In situ generated dual oxyanions enable bifunctional synergy to overcome dehydrogenation limitations in HMF electrooxidation

Abstract

Enhancing the dehydrogenation kinetics of both electrocatalysts and the substrate is crucial for the electrooxidation of biomass-derived 5-hydroxymethylfurfural (HMF) to high-value 2,5-furandicarboxylic acid. Herein, we develop a dual oxyanion co-adsorption strategy by constructing a Mo-doped Ni₃S₂-modified Ni(OH)_x electrocatalyst (denoted as Mo-Ni₃S₂/Ni(OH)_x), enabling the simultaneous adsorption of in situ generated SO₄²⁻ and MoO₄²⁻ oxyanions on the Ni(OH)_x surface. Experimental results and theoretical calculations demonstrate the distinct and synergistic role of co-adsorbed oxyanions in promoting HMF oxidation: (i) SO₄²⁻ primarily lowers the energy barrier for Ni(OH)₂ dehydrogenation, accelerating the formation of active Ni³⁺ species; and (ii) MoO₄²⁻ predominantly interacts with hydrogen atoms of HMF, reducing the adsorption energy of HMF and facilitating its dehydrogenation kinetics. Taking advantage of this bifunctional synergy, Mo-Ni₃S₂/Ni(OH)_x achieves a current density of 100 mA cm⁻² at 1.46 V vs. RHE, 2.5 times higher than that of the Ni₃S₂/Ni(OH)_x reference with only SO₄²⁻ adsorption. This synergy also allows Mo-Ni₃S₂/Ni(OH)_x to outperform Ni₃S₂/Ni(OH)_x in terms of electrochemical activity during the oxidation of other nucleophiles. Remarkably, it demonstrates satisfactory practical applicability in an integrated membrane electrode assembly (MEA) electrolyzer, achieving near 100% FDCA selectivity after 20 cycles at 1.9 V. In contrast, exogenously added oxyanions show a markedly weaker promotional effect due to competitive adsorption with HMF and OH⁻ at the active sites. This work elucidates the regulatory mechanism of similar oxyanions in a complex co-adsorption system for HMF oxidation and offers a rational strategy for designing efficient electrocatalysts for biomass valorization.

---

1 Introduction

Polyethylene terephthalate (PET), one of the most widely produced plastics, contributes substantially to fossil fuel consumption and environmental pollution due to its pervasive use.¹ Biodegradable bio-based polyethylene 2,5-furandicarboxylate (PEF), synthesized from 2,5-furandicarboxylic acid (FDCA) and monoethylene glycol (MEG), represents a promising PET alternative,^2,3 owing to its superior gas barrier properties (O₂, CO₂, and H₂O vapor), enhanced mechanical performance, and improved recyclability.^4,5 However, the large-scale production of PEF remains hindered by the high cost of FDCA manufacturing.⁶ Conventional FDCA synthesis from 5-hydroxymethylfurfural (HMF) requires harsh conditions (100–200 °C, >50 atm) and toxic oxidants (e.g., KMnO₄ and H₂O₂), posing significant safety risks; additionally, the by-product separation entails high energy consumption and cost.^7,8 In contrast, the electrochemical oxidation of HMF (HMFOR) enables efficient FDCA production under ambient conditions, particularly when driven by renewable electricity.^9,10 Nevertheless, the practical implementation of the HMFOR is currently limited by the scarcity of efficient electrocatalysts operating at industrially relevant current densities.¹¹

Nickel-based materials have been extensively studied as electrocatalysts for alkaline HMFOR due to their excellent activity, stability and cost-effectiveness.^12–16 Among these, Ni(OH)₂ shows particular promise due to its intrinsic redox capability and high faradaic efficiency (FE) for FDCA production via an indirect oxidation mechanism.^13,17 Under applied anodic potential, Ni(OH)₂ undergoes a dehydrogenation process to form active Ni(OH)O intermediates; these active intermediates then directly oxidize HMF through the proton-coupled electron transfer (PCET) process.^12,18 However, the sluggish dehydrogenation kinetics of both Ni(OH)₂ and the substrate (i.e., HMF) limit the electrocatalytic activity.^12,19 Furthermore, the weak adsorption of the substrate onto Ni(OH)₂ hinders the PCET process, thereby suppressing the conversion frequency of HMF.^17,19,20 Therefore, promoting the dehydrogenation kinetics while strengthening HMF adsorption is crucial for the HMFOR.

Recent studies have primarily concentrated on accelerating Ni(OH)₂ dehydrogenation by employing methods such as heterostructure construction,^21,22 defect engineering,^20,23 and precious metal incorporation (e.g., Pd and Pt).^12,24 However, conventional nickel-based materials (e.g. Ni(OH)₂ and NiO) are susceptible to irreversible oxidation to high-valent Ni⁴⁺ species, which leads to rapid deactivation due to the loss of reversible redox capability.^25,26 Blindly accelerating the dehydrogenation of Ni(OH)₂ would inevitably result in a trade-off between high activity and poor stability. Therefore, it is crucial to balance the dehydrogenation kinetics of Ni(OH)₂ and the substrate in order to resolve the activity–stability dilemma. Remarkably, embedding phosphate (PO₄³⁻) into Ni(OH)₂ leads to elongation and subsequent cleavage of the O–H bond in HMF, thus promoting the oxidation of HMF.²⁷ Wang et al.²⁸ recently further found that PO₄³⁻ can promote proton deintercalation from lattice hydroxyl groups through the formation of hydrogen bridges to generate Ni³⁺–O active sites. Meanwhile, PO₄³⁻ mediates OH adsorption via hydrogen bridges, thus promoting proton deintercalation from OH and accelerating the oxidation of HMF. Nevertheless, research on oxyanion-mediated modification strategies remains scarce, and the mechanisms of structurally analogous oxyanions in biomass conversion are unclear. Hence, employing various oxyanions to modify nickel-based materials to modulate dehydrogenation kinetics and accelerate biomass conversion remains a highly challenging objective.

