lattice of Ni (111) expands and becomes larger than the standard value of Ni (111) plane spacing. Similarly, the lattice spacing of Nb₂O₅ (001) has shrunk to 0.382 nm, which is smaller than the standard value of Nb₂O₅ (001). Energy-dispersive X-ray (EDX) elemental mappings of Ni/NbO_x reveal that Ni, Nb and O elements are homogeneously distributed in the nanorods (Fig. 1j).

  To further investigate the electronic states of Ni/NbO_x, we adopted X-ray photoelectron spectroscopy (XPS) to explore the electronic interaction between Ni and Nb element. Fig. S5 (Supporting information) shows the survey spectrum of the as-prepared catalyst, consisting of peaks of Ni 2p, Nb 3d and O 1s for all the Ni/NbO_x catalysts, whereas there are no signal of Ni element in Nb₂O₅. As shown is Fig. 2a, Ni 2p spectrum of Ni/NbO_x consists of three types of peaks ascribed to Ni⁰, Ni²⁺ and satellites in Ni 2p_3/2 region, respectively. The high valence Ni²⁺ is attributed to the inevitable surface oxidation of Ni at the nanoscale. And the Ni 2p_1/2 region also reveals the co-existence of Ni⁰, Ni²⁺ and satellites. In addition, Ni/NbO_x exhibits a positive shift in Ni 2p binding energy and a higher Ni²⁺/Ni⁰ ratio compared with that of pure Ni, indicating that Ni valence state in Ni/NbO_x heterostructure increase after introducing NbO_x. Accordingly, Nb 2p binding energy of Ni/NbO_x displays a negative shift compared with Nb₂O₅ sample, evidencing the electn transfer from Ni to NbO_x (Fig. 2b) [17]. Ultraviolet photoelectron spectroscopy (UPS) was carried out to further support this result. As is shown in Fig. 2c, NbO_x displays a higher work function (3.95 eV) compared with Ni (3.50 eV), corresponding to a lower Fermi level (E_F) of NbO_x than that of Ni. Thus, the transfer of electron from Ni to Nb₂O₅ could be achieved, which is consistent with the above XPS results [18]. Furthermore, the O 1s orbital spectra could be assigned to three components for Nb₂O₅. The peaks at 529.92, 530.83 and 532.86 eV are attributed to lattice oxygen (O_lat), adsorption oxygen (O_ads) and adsorbed water (H₂O), respectively [19]. Whereas there is a new peak assigned to oxygen vacancy (O_v) at 531.77 eV emerges for Ni/NbO_x, which demonstrates the existence of oxygen vacancy (Fig. 2d). And the oxygen vacancy content increases with the increase of Nb content, the order of O_v content is Ni/NbO_x-L < Ni/NbO_x < Ni/NbO_x-H, (Fig. S6 in Supporting information), which is 20.22%, 35.14% and 51.06%, respectively. This conclusion is also supported by EPR results (Fig. S7 in Supporting information). And the electron paramagnetic resonance (EPR) was also carried out to further monitor the signal of oxygen vacancy (Fig. 2e), only Ni/NbO_x displays obvious signal at g = 2.003. It was noted that the oxygen vacancy content increased significantly after the formation of the heterostructure, which indicated that the oxygen vacancy was located at the Ni/NbO_x heterogeneous interface, as shown in Fig. 2f. The formation of oxygen vacancy at the interface between Ni and NbO_x is due to the lattice mismatch of Ni and Nb element thereby the coordination numbers between adjacent Ni and Nb atom decrease. Moreover, the content of oxygen vacancy increased with the increase of NbO_x content, which was attributed to the increase of the number of Ni/NbO_x interface, so that the control of oxygen vacancy content was achieved [20,21].

  Figure 2

  

  Figure 2.  (a) Ni 2p spectrum of Ni and Ni/NbO_x. (b) Nb 3d spectrum of Ni/NbO_x and Nb₂O₅. (c) UPS spectrum of Ni and NbO_x, (d) O 1s of XPS spectrum of as-prepared samples. (e) EPR spectrum of Ni, Ni/NbO_x and Nb₂O₅. (f) The schematic diagram of Ni/NbO_x structure.

