Coordination Effect-Promoted Durable Ni(OH)2 for Energy-Saving Hydrogen Evolution from Water/Methanol Co-Electrocatalysis

Highlights A novel Ni(OH)2-based catalyst with ultralow Ni–Ni coordination is produced, exhibiting high activity (100 mA cm−2 at 1.39 V for methanol oxidation reactions) and outstanding stability in an industrial concentration electrolyte (over 500 mA cm−2 in 6 M KOH). Mechanistic studies show that the improved kinetics and durability are primarily due to ultralow Ni–Ni coordination, 3D-networking structures and the Mo dopant. Supplementary Information The online version contains supplementary material available at 10.1007/s40820-022-00940-3.


Introduction
With the rising environmental problems and depletion of fossil fuels, the development and utilization of renewable energies have drawn extensive attention [1]. As clean and renewable energy, hydrogen is considered a potential replacement for fossil fuels owing to its high energy density and wide availability [2]. Electrocatalytic water splitting is a viable technique for obtaining hydrogen [3]. However, it is uneconomical to obtain hydrogen in a traditional electrocatalytic water splitting way. In the process, the cathode catalyzes hydrogen evolution reactions (HER) to produce hydrogen gas (H 2 ), while the anode catalyzes oxygen evolution reactions (OER) to produce oxygen (O 2 ) with high overpotential, leading to a low-economic-value anode product (O 2 ). To solve these problems, the geometry and electronic structures of electrocatalysts have been widely studied to exploit low overpotential catalysts and improve economic efficiency. Such studies include heteroatom doping [4], defect modulation [5], tension modulation [6], surface self-reconstruction [7], metal-atom escape [8], heterointerface construction [9], tunable pore structures [10]. However, the overpotential of the best OER catalyst is still high at around 150 mV, whereas the overpotential of HER catalysts is close to 0 mV [11].
Alternatively, some novel anodic reactions have been used to replace OER, coupling HER reaction for energy-saving hydrogen production [12]. This kind of reaction is attractive because it does not involve OER, which is limited by the theoretical potential of 1.23 V. Most anodic reactions involve simple organic molecules, such as methanol [6,13], ethanol [14,15], glycerol [16,17], urea [11,18,19], amine [20], furfural [21] and 5-hydroxymethylfurfural [22,23] with low working potentials. Compared with traditional OER, anodic reactions based on simple organic molecule oxidation reduce the working voltage and exhibit better reaction kinetics. As the simplest alcohol, methanol can be easily produced by chemical or biomass industrial synthesis [24]. The synthesis of methanol is cheap compared to that of other organic matter (about 350 € per tonne). Furthermore, methanol exhibits very high solubility in water, and methanol oxidation reactions (MOR), which generate value-added formic acid and formate salts (about 539 € per tonne), have fast kinetics [25]. Thus, MOR is an ideal reaction to replace OER. Based on this, a novel electrolyzer can produce H 2 and formate salts with low energy.
For Ni-based materials, mechanistic studies have shown that Ni 2+ is oxidized to Ni 3+ species, which is a real active site for catalyzing methanol to formate [50]. Notably, surface evolution of Ni-based materials, such as oxidation, hydroxylation and reconstruction, usually occurs in the presence of an electrolyzer [34]. This phenomenon can achieve the aforementioned oxidation (Ni 2+ → Ni 3+ ), making them electrocatalytically active. In contrast, excessive evolution can deactivate the electrocatalysts [44]. The durability of Nibased materials is still a challenge in alkaline hydrogenation evolution, especially at high current density in an industrial concentration (6 M KOH), which limits their applications. Thus, there is a need to develop durable Ni-based electrocatalysts at high current density in alkaline media with the industrial concentration.