Herein, we propose a dual-oxyanion chemisorption strategy achieved through the construction of a Mo-doped Ni₃S₂-modified Ni(OH)_x electrocatalyst (Mo-Ni₃S₂/Ni(OH)_x), which facilitates the co-adsorption of SO₄²⁻ and MoO₄²⁻ during electrochemical reconstruction. In situ Raman spectroscopy confirms that Mo and S species within the electrocatalyst evolve into MoO₄²⁻ and SO₄²⁻ anions under operating conditions, which subsequently re-adsorb onto the Ni(OH)_x surface. Electrochemical analysis, combined with density functional theory (DFT) calculations, reveals that under dual-oxyanion co-adsorption, SO₄²⁻ plays a dominant role in promoting Ni(OH)₂ dehydrogenation, thereby accelerating the formation of active Ni³⁺ species. Concurrently, in situ Raman and DFT results indicate that MoO₄²⁻ is more effective in reducing the adsorption energy of HMF and facilitating its dehydrogenation. As a result, the Mo-Ni₃S₂/Ni(OH)_x electrocatalyst achieves complete HMF conversion with a FE of 94% toward FDCA over eight consecutive reaction cycles. This work advances the understanding of the distinct roles and synergistic regulatory mechanisms of dual oxyanions in the HMFOR, offering fundamental insights for the rational design of high-performance electrocatalysts.

2 Results and discussion

2.1. Synthesis and characterization

The synthesis of Mo-Ni₃S₂/Ni(OH)_x supported on nickel foam (NF) is depicted in Fig. 1a. Mo–Ni(OH)_x was initially synthesized hydrothermally to serve as a precursor. The precursor was then immersed in a sodium sulfide solution for a second hydrothermal step, wherein partial replacement of hydroxyl groups by sulfur atoms yielded Mo-Ni₃S₂/Ni(OH)_x. For comparison, Ni₃S₂/Ni(OH)_x supported on NF was synthesized in an identical manner, but without the addition of ammonium molybdate. The optimized Mo-Ni₃S₂/Ni(OH)_x electrocatalyst was obtained by varying the concentrations of ammonium molybdate and sodium sulfide (Table S2, Fig. S1 and Note S1). The scanning electron microscope (SEM) images (Fig. 1b–d) reveal that both Mo-Ni₃S₂/Ni(OH)_x and Ni₃S₂/Ni(OH)_x form uniform, interconnected thin nanosheets on NF. The Ni₃S₂/Ni(OH)_x nanosheets exhibit uniform thickness and sharp corners (Fig. 1b). In contrast, Mo incorporation induces significant morphological changes, resulting in wavy nanosheets that interweave into a high-density 3D structure (Fig. 1c and d). This curved and thin architecture is expected to enhance specific surface area and active site density,²⁹ attributed to abundant defects generated in the Ni₃S₂ or Ni(OH)₂ crystal structures upon Mo introduction.^29,30 Transmission electron microscopy (TEM) imaging confirms the well-defined morphology of Mo-Ni₃S₂/Ni(OH)_x nanosheets (Fig. 1e). High-resolution TEM (HRTEM, Fig. 1f) reveals lattice spacings of 0.204 nm and 0.231 nm, corresponding to the Ni₃S₂ (202) and jamborite (015) planes, respectively. The selected area electron diffraction (SAED) pattern (Fig. 1g) further confirms Ni₃S₂ (PDF #44-1418, (202) ring) and jamborite (PDF #89-7111, (012) and (110) rings). EDS elemental mapping demonstrates homogeneous distribution of Mo, Ni, S, and O throughout the nanosheets, confirming successful Mo incorporation. The inductively coupled plasma-mass spectrometry (ICP-MS) analysis of Mo-Ni₃S₂/Ni(OH)_x indicates a Mo/Ni atomic ratio of 0.12, as shown in Table S3.

Fig. 1 (a) Schematic illustration of the synthesis of Mo-Ni₃S₂/Ni(OH)_x; (b) SEM image of Ni₃S₂/Ni(OH)_x; (c and d) SEM images, (e) TEM image, (f) HRTEM image, (g) SAED image, and (h) EDS elemental mapping images of Mo-Ni₃S₂/Ni(OH)_x.

X-ray diffraction (XRD) was used to analyze the crystal structures of the electrocatalysts. For Ni(OH)_x (Fig. S2a), diffraction peaks at 19.2°, 38.6°, 52.1°, and 62.9° are indexed to hexagonal Ni(OH)₂ (PDF #73-1520),³¹ while peaks at 11.4°, 23.0°, 33.6°, 34.5°, 46.3°, 59.7°, 61.5°, and 65.1° correspond to hexagonal jamborite (Ni(OH)₂/NiOOH hybrid, PDF #89-7111).³² In Mo-Ni(OH)₂, the jamborite peaks persist, whereas Ni(OH)₂ peaks are absent, indicating that the presence of Mo inhibits the formation of Ni(OH)₂. Following the sulfidation of Mo-Ni(OH)₂, new peaks emerge at 21.7°, 31.1°, 44.3°, 49.7°, and 55.1° in Mo-Ni₃S₂/Ni(OH)_x, which are assigned to Ni₃S₂ (PDF #44-1418), confirming successful sulfur incorporation. Notably, comparing Mo-Ni₃S₂/Ni(OH)_x to Ni₃S₂/Ni(OH)_x reveals that the Ni₃S₂ peaks are reduced in intensity and broadened under Mo incorporation, but there is no peak shift. This suggests that Mo incorporation introduces lattice defects and refines the Ni₃S₂ grains.^29,30 X-ray photoelectron spectroscopy (XPS) was employed to probe the elemental composition and electronic structure. The Mo-Ni₃S₂/Ni(OH)_x survey spectrum confirms the coexistence of Mo, Ni, S, and O (Fig. S2b). Detection of Mo 3d spectra in Mo-Ni₃S₂/Ni(OH)_x (absent in Ni₃S₂/Ni(OH)_x) confirms Mo incorporation. Deconvolution shows peaks at 235.1 eV (Mo 3d_3/2) and 232.0 eV (Mo 3d_5/2), assigned to Mo⁶⁺ (235.4, 232.2 eV) and Mo⁵⁺ (234.8, 231.7 eV) (Fig. S2c).³³ Ni 2p spectra of both materials are dominated by Ni²⁺. Significantly, Ni 2p binding energies in Mo-Ni₃S₂/Ni(OH)_x exhibit a 0.2 eV negative shift versus Ni₃S₂/Ni(OH)_x (Fig. S2d), indicating increased electron density at Ni²⁺ ions. S 2p spectra of Mo-Ni₃S₂/Ni(OH)_x (Fig. S2e) show peaks at 161.7 eV (S 2p_3/2) and 163.7 eV (S 2p_1/2) for S²⁻,^7,34,35 and a peak at 169.2 eV for surface-oxidized SO₄²⁻.^7,36 In summary, Mo incorporation successfully modified Ni₃S₂/Ni(OH)_x, promoting the generation of lattice defects while modulating the electronic environments of Ni in Mo-Ni₃S₂/Ni(OH)_x.