  DownLoad: Full-Size Img PowerPoint

  We evaluated the HOR performance of as-prepared catalysts in H₂-saturated 0.1 mol/L KOH solution via standard three-electrode system, using the catalysts grown on nickel foam (NF) with a geometric surface area of ~1 cm⁻² as the working electrode directly. And iR-correction (i, current; R, resistance) was conducted for all the electrochemical data reported here for the uncompensated Ohmic drop. Linear sweep voltammetry (LSV) of Ni/NbO_x exhibit minimal and negligible current in N₂-saturated 0.1 mol/L KOH, whereas obvious current is observed in H₂-saturated 0.1 mol/L KOH, which indicates the increased current comes from H₂ oxidation (Fig. S8 in Supporting information). To explore the intrinsic activity of our catalysts, we obtained corresponding powder from NF substrates through ultrasonication treatment and then loaded them on glassy carbon rotating-disk electrode (RDE) to evaluate their HOR performance. As shown in Fig. 3a, both single Ni and Nb₂O₅ show extremely poor HOR activity, which is consistent with previous research [7]. However, the current of all the Ni/NbO_x catalysts sharply increase in 0.1 mol/L KOH solution, indicating the excellent HOR performance. In addition, there is a volcano-type relationship between Ni/Nb ratio and the HOR activity and the optimal Ni/NbO_x catalyst could reach the diffusion-limiting current at only 60 mV, which even surpasses the state-of-the-art commercial Pt/C catalyst from the kinetic to the diffusion-limiting regions (Fig. 3b). In addition, we noticed that, unlike Ni-based catalysts, which are easily inactivated by oxidation, the Ni/NbO_x catalyst still does not show significant HOR current decay at 0.2 V, and the HOR activity decreases significantly when the potential is close to 0.3 V, indicating that the antioxidant capacity of Ni is improved, which is due to the effective regulation of OH* on the Ni surface by oxygen vacancy (Fig. S9 in Supporting information) [22,23]. To explore the HOR kinetic properties of all the catalysts, we recorded the polarization curves of catalysts at varied rotating speeds from 900 rpm to 2500 rpm (Fig. 3c). The plotted j⁻¹ and ω^−1/2 exhibits a linear function with a slope calculated to be 13.48 cm² mA⁻¹ rpm^−1/2, which is in line with the theoretical value of 14.8 cm² mA⁻¹ rpm^−1/2 of the two-electron HOR process (Fig. S9 in Supporting information) [24,25]. Then we calculated the kinetic current density (j_k) at 50 mV for Ni/NbO_x from the Koutecky–Levich equation, then normalized to mass activity. As shown in Fig. S10 (Supporting information), the optimized Ni/NbO_x perform a high mass activity (j_{k, m}) of 60.5 mA/mg^, far more than pristine Ni. To assess the intrinsic HOR activity of these catalysts, we obtained the exchange current density (j₀) through the linear-fitting of micro-polarization region and fitting of the Butler–Volmer equation in Tafel region (Figs. 3d and e, Tables S2 and S3 in Supporting information). We then tested the electrochemically active surface area (ECSA) of catalysts through integrating the charge of desorption of hydroxyl species by performing cyclic voltammetry (CV) in 0.1 mol/L Ar-saturated KOH solution (Figs. S12 and S13 in Supporting information), and calculated their ECSA-normalized exchange current density (j_{0, s}). As is shown in Fig. 3f, the j_{0, s} of Ni/NbO_x is 0.036 mA/cm², which is higher than pure Ni and the most PGM-free alkaline HOR catalyst reported so far (Fig. 3i).

  Figure 3

  

  Figure 3.  (a) HOR polarization curves of Ni, Nb₂O₅, Ni/NbO_x and Pt/C—Com in H₂-saturated 0.1 mol/L KOH at a rotation rate of 1600 rpm with a scan rate of 1 mV/s. (b) HOR polarization curves of Ni/NbO_x-L, Ni/NbO_x, Ni/NbO_x-H. (c) HOR polarization curves of Ni/NbO_x at varied rotational speeds. (d) Tafel plots of freshly synthesized catalysts. (e) Micro-polarization region fitting of as-prepared catalysts. (f) Comparison of ECSA-normalized exchange current densities (j_{0, s}) of Ni, Ni/NbO_x-L, Ni/NbO_x and Ni/NbO_x-H. (g) HOR polarization curves of Ni/NbO_x before and after durability test, the embedded column chart is the comparison of current density at 50 mV before and after 2000 cycles. (h) Comparison of HOR polarization of Ni/NbO_x and Pt/C—Com with and without the presence of CO. (i) Comparisons of mass activity and specific activity of Ni/NbO_x with that of other Ni-based catalysts reported recently.