In this study, we propose a facile strategy to enhance the kinetics and durability of Ni-based electrocatalysts by synthesizing Mo-doped 3D-networking Ni(OH) 2 catalyst with ultralow Ni-Ni coordination from a NiMoO 4 ·0.75 H 2 O precursor. The obtained electrocatalyst shows excellent MOR activity and high selectivity for value-added formate, especially at high current density in an industrial concentration. A current density of 100 mA cm −2 at 1.39 V is achieved for MOR, delivering 28 mV dec −1 for the Tafel slope. An assembled two-electrode electrolyzer generates 500 mA cm −2 at a cell 1 3 voltage of 2.00 V with 90% Faradaic efficiency. Furthermore, electrolyzer operates for 50 h in an industrial concentration electrolyte (6 M KOH) without obvious deterioration. Mechanistic studies based on density functional theory (DFT) calculations and X-ray absorption spectroscopy (XAS) reveal that the improved kinetics and durability are mainly attributed to the (1) ultralow Ni-Ni coordination, which induces porosity in the structure, increasing the contact area and facilitating the reaction; (2) 3D-networking structures, which increase the density of the active sites; (3) uncompleted dissolution Mo, which strengthens the 3D-networking framework. This study paves a new way for designing electrocatalysts with enhanced activity and durability for industrial energy-saving hydrogen production.

Preparation of Anode and Cathode Electrocatalysts
The anode and cat hode electrocat alysts are LC-Ni(OH) 2 ·xH 2 O and Ni 4 Mo, respectively. To prepare of LC-Ni(OH) 2 ·xH 2 O, 3.2 g NaOH and 1 g (NH 4 ) 2 S 2 O 8 were first dissolved in 16 and 8 mL DI water, respectively, to create solutions. Following that, both solutions were added in 18 mL DI water to generate a chemical reconstruction solution. Subsequently, the prepared precursor NiMoO 4 ·0.75 H 2 O nanorod arrays were immersed in the solution for 30 min before being washed with DI water and dried at 40 °C to obtain LC-Ni(OH) 2 ·xH 2 O. In addition, samples with reconstruction times of 10 and 20 min were also prepared and named as LC-Ni(OH) 2 ·xH 2 O-10 and LC-Ni(OH) 2 ·xH 2 O-20, respectively. For the preparation of Ni 4 Mo, the NiMoO 4 ·xH 2 O nanorod arrays were heated at 500 ℃ for 2 h in a H 2 (5%)/Ar (95%) atmosphere and then obtained the Ni 4 Mo electrocatalyst.

Preparation of Ni(OH) 2
Typically, a hydrothermal process is used to create Ni(OH) 2 on Ni foam. In this instance, 60 mL of DI water was used to dissolve 1.732 g of NiCl 2 ·6H 2 O and 1.1 g of urea. To create a homogeneous solution, the aqueous solution was thoroughly stirred. One piece of the treated Ni foam was then placed into the prepared solution, transferred to a stainless steel autoclave lined with Teflon and kept at 120 °C for 12 h. The finished product was thoroughly cleaned and repeatedly sonicated with DI water and ethanol after naturally cooling to room temperature. The sample was then dried overnight at 60 °C in the air.

Characterization
The morphologies of electrodes were characterized by field-emission scanning electron microscopy (FESEM, SU-70) and field-emission transmission electron microscopy (FETEM, JEM-F200). In addition, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and the Inductively Coupled Plasma-Optical Emission Spectrometer (ICP-OES) were used for detailed analysis. The XRD patterns were recorded using the Bruker D8 Advance (Cu Kα, 50 kV and 360 mA). The XPS was conducted using a Thermo Scientific™ K-Alpha™ + spectrometer equipped with a monochromatic Al Kα X-ray source (1486.6 eV) operating at 100 W. Samples were analyzed under vacuum (P < 10 −8 mbar) with a pass energy of 150 eV (survey scans) or 25 eV (high-resolution scans). All peaks were calibrated with C 1 s peak binding energy at 284.8 eV for adventitious carbon. The ICP-OES was recorded by Agilent 5110. Furthermore, Raman spectra and XAS were used to deeply study the fine structure of the synthesized electrocatalysts. For Raman spectra, it was recorded by Horiba: HR Evolution with a 633-nm laser, whereas the XAS was collected by employing synchrotron radiation light source at BL12B2 beam line of the National Synchrotron Radiation Research Center (NSRRC) in SPring 8 (Japan) at room temperature. The detailed analysis of the XAS is listed in supporting information.