2.2. HMFOR performance assessment

The HMFOR activities of different electrocatalysts were assessed via linear sweep voltammetry (LSV) in 1.0 M KOH with 50 mM HMF (Fig. 2a). Remarkably, Mo-Ni₃S₂/Ni(OH)_x exhibits superior HMFOR activity compared to Ni₃S₂/Ni(OH)_x and NF. At 1.46 V vs. RHE, Mo-Ni₃S₂/Ni(OH)_x achieves 100 mA cm⁻², a 2.5-fold enhancement over Ni₃S₂/Ni(OH)_x. This pronounced activity difference highlights the crucial role of Mo incorporation in boosting HMFOR performance. Furthermore, benchmark testing confirms that Mo-Ni₃S₂/Ni(OH)_x outperforms most reported electrocatalysts at equivalent potentials (Fig. 2b and Table S4). Mo-Ni₃S₂/Ni(OH)_x still exhibits performance comparable to that of some of the most advanced electrocatalysts even at a high HMF concentration of 100 mM (Fig. S4 and Table S5). Tafel analysis further elucidates the kinetics. Mo-Ni₃S₂/Ni(OH)_x exhibits a significantly lower Tafel slope (45.37 mV dec⁻¹, Fig. S5a) than Ni₃S₂/Ni(OH)_x (88.76 mV dec⁻¹) and NF (138.41 mV dec⁻¹), demonstrating accelerated electron transfer kinetics. This finding is corroborated by electrochemical impedance spectroscopy (EIS, Fig. 2c), where Mo-Ni₃S₂/Ni(OH)_x displays a smaller charge transfer resistance (R_ct). Additionally, the double-layer capacitance (C_dl) of Mo-Ni₃S₂/Ni(OH)_x (2.29 mF cm⁻², Fig. S5b–d) is larger than that of Ni₃S₂/Ni(OH)_x (1.25 mF cm⁻²), indicating a greater electrochemically active surface area (ECSA) and exposure of active sites. It should be emphasized that the FDCA selectivity of Mo-Ni₃S₂/Ni(OH)_x is far higher than that of Ni₃S₂/Ni(OH)_x, even though they have close FE of FDCA (Fig. S6a–c). Multiple side reactions in the HMFOR are key factors affecting the FDCA selectivity.^37,38 Compared with Ni₃S₂/Ni(OH)_x, the faster oxidation rate of HMF over Mo-Ni₃S₂/Ni(OH)_x prevents undesirable side reactions, such as self-degradation, self-polymerization,³⁹ and the Cannizzaro⁴⁰ reaction of HMF. Furthermore, chronoamperometry experiments confirm that Mo-Ni₃S₂/Ni(OH)_x cannot oxidize the carboxyl groups in FDCA (Fig. S6d). This avoids overoxidation of the product, leading to high FDCA selectivity.

Fig. 2 (a) LSV curves of different electrocatalysts in 1.0 M KOH with 50 mM HMF (without iR compensation); (b) comparison of the performance of Mo-Ni₃S₂/Ni(OH)_x with those of previously reported electrocatalysts; (c) Nyquist plots of the electrocatalysts in 1.0 M KOH with the addition of 50 mM HMF at 1.45 V vs. RHE; (d) current density of Mo-Ni₃S₂/Ni(OH)_x during chronoamperometry measurements; (e) HMF conversion, FDCA selectivity and FE under 8 successive electrolysis cycles; (f) the visualized MEA setup for the HMFOR and HER (AEM: anion exchange membrane); (g) the LSV curves of Mo-Ni₃S₂/Ni(OH)_x in the MEA; (h) stability measurements for the MEA electrolyzer at 1.9 V over Mo-Ni₃S₂/Ni(OH)_x in 1.0 M KOH with 50 mM HMF; (i) comparison of activities of the electrocatalysts for different nucleophiles.

The stability of Mo-Ni₃S₂/Ni(OH)_x for the HMFOR was evaluated through cyclic tests at 1.45 V vs. RHE in an H-cell, since the competitive oxygen evolution reaction (OER) at higher potentials leads to a reduction in the selectivity and FE of FDCA (Fig. S6e). Notably, high FDCA selectivity and FE (98%, 94%) and 100% HMF conversion were maintained after 8 cycles (Fig. 2e), outperforming most reported electrocatalysts (Table S6) and demonstrating excellent cycling stability. It is observed that during the initial five cycles (with electrocatalyst rinsing between tests), a distinctive oxidation peak emerges, characterized by a sudden increase in current density within 0.2 hours, followed by a gradual decline (Fig. 2d). We attribute this to electrocatalyst reconstruction via dissolution of S and Mo.^41,42 To validate this, a 2.5 h activation step at 1.45 V vs. RHE in 1.0 M KOH was performed prior to the sixth cycle. Activation shows gradual current decay from ≈20 mA to ≈0.3 mA with 8.6C total charge transfer (Fig. S7a). Critically, in situ EIS and LSV analysis (Fig. S7b and c) confirm that the OER onset occurs at 1.48 V vs. RHE (1.0 mA cm⁻²), indicating that charge transfer at 1.45 V vs. RHE reflects reconstruction dynamics, not oxygen evolution. Significantly, the first five cycles exhibit higher charge consumption (180–190C vs. theoretical 174C, Fig. S8), yielding lower FDCA FE (≈88%). After activation, cycles 6–8 show reduced charge (178–185C) and improved FE (≈94%), indicating that continuous reconstruction does not compromise FDCA efficiency. Next, the near-practical performance was assessed in a membrane electrode assembly (MEA) electrolyzer with Mo-Ni₃S₂/Ni(OH)_x as the anode (3.0 × 3.0 cm²) and NF as the cathode (Fig. 2f). It is worth noting that a current of 1.0 A was attained at 1.96 V with 50 mM HMF, whereas a voltage of 280 mV was required in the absence of HMF (Fig. 2h). Moreover, the onset potential of the HMFOR is lower than that of the OER, indicating that the HMFOR is thermodynamically more favorable than the OER.⁴³ These results indicate that the HMFOR can effectively supplant the energy-intensive OER, thereby facilitating valuable chemical production with reduced energy input. Furthermore, chronoamperometric tests were conducted at a constant potential of 1.9 V in a 50 mM HMF solution to evaluate the stability of the electrocatalyst over 20 cycles (Fig. S9). As illustrated in Fig. 2h, Mo-Ni₃S₂/Ni(OH)_x retains 100% HMF conversion and nearly 100% FDCA selectivity after 20 cycles in the MEA electrolyzer. Due to the more favorable thermodynamics of the HMFOR and its significantly higher current density compared to the OER at 1.9 V (Fig. S10), competing OER does not affect the high selectivity towards FDCA. This result underscores its outstanding electrocatalytic stability and potential for large-scale FDCA production. Moreover, to evaluate the versatility of the electrocatalyst, we assessed the electrocatalytic performance in the electrooxidation of other nucleophiles such as urea, glycerol, propylene glycol (PG), glycol, furfural, furfuryl alcohol (FA), and ethanol using LSV (Fig. 2i and S11). For all substrates, Mo-Ni₃S₂/Ni(OH)_x exhibited higher electrocatalytic activity, indicating that the dissolution of Mo and S significantly affects the electrooxidation of various nucleophiles.