  DownLoad: Full-Size Img PowerPoint

  In order to further explore the potential for practical applications, the accelerated durability test (ADT) was carried out. As shown in Fig. 3g, the HOR performance of Ni/NbO_x maintains well after 2000 cycles cyclic voltammetry in H₂-saturated 0.1 mol/L KOH solution, with only minimal current loss of 6% at 50 mV (the inset bar plot), indicating the remarkable stability of Ni/NbO_x catalyst. Another evaluation indicator of HOR catalyst for practical utilization is CO tolerance since the anode catalyst is easily poisoned by even trace of CO impurity in the reforming hydrogen [26-28]. To our excitement, the Ni/NbO_x catalyst exhibits excellent HOR performance even in the 1000 ppm CO/H₂ mixture with only a small attenuation of 13% in HOR current, whereas commercial Pt/C catalyst almost lost 72.5% current in the same condition (Fig. 3h and Fig. S14 in Supporting information). This result indicates the Ni/NbO_x possess excellent CO tolerance. After the accelerated durability test, there was no significant change in the Ni 2p spectrum of Ni/NbO_x catalyst, which indicated that the valence state of Ni remained stable (Fig. S15a in Supporting information). The O 1s spectrum shows that the proportion of oxygen vacancy is close to that of the fresh catalyst after the durability test, indicating the stability of the oxygen vacancy (Fig. S15b in Supporting information). EPR results further support this conclusion, the EPR curve after the 2000 cycles CV test almost coincided with the initial curve, indicating that there was no significant change in the oxygen vacancy in the Ni/NbO_x catalyst (Fig. S16 in Supporting information). CO-TPD was carried out to investigate the reasons for increased CO tolerance, the CO deposition temperature of Ni/NbO_x is much lower than that of commercial Pt/C, indicating weaker CO adsorption of Ni/NbO_x than that of commercial Pt/C, which could reduce CO poisoning effect on catalyst (Fig. S17 in Supporting information).

  Although the decisive mechanism of alkaline HOR remains controversial, recent studies generally agree that HBE and OHBE are important descriptors of HOR reaction in alkaline media [19,28-31]. Thus, various characterizations and density functional theory (DFT) were conducted to further explore the distinction of HBE and OHBE between Ni/NbO_x and other catalysts. The differential charge density distribution results reveal electron transfer from Ni to NbO_x support, and the reinforced interfacial interaction could redistribute d electron thereby regulating d band structure and adsorption properties (Fig. 4a). The H* adsorption strength was first evaluated by H₂ temperature program deposition (H₂-TPD). The results indicate the main H₂ deposition peak of Ni and Ni/NbO_x is located at about 324 ℃ and 293 ℃, respectively, proving weaker hydrogen binding strength of Ni/NbO_x than that of Ni (Fig. S18 in Supporting information). Therefore, NbO_x could regulate the electronic structure of Ni metal so that decrease the HBE of metal Ni after the formation of Ni/NbO_x heterostructure. And DFT results further supported this view, the HBE value on Ni/NbO_x is calculated to be −0.42 eV, which is positive than that of Ni metal (Fig. 4e). Previous study considered that HBE of metal Ni was too strong, resulting in too high water formation energy [24], so more positive HBE of Ni/NbO_x could accelerate Volmer step thereby facilitating the whole HOR process.

  Figure 4

  

  Figure 4.  (a) The differential charge density distribution results of Ni/NbO_x. (b) The Raman spectrum of Ni/NbO_x in N₂-saturated 0.1 mol/L KOH solution. (c) The Raman spectrum of Ni/NbO_x in H₂-saturated 0.1 mol/L KOH solution. (d) The projected density of states of NbO_x and Nb₂O₅. (e) HBE and OHBE of Ni, Ni/NbO_x. (f) Reaction pathways of Ni/NbO_x and Ni for alkaline HOR. (g) The relationship between HOR activity, OHBE, oxygen vacancy content. (h) The illustration of H* and OH* adsorption on Ni, Ni/NiO_x and Nb₂O₅, respectively.