Electrochemical Measurements
All electrochemical measurements were taken with a electrochemical workstation (CHI760E, CH instruments Inc., Shanghai) at room temperature. The electrochemical measurements were taken in a three-electrode system. The catalyst-loaded Ni foam (1 × 1 cm 2 ) was used as working electrode, and the Hg/HgO was used as reference electrode. Pt sheet (1 × 1 cm 2 ) was used as a counter electrode, while Scheme 1 Preparation of low coordination Ni(OH) 2 ·xH 2 O nanorod arrays and their assembly of the electrolytic cell coupling HER and MOR a graphite rod was used in case of HER to avoid potential contamination of Pt. 1 M KOH with or without 0.5 M methanol was used as electrolytes, and polarization curves were collected at a scan rate of 5 mV s −1 . The potential drop (iR) loss due to the solution/system resistance was applied according to the equation: E corr = E mea -iR. All potentials presented in this work were calibrated to the reversible hydrogen electrode (RHE) according to the equation: E RHE = E Hg/HgO + 0.059pH + 0.098. Doublelayer capacitance (C dl ) of the as-prepared electrode was measured by cyclic voltammetry in a potential range of 0.978 − 1.078 V vs RHE at scan rate of 10 − 50 mV s −1 . A large current density MOR experiment was carried out in a two-electrode system. With 6 M KOH and 3 M methanol serving as the electrolyte, the LC − Ni(OH) 2 ·xH 2 O loaded on Ni foam (1 × 0.5 cm 2 ) was employed as the anode, while the Ni 4 Mo served as the cathode.

Product Analysis
Ion chromatography (Technology Co. Ltd., Qingdao, China), which was equipped with organic anion columns containing the leachate of 2.4 mmol Na 2 CO 3 and 6 mmol NaHCO 3 , was employed for the quantification of products from electrochemical oxidation of methanol. Before chronoamperometry measurements, 100 uL of electrolyte was collected and diluted with DI water with a ratio of 1:100. The measurement of each sample was repeated three times, and the concentration of formate ion was calibrated based on standard solutions with known concentrations. The detailed calculations for the Faradaic efficiency (FE) and energy consumptions are listed in the supporting information.

Materials Characterization
The SEM, high-resolution TEM (HRTEM) and energy-dispersive X-ray spectroscopy (EDS) are performed to investigate the changes in the morphology and composition of the NiMoO 4 ·0.75H 2 O precursor and LC-Ni(OH) 2 ·xH 2 O nanorod array. Notably, the transition products from the precursors to the final electrocatalysts, namely LC-Ni(OH) 2 ·xH 2 O-10 and LC-Ni(OH) 2 ·xH 2 O-20, which are chemically constructed in 10 and 20 min, are also analyzed in detail to determine the chemical evolution process. The SEM images of the NiMoO 4 ·0.75 H 2 O precursor ( Fig. S1a-b) show wellcrystallized cuboids with sizes ranging from 0.5 to 1.5 μm and lengths of ~ 10 μm. As chemical reconstitution proceeds, the structure of the micron cuboids could be observed, but the previously smooth surface changes to a crude surface for LC-Ni(OH) 2 ·xH 2 O-10 ( Fig. S2), LC-Ni(OH) 2 ·xH 2 O-20 ( Fig. S3) and LC-Ni(OH) 2 ·xH 2 O ( Fig. 1a-b). HRTEM reveals that a stubbly substance grows on the surface of the electrocatalyst (Fig. 1c). With higher magnification, the lattice fringes become irregular, and dark spots could be observed (Fig. 1d), indicating that the electrocatalyst becomes more amorphous due to chemical remodeling processes. The EDS element mapping images (Fig. 1e-h) show a decrease in the Mo element content, which is lower than that of NiMoO 4 ·0.75H 2 O (Fig. S1), LC-Ni(OH) 2 ·xH 2 O-10 ( Fig. S2) and LC-Ni(OH) 2 ·xH 2 O-20 (Fig. S3). It indicates that most of the Mo atoms are dissolved. However, some small amount Mo may be still in the LC-Ni(OH) 2 ·xH 2 O, acting as dopants. As a result, the precursor NiMoO 4 ·0.75H 2 O has evolved into LC-Ni(OH) 2 ·xH 2 O with small amount of Mo as dopants.