2.3. Dynamic dissolution and re-adsorption of MoO₄²⁻ and SO₄²⁻

The phase reconstruction of Mo-Ni₃S₂/Ni(OH)_x during the HMFOR was investigated using multiple characterization techniques. The post-reaction SEM (Fig. S12) confirms that the wrinkled nanosheet morphology remains intact after 22 h of HMFOR operation. In addition, the EDS mapping reveals persistent but diminished surface concentrations of Mo and S (Fig. S12c and f), indicating partial dissolution during HMFOR operation. XRD patterns (Fig. S13) confirm that the electrocatalyst retains the Ni(OH)_x structure after 22 h HMFOR. Critically, in situ Raman spectroscopy tracks structural evolution. For Mo-Ni₃S₂/Ni(OH)_x (Fig. 3b) under open-circuit potential (OCP) conditions, weak peaks at 200–400 cm⁻¹ correspond to Ni–S bonds,²¹ while the peak at 495 cm⁻¹ is assignable to disordered Ni(OH)₂ (Ni²⁺–O vibration).⁴⁴ Upon applying 1.45 V vs. RHE, partial disappearance of Ni–S and the presence of Ni²⁺–O signatures indicate the conversion of Ni₃S₂ to Ni(OH)₂.⁴⁵ Furthermore, between OCP and 1.65 V vs. RHE, the peak at ≈1025 cm⁻¹ matches SO₄²⁻,⁴⁵ while the feature at ≈894 cm⁻¹ (observed at 1.25–1.45 V vs. RHE) corresponds to Mo=O symmetric stretching in MoO₄²⁻.⁴⁶ These observations indicate that S and Mo leach from the electrocatalyst lattice during the HMFOR, subsequently adsorbing as SO₄²⁻ and MoO₄²⁻ surface species (Fig. 3a).

Fig. 3 (a) Schematic diagram of Mo-Ni₃S₂/Ni(OH)_x reconstruction; (b–d) in situ Raman spectra of electrocatalysts in 1.0 M KOH with or without 50 mM HMF; (e) Mo 3d, (f) S 2p, and (g) Ni 2p XPS spectra of Mo-Ni₃S₂/Ni(OH)_x before and after the HMFOR.

Similar structural evolution occurs during the OER on Mo-Ni₃S₂/Ni(OH)_x. Specifically, Raman peaks at 475 and 557 cm⁻¹ emerge at 1.40 V vs. RHE, corresponding to Ni³⁺–O bending and stretching vibrations,⁴⁷ confirming transformation of Ni₃S₂/Ni(OH)_x into NiOOH with concurrent surface adsorption of MoO₄²⁻ and SO₄²⁻. Notably, the NiOOH formation during the HMFOR requires 1.50 V vs. RHE. This difference implies that Ni³⁺ species generated at 1.40–1.45 V vs. RHE undergo rapid reaction with HMF via a PCET mechanism (Ni³⁺–O + HMF → Ni²⁺–OH + HMFCA),^48,49 reverting to Ni(OH)₂ before accumulating. The OCP measurements corroborate this mechanism. As shown in Fig. S15, after 5 min pre-oxidation at 1.45 V vs. RHE, accumulated Ni³⁺ species reduce to Ni²⁺ within ≈40 min in 50 mM HMF-containing electrolyte, whereas reduction requires ≈20 h in 1.0 M KOH. This accelerated reduction kinetics confirms spontaneous reduction of Ni³⁺–O by HMF. These results indicate that the Ni³⁺–O species serve as the active centers for the HMFOR.

Therefore, during the HMFOR, NiOOH within Ni(OH)_x (i.e., the NiOOH/Ni(OH)₂ hybrid) spontaneously reacts with HMF, reducing to Ni(OH)₂. Concurrently, Ni(OH)₂ in Mo-Ni₃S₂/Ni(OH)_x undergoes potential-driven dehydrogenation to form Ni³⁺ intermediates; these subsequently regenerate Ni(OH)₂ via a PCET process. This dynamic cycle enables rapid HMF oxidation. Notably, while both Ni₃S₂/Ni(OH)_x and Mo-Ni₃S₂/Ni(OH)_x exhibit a NiOOH formation potential of 1.40 V vs. RHE during the OER (Fig. S16 and 3c), Mo-Ni₃S₂/Ni(OH)_x requires a higher potential (1.50 V vs. RHE) for NiOOH formation during the HMFOR compared to Ni₃S₂/Ni(OH)_x (1.40 V vs. RHE). This shift indicates that re-adsorbed MoO₄²⁻ anions facilitate the PCET process of HMF dehydrogenation, thereby enhancing HMFOR activity and avoiding irreversible oxidation of Ni(OH)₂.