  DownLoad: Full-Size Img PowerPoint

  In addition, we also noted that the adsorption strength of OH* on our catalysts is much stronger than that of H*, so we can reasonably speculate that catalysts are more likely to preferentially adsorb OH* rather than H* species (Table S4 in Supporting information). As a result, the adsorption strength of OH* would greatly affect H* adsorption and subsequent Volmer step. Thus, the proper adsorption of hydroxyl may be more important than that of hydrogen for enhanced HOR performance. And numerous studies suggested that hydroxyl adsorption strength play a critical role in the whole alkaline HOR process recently [32-34]. To gain deer understanding of the critical role of hydroxyl in alkaline HOR, we carried out in situ surface-enhanced Raman scattering spectroscopy (in situ SERS) to monitor reaction process in real time. Spectroscopic signals were collected in N₂-or H₂-saturated 0.1 mol/L KOH solution. There exist obvious Ni-O and Nb-O peaks in the Raman spectroscopy of Ni and Nb₂O₅, respectively (Fig. S19 in Supporting information). Compared to Nb₂O₅, the Nb-O peaks in Ni/NbO_x become wider and redshift to lower wavenumbers, which is due to the more oxygen vacancies in the Ni/NbO_x sample, so the Nb-O peak of Ni/NbO_x was split into Nb⁴⁺-O and Nb⁵⁺-O peaks, and the Raman shift of the Nb⁴⁺-O peak was lower than that of the Nb⁵⁺-O peak, thereby showing that the Nb-O bond was broadened and redshifted than that of Nb₂O₅. For Ni/NbO_x, when tested in N₂-saturated solution, the Nb-O peak blue shift to higher wavenumber with an increase in the potential in the range of 0–0.15 V, which represents the dynamic process by which OH* fill in oxygen vacancies and Nb⁴⁺-O is transformed into Nb⁵⁺-O. From 0.2 V to 0.4 V, the Nb-O peak remains stable, there appear obvious Ni-O peak simultaneously, which means that the oxygen vacancies are saturated with OH*, and OH* begins to attack the Ni surface to form Ni-O (Fig. 4b). However, when tested in H₂-saturated solution, there is no significant change in Nb-O peak signal from 0 to 0.2 V, which is ascribed to HOR process that consumes OH* adsorbed on oxygen vacancy in the low potential range. Above 0.2 V, the gradual oxidation of nickel surface hinders the dissociation and adsorption of hydrogen, and the OH* filled in the vacancy is not easily consumed, so the Nb-O peak is gradually blue-shifted thereby HOR activity begin to decay (Fig. 4c). The reason for the enhanced stability of Ni/NbO_x catalyst can also be explained by this idea. The deactivation of Ni-based HOR catalysts is due to oxidation of Ni. The in-situ Raman results showed that the Ni-O peak occurred at a potential of 0.05 V for pure nickel, indicating that Ni is oxidized at 0.05 V (Fig. S19a in Supporting information). However, for Ni/NbO_x catalyst, a weak Ni-O peak does not begin to appear until 0.2 V (Fig. 4b), which means that the antioxidant capacity of Ni/NbO_x catalyst is improved thereby showing excellent stability.

  All the Raman spectroscopy results demonstrate that OH* and oxygen vacancies are involved in the alkaline HOR process and DFT results further demonstrate the significant role of oxygen vacancy for alkaline HOR performance. Projected density of state (PDOS) results demonstrates that the d band center of Nb downshift from 1.392 eV to 0.376 eV after introducing oxygen vacancy (Fig. 4d). According to D-band theory, a lower D-band center (far away from Fermi level) leads to more filling antibonding states with electron which means weaker adsorption between surface adsorbed species and catalysts [35]. As a result, the value of OHBE on oxygen vacancy is calculated to be −4.23 eV, which is more positive than that of Nb₂O₅ (Fig. 4e), proving that oxygen vacancy could optimize the adsorption of hydroxyl thereby accelerate the HOR reaction. Indeed, the formation energy of H₂O* decrease from 1.12 eV to 0.49 eV after introducing NbO_x, which is attributed to optimized OH* binding strength on oxygen vacancy. On the contrary, the adsorption of OH* on Nb₂O₅ and Ni is too strong and too weak, respectively, which is not conducive to the occurrence of subsequent steps (Figs. 4f and h). This result can be explained by Sabatier principle, that is highly active catalyst should have moderate adsorption strength for the intermediates. Specifically, if the OHBE of catalyst is too weak, very small amounts of OH* adsorbed on catalyst surface are hardly difficult to affect the entire HOR process. On the contrary, excess OH* may occupy the adsorption site of the hydrogen and thus poison the catalyst if the OHBE is too strong [36].