The chemical reconstruction is also studied by XPS and ICP-OES. As shown in Fig. 2a (Fig. 1d).
Furthermore, Raman spectroscopy is conducted to analyze the precursor, transition products and final electrocatalyst. The Raman spectra (Fig. 2f)

Electrochemical Performance
In the presence of 0.5 M methanol, the polarization curves of LC-Ni(OH) 2 ·xH 2 O, Ni(OH) 2 , NiMoO 4 ·0.75H 2 O and Ni foam are compared in 1 M KOH. The onset potential of Ni(OH) 2 ·xH 2 O against reversible RHE for methanol oxidation processes is 1.35 V, which is better than that of NiMoO 4 ·0.75H 2 O and Ni(OH) 2 (Fig. 3a). Furthermore, at a potential of 1.43 V (vs. RHE), LC-Ni(OH) 2 ·xH 2 O shows a current density of 200 mA cm −2 , which is 192 mV lower than that of traditional OER (Fig. 3b), demonstrating the high potential of LC-Ni(OH) 2 ·xH 2 O for energy-saving hydrogen production. Notably, an oxidation peak appeared at 1.39 V vs. RHE in OER (Fig. 3b), which can be assigned to the oxidation of Ni 2+ to Ni 3+ [62]. To verify this, a cycle voltammetric measurement of LC-Ni(OH) 2 ·xH 2 O is taken (Fig. S5), which confirms the observation. Figure 3c reveals that the Tafel slope for LC-Ni(OH) 2 ·xH 2 O is 28 mV dec −1 , which is substantially slower than that of NiMoO 4 ·0.75 H 2 O (64 mV dec −1 ) and Ni(OH) 2 (40 mV dec −1 ), showing that the dissolution of Mo species significantly accelerates the kinetics of MOR.
To further examine the electrode kinetics and the ionic and charge transport resistance of MOR, electrochemical impedance spectroscopy (EIS) is used at 1.37 V (vs. RHE) in 1 M KOH and 0.5 M methanol, with the associated Nyquist plots (Fig. 3d). The equivalent circuit is composed of a resistor representing the Ohmic resistance (R s ) and a parallel combination, including a resistor reflecting the charge transfer resistance (R ct ) and a constant-phase element (CPE-1) [63]. Compared to LC-Ni(OH) 2   The current density declines from ~ 130 to 95 mA cm −2 due to methanol consumption, as shown in Fig. 3f. After refreshing the electrolyte, the current density increases back to ~ 130 mA cm −2 without obvious deterioration, showing its high potential for energy-saving hydrogen production. There is no significant difference in the LSV curves of LC-Ni(OH) 2 ·xH 2 O before and after the stability test (Fig. 3g). In addition, the SEM image ( Fig. S6a-b), Raman shift (Fig. S6c) and ICP-OES (Fig. S6d) after the stability test also show no significant change, which confirms that it exhibits excellent stability. Ion chromatography (Fig. S7) is used to study the Faradaic efficiency of MOR catalyzed by LC-Ni(OH) 2 ·xH 2 O at various potentials. The Faradaic efficiency declines from 100% (1.4 V versus RHE) to 95% (1.8 V vs. RHE) as the potential increased (Fig. 3h). This can be attributed to the significant likelihood of OER competition. Further, the double-layer capacitance (C dl ) measured by cyclic voltammetry (Fig. S8) is employed to compare the electrochemical active surface area (ECSA). However, ECSA = C dl /C s , where C s is a constant and a common factor. So, it is reasonable to use C dl to replace ECSA in the qualitative comparison of intrinsic activity. Ni(OH) 2 has a much higher C dl (2.99 mF cm −2 ) than LC-Ni(OH) 2 •xH 2 O (1.85 mF cm −2 ) (Fig. S9). However, the C dl normalized activity of LC-Ni(OH) 2 ·xH 2 O is higher than that of Ni(OH) 2 , indicating that it has a much higher density of active sites (Fig. 3i). We compare its performance to that of recently reported catalysts for methanol/water co-electrolysis with higher current density at 100 mA cm −2 , and we find that LC-Ni(OH) 2 ·xH 2 O is similar to the best HCl-modified Ni(OH) 2 and more active than others, including NiSe, NiFe, NiMoO 4 and Ni 3 S 2 (Fig. 3j).