XPS analysis further supports this structural transformation mechanism. As shown in Fig. 3e, the significant decreases in Mo⁵⁺ and Mo⁶⁺ peak intensities indicate extensive Mo dissolution, with residual Mo likely adsorbed as MoO₄²⁻. Additionally, the complete disappearance of the Ni–S bond signal in the S 2p spectra (Fig. 3f) demonstrates electrochemical dissolution of surface S from Mo-Ni₃S₂/Ni(OH)_x. Notably, Ni 2p spectra after 8 electrolysis cycles of HMF remain dominated by Ni²⁺, consistent with the initial state, though a minor Ni³⁺ presence persists. This residual Ni³⁺ likely arises from (1) hindered reduction back to Ni²⁺ due to low HMF concentration in later electrolysis stages, and (2) inevitable surface oxidation by air.⁵⁰

2.4. Bifunctional synergy of co-adsorbed MoO₄²⁻/SO₄²⁻ in dehydrogenation steps

As confirmed above, during the HMFOR process, the S and Mo in Mo-Ni₃S₂/Ni(OH)_x progressively oxidize to thermodynamically stable SO₄²⁻ and MoO₄²⁻. These oxyanions initially dissolve into the electrolyte, while concurrently undergoing partial, rapid and stable re-adsorption onto the Ni(OH)_x surface. This clear reconstruction behavior likely contributes significantly to the exceptional HMFOR performance of Mo-Ni₃S₂/Ni(OH)_x. To test this hypothesis, we examined the redox behavior of the electrocatalysts via cyclic voltammetry (CV). All samples exhibit reversible Ni²⁺/Ni³⁺ redox peaks, with the closed CV curve area progressively increasing during activation (Fig. S17). Notably, after 25 cycles, the CV curve area followed the order Mo-Ni₃S₂/Ni(OH)_x > Ni₃S₂/Ni(OH)_x > Ni(OH)_x. This suggests that the oxidation and dissolution of both Mo and S promotes the reconstruction, thereby exposing additional convertible Ni sites for the HMFOR.

To elucidate the distinct roles of two structurally similar oxyanions in the HMFOR, CV was employed to investigate how adsorbed MoO₄²⁻ and SO₄²⁻ influence the formation of Ni³⁺ species, which are catalytically critical for the HMFOR.^35,49 As depicted in Fig. 4a, the onset potential for Ni²⁺ oxidation shifts negatively (i.e., decreases) with increasing SO₄²⁻ concentration in the electrolyte, while the redox peak area significantly increases. This confirms that SO₄²⁻ adsorption facilitates Ni(OH)₂ oxidation to active Ni³⁺ species, thereby enhancing HMFOR activity (Fig. S18a). In the classical OER mechanism, EIS distinguishes two electron-transfer steps (electrocatalyst oxidation and the OER itself). For the Bode plots, the high-frequency region corresponds to electrode surface oxidation, whereas the low-frequency region relates to the OER interface.⁵¹ Critically, introducing 0.1 M SO₄²⁻ into 1.0 M KOH reduces the high-frequency phase angle (Fig. 4d), indicating that SO₄²⁻ adsorption accelerates Ni(OH)₂ oxidation. This conclusion is supported by the decreased diameter of the Nyquist semicircle upon SO₄²⁻ addition (Fig. S19a–c).

Fig. 4 CV curves of Ni(OH)_x in pure 1.0 M KOH electrolyte (Fe-free) containing different concentrations of (a) K₂SO₄, (b) Na₂MoO₄, and (c) K₂SO₄ + Na₂MoO₄; Bode plots of Ni(OH)_x in pure 1.0 M KOH with or without (d) 0.1 M K₂SO₄ and (e) 0.1 M Na₂MoO₄; (f) Bode plots of Ni(OH)_x in pure 1.0 M KOH containing 100 mM HMF, with or without 0.01 M Na₂MoO₄ + 0.01 M K₂SO₄; (g) free energies for the generation of Ni(OH)O from Ni(OH)₂, Ni(OH)₂–SO₄²⁻, Ni(OH)₂–MoO₄²⁻ and Ni(OH)₂–SO₄²⁻–MoO₄²⁻, and the corresponding optimized structures; (h) adsorption free energy of HMF on Ni(OH)O, Ni(OH)O–SO₄²⁻, Ni(OH)O–MoO₄²⁻ and Ni(OH)O–SO₄²⁻–MoO₄²⁻; (i) in situ Raman spectra of Mo-Ni₃S₂/Ni(OH)_x in 1.0 M KOH (disconnection indicates circuit disconnection); (j) schematic diagram of bifunctional synergy between co-adsorbed SO₄²⁻/MoO₄²⁻ for facilitating two dehydrogenation steps.

Notably, the influence of adsorbed MoO₄²⁻ on Ni(OH)₂ oxidation stands in stark contrast to that of SO₄²⁻, despite both species enhancing HMFOR performance (Fig. S18b). As shown in Fig. 4b, increasing MoO₄²⁻ concentration induces a positive shift in the onset potential for Ni(OH)₂ dehydrogenation, suggesting that MoO₄²⁻ adsorption renders Ni(OH) oxidation either thermodynamically less favorable or kinetically hindered.⁴³ Nevertheless, higher MoO₄²⁻ concentrations substantially increase the integrated redox peak area in the CV profiles, indicating a greater extent of Ni²⁺ to Ni³⁺ conversion. To investigate this seemingly contradictory phenomenon, we analyze the underlying mechanisms from both kinetic and thermodynamic perspectives. As shown in Fig. 4e and S19d–f, the Nyquist and Bode plots reveal that MoO₄²⁻ reduces the diameter of the Nyquist semicircle and decreases the phase angle of the peaks in the high-frequency region, indicating that its adsorption kinetically accelerates charge transfer and enhances the Ni(OH)₂ oxidation rate. Furthermore, the lower Tafel slope for Ni(OH)_x + MoO₄²⁻ versus Ni(OH)_x corroborates this kinetic enhancement (Fig. S21). These results suggest that the MoO₄²⁻ adsorption primarily imposes thermodynamic inhibition on Ni(OH)₂ dehydrogenation, which was later confirmed by DFT calculations. However, Ni(OH)₂ dehydrogenation is an electrochemical process,⁴⁸ so increasing the applied potential overcomes this slight inhibitory effect. Specifically, at potentials exceeding ≈1.40 V vs. RHE, the suppressed Ni²⁺ sites can convert to Ni³⁺.