  To further support the above view, we explore the relationship between OHBE and oxygen vacancy content of Ni/NbO_x, as PDOS results demonstrate that oxygen vacancy can modulate d-band center of Nb thereby regulate OH* adsorption. So we took CO stripping experiment to determine OH* adsorption strength because the enhanced OH* adsorption could boost removal of CO adsorbed on catalyst surface [12]. According to CO stripping curves, the CO stripping potential peak of Ni/NbO_x-L, Ni/NbO_x, Ni/NbO_x-H are located at 0.636 V, 0.642 V and 0.665 V, respectively (Fig. S20 in Supporting information). The CO oxidation potential is more positive, the OH* adsorption strength is weaker [37]. Thus, the order of OHBE is Ni/NbO_x-L > Ni/NbO_x > Ni/NbO_x-H. Whereas the order of O_v content is Ni/NbO_x-L < Ni/NbO_x < Ni/NbO_x-H. Thus, we could confirm that OHBE and oxygen vacancy content were negatively correlated and we attributed the varied OHBE to different oxygen vacancy level in the as-prepared catalysts, which is in line with PDOS results. Combined with HOR activity regularity, we established a volcan-shape relationship between HOR performance and OHBE (Fig. 4g), indicating that appropriate OHBE is essential for the HOR process, further support the above results. Take together, combining experimental results and DFT calculations, we have explained the mechanism of the increase in Ni/NbO_x activity, that is, the precise hydroxyl adsorption strength is customized through oxygen vacancies, and the oxygen vacancies are very easy to bind to hydroxyl groups at the operating voltage, and the hydroxyl groups adsorbed at oxygen vacancies are easy to react with H*, thus facilitating rate-determining Volmer step in alkaline HOR process.

  In summary, we demonstrate Ni/NbO_x catalyst with well-controlled oxygen vacancy towards enhanced HOR performance. Benefit from well-regulated OH* adsorption strength via oxygen vacancy, the formation energy of water is significantly reduced. As a result, the optimal Ni/NbO_x catalyst exhibits exciting HOR performance with a high j_{0, s} of 0.036 mA/cm², which is 4-fold higher than that of single Ni. Notably, we also take in situ Raman spectroscopy to monitor the dynamics of the catalyst during the reaction and demonstrate that the OH* adsorbed on oxygen vacancy could easily remove adjacent H* adsorbed on Ni under working potential. This work reveals the vital role of OH* in alkaline HOR process and in situ evolution during the reaction, emphasizes moderate OH* binding strength contribute to alkaline HOR performance, which provides new insight for designing highly active HOR catalysts in alkaline environment.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Guo Yang: Writing – review & editing, Writing – original draft, Formal analysis, Data curation, Conceptualization. Kai Li: Resources, Investigation. Hanshi Qu: Validation. Jianbing Zhu: Resources. Chunyu Ru: Visualization. Meiling Xiao: Writing – review & editing. Wei Xing: Writing – review & editing, Funding acquisition. Changpeng Liu: Writing – review & editing, Supervision, Funding acquisition.

  Acknowledgments

  This work was supported by Jilin Province Science and Technology Development Program (Nos. 20200201001JC, 20210502002ZP, 20230101367JC, 20220301011GX) and Jilin Province Science and Technology Major Project (No. 222648GX0105103875).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110150.