Electronic Structure and Mechanisms Analysis
HRTEM and XRD reveal that the synthesized catalyst has an amorphous structure, and XPS reveals that the amorphous structure has the similar structure as Ni(OH) 2 , but its morphologies and XRD peaks do not match those of Ni(OH) 2 . However, it outperforms Ni(OH) 2 ; thus, there is a need to understand the structure of the catalyst to accelerate the high-performance Ni-based electrocatalysts design. Figure 4a shows the normalized X-ray absorption nearedge spectroscopy (XANES) of NiMoO 4 ·0.75H 2 O and LC-Ni(OH) 2  The Ni 2+ species showed a pre-edge peak at 8333 eV, which is attributed to the 1 s → 3d quadrupole transition [64]. In addition, NiMoO 4 ·0.75H 2 O and NiO showed characteristic peaks at 8354 and 8365 eV, respectively. Thus, the curves for LC-Ni(OH) 2 ·xH 2 O are very similar to that of Ni(OH) 2 , indicating that they have similar properties, which is in well agreement with the XPS observations. Figure 4b shows the Fourier transformed curves of the extended X-ray absorption fine structure (EXAFS) in R space. There is a peak at 1.66 Å for LC-Ni(OH) 2 2 shows a characteristic peak at 5.8 Å, which could be assigned to the second shell of the Ni-Ni path, whereas LC-Ni(OH) 2 ·xH 2 O shows no peak at 5.8 Å. This further indicates the unique Ni-Ni coordination structure for this novel electrocatalyst.
Wavelet transform EXAFS (wt-EXAFS) of the electrocatalysts is shown in Fig. 4c-e, which show four, three and two shells, respectively. The maximum for these samples is at ~ 2.81 Å on the R-axis and at 9.27, 8.21 and 7.33 Å −1 on the k-axis. The peak of LC-Ni(OH) 2 ·xH 2 O at 2.81 Å shifts from the high-to the low-k region compared with that of the other two species, indicating that the decrease in the peak (Fig. 4b) is attributed to the absence of paths for heavy atoms, such as Mo and Ni, rather than light atoms, such as O. This further indicates the dissolution of Mo species and the low coordination number of the Ni-Ni path.
Based on physical and electronic characterizations, we propose the chemical construction mechanisms (Fig. 5a) (Fig. S10). Thus, the crystalline structure contains uncertain crystalline H 2 O molecules, making it amorphous. As a result, XRD could not obtain the crystalline structure. Furthermore, because Mo cannot be completely dissolved, residual Mo remains as a dopant in the form of Mo-O-Ni (Fig. 5a). Thus, the newly formed catalyst can be deduced as Mo-doped Ni(OH) 2 containing uncertain crystalline water.