To elucidate the effect of co-adsorbed MoO₄²⁻/SO₄²⁻ on Ni(OH)₂ dehydrogenation, a mixture of the two oxyanions was introduced into pure 1.0 M KOH electrolyte. As shown in Fig. 4c, the onset potential for Ni(OH)₂ dehydrogenation exhibits a non-monotonic trend with increasing total oxyanion concentration, i.e., initially shifting positively, and then negatively, before rising again. This complex behavior suggests competitive adsorption at Ni sites between the two anions. Notably, the mixed additives enhance the redox peak areas, although the increase is not strictly proportional to concentration, indicating that while co-adsorbed oxyanions promote Ni³⁺ generation, MoO₄²⁻ retains an inhibitory effect on Ni(OH)₂ dehydrogenation within the dual-oxyanion system. Only the low-concentration mixture (0.01 M of each component) leads to improved HMFOR performance (Fig. S18c). In contrast, higher concentrations suppress HMFOR activity, likely due to excessive surface site occupation by exogenously added oxyanions, which hinders HMF adsorption. EIS further supports this interpretation. In 1.0 M KOH containing 100 mM HMF and low-concentration additives, a distinct Nyquist semicircle is observed at 1.35 V vs. RHE, corresponding to the onset of HMF oxidation.⁷ Specifically, in the potential range of 1.35–1.45 V vs. RHE, the Ni(OH)_x electrode with co-adsorbed MoO₄²⁻ and SO₄²⁻ exhibits smaller semicircles than the unmodified Ni(OH)_x electrode (Fig. S19g–i), indicating reduced charge-transfer resistance. The corresponding Bode plots support this finding: the Ni(OH)_x + MoO₄²⁻ + SO₄²⁻ electrode displays a phase-angle inflection at 1.35 V vs. RHE and consistently lower phase angles than Ni(OH)_x across the measured frequency range (Fig. 4f). These results demonstrate that MoO₄²⁻/SO₄²⁻ co-adsorption accelerates HMFOR kinetics by facilitating charge transfer at the electrode/electrolyte interface.

In situ Raman spectroscopy was conducted to elucidate the adsorption mode of autogenous MoO₄²⁻/SO₄²⁻. First, Mo-Ni₃S₂/Ni(OH)_x was oxidized in 1.0 M KOH at 1.6 V vs. RHE for 10 min to produce oxyanions. Subsequently, the circuit was disconnected and the electrolyte was immediately replaced with fresh 1.0 M KOH to rinse the electrode in the flow cell (Fig. S22). After 10 min, the electrolyte was replaced a second time and rinsing continued for a further 10 min. Fig. 4i shows that, after oxidation at 1.6 V vs. RHE for 10 min, the characteristic peaks of MoO₄²⁻ and SO₄²⁻ exhibit high intensity. This indicates that a large number of oxyanions are adsorbed onto the electrode surface. Following circuit disconnection, the intensity of the two characteristic peaks decreased significantly, indicating substantial desorption. This confirms that the applied electric field facilitates oxyanion adsorption and that weakly electrostatically adsorbed species desorb upon field removal. Notably, even after rinsing the electrocatalyst with oxyanion-free electrolyte for 20 min under disconnected conditions, no further decrease in intensity was observed for the MoO₄²⁻/SO₄²⁻ peaks. This suggests that some of the oxyanions are firmly anchored to the electrode surface through specific adsorption rather than through the applied electric field or dynamic equilibrium with dissolved oxyanions. Due to the instability of electrostatic adsorption,⁵² these stable chemisorbed oxyanions are key to significantly enhancing the HMFOR kinetics. To estimate the concentration of autogenous oxyanions, electrolysis was conducted using a Mo-Ni₃S₂/Ni(OH)_x electrode at 1.45 V vs. RHE for 3 h (10 mM HMF, 30 mL). ICP-MS analysis of the electrolyte yielded Mo and S concentrations of 0.075 and 0.11 mM, respectively. Since a portion of the oxyanions are stably adsorbed onto the electrocatalyst surface, their local concentration on the electrode surface during operation is expected to be higher than the measured value.

To elucidate the intrinsic mechanism of co-adsorbed MoO₄²⁻ and SO₄²⁻ in the HMFOR, DFT calculations were performed. Given that deintercalation of lattice hydroxyl groups from Ni-based hydroxides during the HMFOR likely exposes underlying Ni atoms,²⁰ and in line with reported chemisorption modes of oxyanions on NiOOH,^{41,45,52–55} adsorption models for MoO₄²⁻ and SO₄²⁻ on Ni(OH)₂ were rationally constructed. The optimized structures show that both oxyanions bridge surface Ni atoms via oxygen ligands (Fig. S23). Calculated free energy changes reveal distinct roles for each oxyanion in Ni(OH)₂ dehydrogenation. Specifically, the dehydrogenation free energy of Ni(OH)₂ with adsorbed SO₄²⁻ is 2.26 eV, lower than that of pristine Ni(OH)₂ (2.50 eV), indicating that SO₄²⁻ reduces the thermodynamic barrier for Ni³⁺ formation (Fig. 4g). In contrast, Ni(OH)₂ with adsorbed MoO₄²⁻ exhibits a higher dehydrogenation free energy (2.60 eV), confirming that MoO₄²⁻ adsorption thermodynamically disfavors Ni(OH)₂ oxidation. Notably, in the co-adsorption configuration (Ni(OH)₂–MoO₄²⁻–SO₄²⁻), the dehydrogenation free energy is 2.39 eV, which is intermediate between the two individual cases but higher than that of SO₄²⁻ alone. This indicates that, when both oxyanions are co-adsorbed, MoO₄²⁻ partially counteracts the promoting effect of SO₄²⁻ on Ni(OH)₂ dehydrogenation. The preferential promotion by SO₄²⁻ in the co-adsorbed system is attributed to the formation of six hydrogen bonds between the oxygen atoms of adsorbed SO₄²⁻ and adjacent hydrogen atoms on the Ni(OH)₂ surface (Fig. 4g). These electronic interactions activate the hydrogen atoms, thereby facilitating dehydrogenation. Based on the relationship between the conjugate acid–base pair (K_aK_b = K_w), the pK_a of HSO₄⁻ in water is 1.9.⁵⁶ From this, the pK_b of SO₄²⁻ can be calculated to be 12.1, which indicates that SO₄²⁻ is an extremely weak base. Therefore, under alkaline conditions, SO₄²⁻ is unlikely to form HSO₄⁻. It can therefore be inferred that the weak hydrogen bond between SO₄²⁻ and Ni(OH)₂ merely activates the H atom in Ni(OH)₂ rather than directly abstracting H to form HSO₄⁻. This mechanism differs fundamentally from that of PO₄³⁻, which promotes the dehydrogenation of Ni(OH)₂ by forming HPO₄²⁻.²⁸