  
  
• 1. [1]

     I. Staffell, D. Scamman, A. Velazquez Abad, et al., Energy Environ. Sci. 12 (2019) 463–491. doi: 10.1039/c8ee01157e

  2. [2]

     Y. Jiao, Y. Zheng, M. Jaroniec, et al., Chem. Soc. Rev. 44 (2015) 2060–2086.

  3. [3]

     Y. Wang, D.F. Ruiz Diaz, K.S. Chen, et al., Mater 32 (2020) 178–203.

  4. [4]

     Y. Yang, C.R. Peltier, R. Zeng, et al., Chem. Rev. 122 (2022) 6117–6321. doi: 10.1021/acs.chemrev.1c00331

  5. [5]

     B.P. Setzler, Z. Zhuang, J.A. Wittkopf, et al., Nat. Nanotechnol. 11 (2016) 1020–1025. doi: 10.1038/nnano.2016.265

  6. [6]

     M. Wang, H. Yang, J. Shi, et al., Angew. Chem, Int. Ed. 60 (2021) 5771–5777. doi: 10.1002/anie.202013047

  7. [7]

     X. Zhao, X. Li, L. An, et al., Energy Environ. Sci. 15 (2022) 1234–1242. doi: 10.1039/d1ee03482k

  8. [8]

     S. Kabir, K. Lemire, K. Artyushkova, et al., J. Mater. Chem. A 5 (2017) 24433–24443.

  9. [9]

     Y. Duan, Z.Y. Yu, L. Yang, et al., Nat. Commun. 11 (2020) 4789.

  10. [10]

      A. Roy, M.R. Talarposhti, S.J. Normile, et al., Sustain. Energy Fuels 2 (2018) 2268–2275. doi: 10.1039/c8se00261d

  11. [11]

      W. Sheng, A.P. Bivens, M. Myint, et al., Energy Environ. Sci. 7 (2014) 1719–1724.

  12. [12]

      S. Qin, Y. Duan, X.L. Zhang, et al., Nat. Commun. 12 (2021) 2686.

  13. [13]

      I.T. McCrum, M.T.M. Koper, Nat. Energy 5 (2020) 891–899. doi: 10.1038/s41560-020-00710-8

  14. [14]

      S. Deng, X. Liu, X. Guo, et al., J. Energy Chem. 54 (2021) 202–207.

  15. [15]

      Y. Yang, X. Sun, G. Han, et al., Angew. Chem. Int. Ed. 58 (2019) 10644– 10649. doi: 10.1002/anie.201905430

  16. [16]

      Y. Duan, X.L. Zhang, F.Y. Gao, et al., Angew. Chem. Int. Ed. 62 (2023) e202217275.

  17. [17]

      N. Yao, R. Meng, F. Wu, et al., Appl. Catal. B: Environ. 277 (2020) 119282.

  18. [18]

      L. Yang, S. Hou, S. Zhu, et al., ACS Catal. 12 (2022) 13523–13532. doi: 10.1021/acscatal.2c04158

  19. [19]

      F. Yang, X. Bao, P. Li, et al., Angew. Chem. Int. Ed. 58 (2019) 14179–14183. doi: 10.1002/anie.201908194

  20. [20]

      J. Zhang, J. Qian, J. Ran, et al., ACS Catal. 10 (2020) 12376–12384. doi: 10.1021/acscatal.0c03756

  21. [21]

      J. Yin, Y. Li, F. Lv, et al., ACS Nano 11 (2017) 2275–2283. doi: 10.1021/acsnano.7b00417

  22. [22]

      Y. Cong, F. Meng, H. Wang, et al., J. Energy Chem. 83 (2023) 255–263.

  23. [23]

      Y. Yang, Y. Huang, S. Zhou, et al., J. Energy Chem. 72 (2022) 395–404.

  24. [24]

      Z. Zhuang, S.A. Giles, J. Zheng, et al., Nat. Commun. 7 (2016) 10141.

  25. [25]

      J. Zheng, S. Zhou, S. Gu, et al., J. Electrochem. 163 (2016) F499. doi: 10.1149/2.0661606jes

  26. [26]

      G. Shi, H. Yano, D.A. Tryk, et al., ACS Catal. 7 (2017) 267–274. doi: 10.1021/acscatal.6b02794

  27. [27]

      F. Song, W. Li, J. Yang, et al., ACS Energy Lett. 4 (2019) 1594–1601. doi: 10.1021/acsenergylett.9b00738

  28. [28]

      Y. Zhou, Z. Xie, J. Jiang, et al., Nat. Catal. 3 (2020) 454–462.

  29. [29]

      D. Strmcnik, M. Uchimura, C. Wang, et al., Nat. Chem. 5 (2013) 300–306. doi: 10.1038/nchem.1574

  30. [30]