To verify this deduction, we perform simulations using these models and the EXAFS data. Surprisingly, the experimental EXAFS results are well reproduced (Fig. S11 and Table S3). The signals for the Ni-Mo bonds disappear during the processes, which is in agreement with the ICP-OES, XPS, Raman and EXAFS results, indicating that Mo species in NiMoO 4 ·0.75H 2 O are selectively dissolved, forming a novel nickel hydroxide species. Compared with the wellcrystallized Ni(OH) 2 with a Ni-Ni coordination number of 6, LC-Ni(OH) 2 ·xH 2 O showed an ultralow Ni-Ni coordination number of 1.5 (Table S3 and Fig. S12). In addition, according to the ICP and XPS results, approximately 1 to 2% Mo remained in the crystalline structure. Because Mo is linked to Ni in the NiMoO 4 ·0.75H 2 O structure, we infer that it links with Ni in the novel Ni(OH) 2 structures through bridged oxygen. Based on these results, we conclude that a new Mo-doped Ni(OH) 2 containing uncertain crystalline water, namely, LC-Ni(OH) 2 ·xH 2 O, is formed.
Crystal orbital Hamilton populations (COHPs) are analyzed based on DFT calculations to further understand the performance enhancement. As shown in Fig. 5b, COHP describes the bonding and antibonding of Ni-O bonds in Ni(OH) 2 , LC-Ni(OH) 2 ·2.5H 2 O and LC-Ni(OH) 2 ·2.75H 2 O. As the number of H 2 O molecules increases, the composition of antibonding orbitals below the Fermi level decreases, lowering the system's energy. This means that the addition of a water molecule strengthens the entire Ni-O bonds of the system, making the catalyst more stable. For comparison, we calculated COHP for the active Ni sites, i.e., the Ni-O bonds (Ni-OH 2 ) formed by Ni and H 2 O molecules for NiMoO 4 and LC-Ni(OH) 2 ·2.75H 2 O (Fig. 5c) since they are loosely bonded and more easily attacked by OH − groups than other Ni-O bonds (Table S4). Surprisingly, the composition of the Ni-O antibonding orbitals in either NiMoO 4 ·0.75H 2 O or LC-Ni(OH) 2 ·2.75H 2 O is similar to that of Ni(OH) 2 below the Fermi level. This indicates that although the overall stability is improved, the activity of the active site is still comparable to that of Ni(OH) 2 . However, LC-Ni(OH) 2 ·2.75H 2 O has a 3D-networking structure and hydrogen bonding H 2 O, which can expose more active sites by releasing the H 2 O and enhance the density of the active sites; thus, the total activity is higher than that of Ni(OH) 2 , which explains why LC-Ni(OH) 2 ·xH 2 O has a low ECSA (Fig. S6b) but high normalized ECSA activity, as observed in the experiment (Fig. 3i).
Further, we investigated the d orbital occupations of the active Ni sites to understand the activity enhancement for LC-Ni(OH) 2 ·xH 2 O (Fig. 6a). Five of the six Ni-O bonds at the active Ni sites in NiMoO 4 ·0.75H 2 O originate from Ni-O-Mo bonds, and the other originates from Ni-H 2 O bonding. Thus, the d orbitals have relatively high symmetry, with one group consisting of double degenerated e g orbitals and the other consisting of three degenerated t 2g orbitals. When LC-Ni(OH) 2 ·xH 2 O is formed after Mo is dissolved out, three Ni-O bonds in the active Ni center originate from OH groups, and the other three originate from H 2 O molecules, which significantly changes the electron state of its d orbitals, thereby breaking the symmetry and lowering the energy of the system. When OH − attacks the Ni active center and releases an electron, i.e., through a hydroxide-ion-coupling electron transfer (HCET) process, one of the high occupied d electrons is stimulated to a higher energy antibonding orbital, increasing the antibonding composition of the Ni-O bond. Therefore, the valence Ni atom shifts from + 2 to + 3, activating the Ni for electrocatalytic reactions.