In addition, the Ni(OH)₂–MoO₄²⁻–SO₄²⁻ system exhibits a lower dehydrogenation free energy than pristine Ni(OH)₂, suggesting that co-adsorption thermodynamically facilitates Ni³⁺ formation. However, this computational prediction appears to contradict the experimental observation that an equimolar mixture of MoO₄²⁻ and SO₄²⁻ increases the onset potential for Ni(OH)₂ oxidation. We attribute this discrepancy to the superior electronegativity and basicity (pK_b) of MoO₄²⁻,^52,57 which favors its preferential chemisorption over SO₄²⁻ at Ni sites. As a result, under equimolar additive conditions, MoO₄²⁻ achieves higher surface coverage, and its inherent inhibitory effect on Ni(OH)₂ dehydrogenation dominates the electrochemical response, offsetting the promoting effect of co-adsorbed SO₄²⁻ and leading to an increased onset potential.

The adsorption of HMF on the electrocatalyst surface is a critical step in the PCET process and therefore essential for HMFOR performance.²¹ In principle, the transformation of Ni₃S₂ to Ni(OH)₂ during electrochemical reconstruction could weaken HMF adsorption and compromise electrocatalytic activity, given the intrinsically low affinity of hydrophilic Ni(OH)₂ for organic substrates such as HMF.^35,58 However, in this study, the reconstructed Mo-Ni₃S₂/Ni(OH)_x retained high HMFOR activity, suggesting that the co-adsorbed oxyanions may actively participate in HMF binding. To test this hypothesis, we performed OCP measurements, which are sensitive to changes in adsorbate composition within the Helmholtz layer.⁵⁹ Prior to OCP testing, the post-reaction Mo-Ni₃S₂/Ni(OH)_x electrode was activated by five CV cycles in 1.0 M KOH (0–0.8 V vs. Hg/HgO) to generate surface-adsorbed oxyanions. As shown in Fig. S24, the introduction of HMF induced an OCP shift (Δ_OCP) of 226 mV for the post-reaction electrode, compared to 198 mV for the pristine Mo-Ni₃S₂/Ni(OH)_x electrode. The larger Δ_OCP observed for the post-reaction sample indicates enhanced HMF adsorption, which we preliminarily attribute to the presence of self-generated oxyanions on the surface.

To gain deeper insight into HMF adsorption, we performed DFT calculations. As shown in Fig. 4h, the calculated adsorption energies of HMF on Ni(OH)O, Ni(OH)O–SO₄²⁻, Ni(OH)O–MoO₄²⁻ and Ni(OH)O–SO₄²⁻–MoO₄²⁻ are −0.72, −1.10, −1.36, and −1.48 eV, respectively. These values indicate that dual chemisorbed oxyanions significantly promote HMF adsorption, thereby facilitating subsequent dehydrogenation steps. Notably, the adsorption energy of HMF on Ni(OH)O–MoO₄²⁻ is lower than that on Ni(OH)O–SO₄²⁻. Furthermore, in the optimized configuration of HMF adsorbed on Ni(OH)O–SO₄²⁻–MoO₄²⁻, the H atom of HMF preferentially forms a hydrogen bond with the O atom of MoO₄²⁻ (Fig. 4h). These results suggest that under co-adsorption conditions, the O atoms of MoO₄²⁻ act as new adsorption sites for HMF, activating its H atoms via hydrogen bonding and promoting its dehydrogenation. This mechanism differs from that of PO₄³⁻, which mediates OH adsorption via a hydrogen bridge and promotes HMF oxidation.

To further verify the effect of MoO₄²⁻ on promoting the kinetics of HMF dehydrogenation, in situ Raman spectroscopy was used to record the dynamic variations of Ni³⁺ throughout the PCET process (Fig. S25). First, Ni(OH)_x was treated in 1.0 M KOH at 1.55 V vs. RHE for 5 min to oxidize Ni²⁺–OH to Ni³⁺–O. Then, 10 mM HMF was injected to react with the Ni³⁺ species and the changes to the characteristic peaks of Ni³⁺–O on the electrocatalyst surface were monitored. The results show that, following oxidation treatment, the characteristic peaks of Ni³⁺–O appears in all samples, indicating the accumulation of Ni³⁺OOH. After the addition of HMF, the Ni³⁺–O peaks gradually diminish and eventually disappear. In the absence or presence of SO₄²⁻, a significant amount of Ni³⁺–O remains on Ni(OH)_x after 3 min. In contrast, following the addition of MoO₄²⁻, the Ni³⁺ species that accumulate on Ni(OH)_x almost completely disappeared after 2.5 min. This suggests that MoO₄²⁻ markedly promotes HMF dehydrogenation and Ni³⁺ reduction, whereas SO₄²⁻ has little effect on this process.