      Y. Qiu, L. Xin, Y. Li, et al., J. Am. Chem. Soc. 140 (2018) 16580–16588. doi: 10.1021/jacs.8b08356

  31. [31]

      W. Ni, A. Krammer, C.S. Hsu, et al., Angew. Chem. Int. Ed. 58 (2019) 7445– 7449. doi: 10.1002/anie.201902751

  32. [32]

      S. Zhu, X. Qin, F. Xiao, et al., Nat. Catal. 4 (2021) 711–718. doi: 10.1038/s41929-021-00663-5

  33. [33]

      Y.H. Wang, X.T. Wang, H. Ze, et al., Angew. Chem. Int. Ed. 60 (2021) 5708– 5711. doi: 10.1002/anie.202015571

  34. [34]

      J. Li, S. Ghoshal, M.K. Bates, et al., Angew. Chem. Int. Ed. 56 (2017) 15594–15598. doi: 10.1002/anie.201708484

  35. [35]

      Z. Chen, Y. Song, J. Cai, et al., Angew. Chem. Int. Ed. 57 (2018) 5076–5080. doi: 10.1002/anie.201801834

  36. [36]

      M. Xu, S. He, H. Chen, et al., ACS Catal. 7 (2017) 7600–7609. doi: 10.1021/acscatal.7b01951

  37. [37]

      Y. Men, X. Su, P. Li, et al., J. Am. Chem. Soc. 144 (2022) 12661–12672. doi: 10.1021/jacs.2c01448

• 

  Figure 1  (a) Schematic illustration of the synthesis of Ni/NbO_x. (b) XRD spectra of Ni/NbO_x-L, Ni/NbO_x, Ni/NbO_x-H. (c) TEM images of Ni/NbO_x. (d) SAED of Ni/NbO_x. (e) HRTEM image of Ni-NbO_x. (f-i) IFFT images and lattice fringes corresponding to Ni and NbO_x in (e), respectively. (j) EDX elemental mapping of Ni/NbO_x.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) Ni 2p spectrum of Ni and Ni/NbO_x. (b) Nb 3d spectrum of Ni/NbO_x and Nb₂O₅. (c) UPS spectrum of Ni and NbO_x, (d) O 1s of XPS spectrum of as-prepared samples. (e) EPR spectrum of Ni, Ni/NbO_x and Nb₂O₅. (f) The schematic diagram of Ni/NbO_x structure.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) HOR polarization curves of Ni, Nb₂O₅, Ni/NbO_x and Pt/C—Com in H₂-saturated 0.1 mol/L KOH at a rotation rate of 1600 rpm with a scan rate of 1 mV/s. (b) HOR polarization curves of Ni/NbO_x-L, Ni/NbO_x, Ni/NbO_x-H. (c) HOR polarization curves of Ni/NbO_x at varied rotational speeds. (d) Tafel plots of freshly synthesized catalysts. (e) Micro-polarization region fitting of as-prepared catalysts. (f) Comparison of ECSA-normalized exchange current densities (j_{0, s}) of Ni, Ni/NbO_x-L, Ni/NbO_x and Ni/NbO_x-H. (g) HOR polarization curves of Ni/NbO_x before and after durability test, the embedded column chart is the comparison of current density at 50 mV before and after 2000 cycles. (h) Comparison of HOR polarization of Ni/NbO_x and Pt/C—Com with and without the presence of CO. (i) Comparisons of mass activity and specific activity of Ni/NbO_x with that of other Ni-based catalysts reported recently.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) The differential charge density distribution results of Ni/NbO_x. (b) The Raman spectrum of Ni/NbO_x in N₂-saturated 0.1 mol/L KOH solution. (c) The Raman spectrum of Ni/NbO_x in H₂-saturated 0.1 mol/L KOH solution. (d) The projected density of states of NbO_x and Nb₂O₅. (e) HBE and OHBE of Ni, Ni/NbO_x. (f) Reaction pathways of Ni/NbO_x and Ni for alkaline HOR. (g) The relationship between HOR activity, OHBE, oxygen vacancy content. (h) The illustration of H* and OH* adsorption on Ni, Ni/NiO_x and Nb₂O₅, respectively.

  下载: 全尺寸图片 幻灯片
•