The mechanism of methanol oxidation is depicted in Fig. 6b, where the four-electron process of OER is replaced by a four-electron process of methanol oxidation. The four-electron methanol oxidation process involves two HCET and two proton-coupled electron transfer (PCET) processes. As aforementioned, when OH − attacks the active center of Ni atoms, Ni 2+ is oxidized to Ni 3+ through the HCET process. Following that, methanol reacts with the hydroxide attached to Ni 3+ to form -OCH 3 . In the presence of an electric field, -OCH 3 releases a proton and an electron through the PCET process to form -OCH 2 . Similarly, in the presence of an electric field, -OCH 2 releases a proton and an electron to form -OCH. Then, OH − continues to attack -OCH through an HCET process, releasing an electron and forming formic acid. The formic acid quickly neutralizes with the alkali in a strongly alkaline environment (6 M KOH) and then generates formate.
Collectively, Mo species in NiMoO 4 ·0.75 H 2 O nanorods are selectively dissolved, forming Mo-doped Ni(OH) 2 . Due to the low coordination effect, amorphous and 3D-networking LC-Ni(OH) 2 ·xH 2 O is formed. For such a special structure, there are more H 2 O and OH − groups around Ni instead of Ni atoms, which increase the contact area between OH − and CH 3 OH and the catalytic active center; thus, the electrochemical properties are increased. Meanwhile, it maintains the Ni(OH) 2 activity for a single site; therefore, overall, the LC-Ni(OH) 2 ·xH 2 O activity is higher than that of Ni(OH) 2 . In addition, part of the Mo that remains and fails to dissolve plays a role in doping therein. It exists in the form of a Mo-O-Ni chemical bond, which is stronger than the hydrogen bond, making the active Ni(OH) 2 unit combine more firmly and improve the stability. Therefore, the coordination effect improves not only the activity but also the stability.

Electrolytic Cell Performance
A two-electrode system with Ni 4 Mo-MoO 2 as the cathode and LC-Ni(OH) 2 ·xH 2 O as the anode is used to produce hydrogen and value-added formate simultaneously. The Ni 4 Mo-MoO 2 is synthesized by heating NiMoO 4 ·0.75H 2 O in an H 2 /Ar atmosphere following the literature [65], and the SEM image shows that Ni 4 Mo alloy nanoparticles grow on the surface of the MoO 2 nanorods (Fig. S13a-b). The XRD confirms the structure of Ni 4 Mo-MoO 2 nanoparticles as indicated in Fig.  S13c. After forming the electrolytic cell, the polarization curves of 0.5 M methanol and no methanol show a significant difference in 1 M KOH (Fig. 7a). In the presence of methanol, MOR dominates, whereas OER dominates in the absence of methanol. The stability of the two-electrode cell system is examined for 100,000 s ( Fig. 7b-c). In the beginning, a voltage of 1.52 V was applied to the cell, resulting in a current density of 130 mA cm −2 . Because of the consumption of methanol and OH − , the current density gradually decreased to 95 mA cm −2 . After refreshing the electrolyte, the current density increased back to 130 mA cm −2 . The coincidence of the LSV curves before and after the stability test also indicates that the two-electrode system is stable (Fig. 7c). To investigate the application of this system under the industrial conditions of high current density and high alkaline electrolyte, we perform electrolysis in 6 M KOH and 3 M methanol at 2.00 V (cell voltage) with over 500 mA cm −2 (without iR correction) for more than 50 h. The current density remains almost constant throughout the tests (Fig. 7d), indicating high activity and stability in the industrial concentration. Furthermore, the Faradaic efficiency remains above 90% (Fig. 7e). When we compare the HER activity of Ni 4 Mo-MoO 2 with and without methanol (Fig. 7f), we discover that the enhanced kinetics of methanol/water co-electrolysis is mostly due to LC-Ni(OH) 2 ·xH 2 O.
Further, we compared the performance of the obtained electrocatalyst with that of recently reported catalysts for small organic molecule co-electrolysis (Fig. 7g). The Ni 4 Mo-MoO 2 || LC-Ni(OH) 2 ·xH 2 O cell obtained herein is the only catalysts could produce a current density of 500 mA cm −2 for small organic molecule co-electrolysis to the best of our knowledge, indicating that the obtained catalyst has potential industrial applications. In the case of the two-electrode system of Ni 4 Mo-MoO 2 || LC-Ni(OH) 2 ·xH 2 O in 1 M KOH and 0.5 M methanol, only 1.49 V is required to obtain a current density of 150 mA cm −2 (Fig. 7a). In the absence of methanol, a higher cell voltage of 1.64 V is required to obtain the same current density. This voltage gap between the cells with and without methanol indicates that energy can be saved in hydrogen production by replacing OER with MOR. For 1 mol H 2 , the energy cost for overall water splitting (HER||OER) is 316 kJ, whereas that of the methanol valueadded boost H 2 production (HER||MOR) is only 287 kJ, indicating that 8.3-11.2% energy can be saved at current densities of 50-300 mA cm −2 , as demonstrated in Fig. 7hi. Furthermore, MOR produces formate (about 550 € per tonne), which is more expensive than methanol (about 350 € per tonne), making this reaction more valuable.
In summar y, electrolytic cells composed of LC-Ni(OH) 2 ·xH 2 O and Ni 4 Mo-MoO 2 can co-electrolyze methanol/water to produce hydrogen at industrial concentrations with low-energy demand owing to the enhanced activity and durability of the LC-Ni(OH) 2 ·xH 2 O electrocatalyst. This proposed hydrogen production method can save 8.3-11.2% energy compared to the traditional direct water electrolysis. In addition, the proposed reaction would be more useful because it can produce high-value chemical formate. Thus, this strategy is economically viable for practical industrial applications.

Conclusions
A novel Mo-doped Ni(OH) 2 containing uncertain H 2 O molecules electrocatalyst LC-Ni(OH) 2 ·xH 2 O was synthesized through the selective dissolution of Mo species from NiMoO 4 ·0.75H 2 O. It shows enhanced kinetics and durability for energy-saving hydrogen production with the co-generation of a highly selective value-added product (formate) for water/methanol co-catalysis. TEM, XPS and EXAFS revealed that dense NiMoO 4 ·0.75H 2 O is in situ  Table S5 of the Supporting Information. h Percentage energy saving for hydrogen production at different current densities when OER is replaced by methanol oxidation reaction. i Energy cost for generating the same amount of H 2 (1 mol) integrated with OER or MOR at constant current densities. The right tick label indicates the required reaction time. The energy costs and percentage energy savings were calculated based on the data in Fig. 7a converted to 3D-networking LC-Ni(OH) 2 ·xH 2 O. Meanwhile, the hydrogen evolution with value-added formate co-generation is boosted at a large current density of more than 500 mA cm −2 and a cell voltage of 2.00 V. The Faradaic efficiency is more than 90% at the current density of more than 500 mA cm −2 with excellent stability for 50 h in a high-concentration electrolyte (6 M KOH). Although XRD could not reveal the structure of LC-Ni(OH) 2 ·xH 2 O because the composition of crystalline H 2 O is uncertain, the fine structure is resolved using XAS and DFT based on the environmental variables, such as the Ni bond length and coordination number. Further mechanistic studies based on DFT revealed that the improved kinetics and durability are mainly attributed to the ultralow Ni-Ni coordination effect for active Ni sites, which results in 3D-networking structures. The ultralow Ni-Ni coordination, 3D-networking structures and Mo dopants improve the intrinsic catalytic activity, increase the active site density and strengthen the binding of 3D-networking structure, respectively. By replacing OER with MOR at the anode, the voltage of the cell consisting of Ni 4 Mo-MoO 2 as the cathode and LC-Ni(OH) 2 ·xH 2 O as the anode decreased from 1.64 to 1.49 V, significantly lowering the energy consumption for hydrogen production. This study paves a new way for realizing low-energy electrolysis of water in industrial alkaline conditions for hydrogen production. up Fund of Shenzhen University (000263 and 000265). We sincerely acknowledge the Instrumental Analysis Center of Shenzhen University (Xili Campus) for HRTEM measurements and thank Shiyanjia Lab (www. shiya njia. com) for the ICP-MS, XPS and XAS experiments.
Funding Open access funding provided by Shanghai Jiao Tong University.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.