In summary, during the HMFOR, Mo and S species in Mo-Ni₃S₂/Ni(OH)_x undergo in situ conversion to MoO₄²⁻ and SO₄²⁻, which subsequently co-adsorb onto the Ni(OH)_x surface. Despite their similar structures and adsorption configurations, these two oxyanions function as independent promoters, each playing a distinct and synergistic role in enhancing HMF oxidation: (i) the O atoms of SO₄²⁻ primarily interact with the H atoms of Ni(OH)₂ via hydrogen bonding, promoting its dehydrogenation and the formation of active Ni³⁺ species; (ii) the O atoms of MoO₄²⁻ preferentially interact with the H atoms of HMF through hydrogen bonding, strengthening HMF adsorption and accelerating its dehydrogenation kinetics. This cooperative decoupling of proton transfer from both the electrocatalyst and the substrate, enabled by co-adsorbed dual oxyanions, is defined as “bifunctional synergy” (Fig. 4j). Importantly, HPLC and ATR-FTIR spectroscopy confirm that the dual-oxyanion adsorption does not alter the intrinsic reaction pathway of the HMFOR (Fig. S26–28 and Note S4), underscoring that the observed enhancement arises from synergistic modulation of reaction kinetics rather than a change in the mechanism. For typical Ni(OH)₂, the formation of Ni³⁺ species is a prerequisite for HMF oxidation. However, Ni(OH)_x contains a certain proportion of NiOOH. Consequently, the two dehydrogenation steps occur in parallel rather than sequentially on Mo-Ni₃S₂/Ni(OH)_x. In the HMFOR reaction pathway, NiOOH in Ni(OH)_x rapidly abstracts H atoms from HMF with the promotion of MoO₄²⁻. Meanwhile, Ni(OH)₂ in Ni(OH)_x is oxidized to Ni³⁺–O species under the effect of SO₄²⁻, which further oxidizes HMFCA. FFCA is subsequently oxidized to FDCA through the same PCET process. As the PCET process of HMF dehydrogenation is the rate-determining step of the HMFOR, it plays a more significant role in reaction efficiency than Ni(OH)₂ oxidation.⁴⁸ Furthermore, facilitating the reduction of Ni³⁺ is beneficial in avoiding the excessive oxidation of Ni³⁺ to Ni⁴⁺. Consequently, the effect of MoO₄²⁻ on HMF dehydrogenation is more significant than the effect of SO₄²⁻ on Ni(OH)₂ dehydrogenation.

2.5. Autogenetic oxyanions vs. externally added oxyanions

While the autogenetic dual MoO₄²⁻ and SO₄²⁻ underpin the superior HMFOR activity of Mo-Ni₃S₂/Ni(OH)_x, externally adding these oxyanions to the Ni(OH)_x electrode (denoted as Ni(OH)_x + MoO₄²⁻ + SO₄²⁻) does not yield the same performance. As shown in Fig. 5a, the LSV curves indicate that Mo-Ni₃S₂/Ni(OH)_x achieves 163 mA cm⁻² at 1.60 V vs. RHE for HMF oxidation, surpassing Ni(OH)_x + MoO₄²⁻ + SO₄²⁻ by 108 mA cm⁻². These results emphasize the important role of self-generated oxyanions in improving the HMFOR kinetics and raise a central question: why do externally added oxyanions fail to replicate the efficacy of their in situ-generated counterparts?

Fig. 5 (a) LSV curves of Mo-Ni₃S₂/Ni(OH)_x and Ni(OH)_x + MoO₄²⁻ + SO₄²⁻ in 1.0 M KOH with 10 mM HMF; (b) in situ ATR-FTIR spectra of the electrocatalysts after varying times at 1.45 V vs. RHE during the HMFOR. (c and d) Open-circuit potentials of electrocatalysts in 1.0 M KOH with and without 10 mM HMF; PDOS of (e) Ni₃S₂/Ni(OH)_x and (f) Mo-Ni₃S₂/Ni(OH)_x; (g) schematic illustration of different roles of additive vs. self-generated oxyanions in the HMFOR.

In general, in situ-dissolved oxyanions adsorb rapidly and stably onto the near-surface of Ni sites,^41,52 whereas externally added oxyanions must first migrate through the bulk electrolyte into the electric double layer, a process that is constrained by mass transfer and the structure of the electric double layer. Consequently, externally added MoO₄²⁻/SO₄²⁻ coexist with OH⁻ and HMF in the bulk electrolyte, inducing competitive adsorption at limited active sites during the HMFOR. Although oxyanion adsorption is reversible, the persistent presence of external MoO₄²⁻/SO₄²⁻ suggests that their adsorption energies may be comparable to or lower than those of HMFOR intermediates.⁶⁰ Furthermore, anodic potentials accelerate oxyanion migration towards electrode interfaces,⁴¹ which could inhibit the adsorption of HMF and OH⁻.

To validate this hypothesis, we probed OH⁻ adsorption on post-reaction Mo-Ni₃S₂/Ni(OH)_x versus Ni(OH)_x + MoO₄²⁻ + SO₄²⁻ using in situ ATR-FTIR spectroscopy. During the HMFOR, spectra reveal characteristic O–H stretching vibrations (3000–3400 cm⁻¹) for adsorbed OH⁻ (Fig. 5b).³⁵ Notably, both systems exhibit comparable OH⁻ peak intensities, indicating that externally added oxyanions do not impede OH⁻ adsorption. This arises because (1) OH⁻ concentration (1.0 M) vastly exceeds oxyanions (0.01 M) in alkaline electrolyte; (2) the applied electric field drives electromigration of anions, thereby accelerating OH⁻ transport to the interface.

We then employed OCP measurements to quantify the oxyanion effects on HMF adsorption strength. As shown in Fig. 5c, the HMF adsorption on Ni(OH)_x + MoO₄²⁻ + SO₄²⁻ yields a Δ_OCP of only 44 mV, significantly weaker than the post-reaction Mo-Ni₃S₂/Ni(OH)_x (226 mV). This confirms that externally added oxyanions severely suppress HMF adsorption. We deduce that the primary mechanism involves the oxyanions and OH- migrating faster than HMF to the reaction interface and thereby preferentially occupying the active sites.

Furthermore, Mo-Ni₃S₂/Ni(OH)_x (after 8 cycles, rinsed/dried) was re-activated at 1.45 V vs. RHE in 1.0 M KOH. It is observed that the current density decayed from 16 to 0.3 mA (cumulative charge: 5.1C, Fig. S29), indicating incomplete reconstruction even after 22 h constant potential electrolysis of HMF. This suggests partial retention of unoxidized Mo/S atoms in Mo-Ni₃S₂/Ni(OH)_x rather than full conversion to oxyanions. Consequently, it is necessary to investigate the ongoing impact of Mo on HMF adsorption and the electronic environments of Ni.

OCP measurements reveal that HMF adsorption strength follows the order Ni₃S₂/Ni(OH)_x (240 mV) > Mo-Ni₃S₂/Ni(OH)_x (198 mV) > Ni(OH)_x (164 mV) (Fig. 5d). The highest affinity of Ni₃S₂/Ni(OH)_x stems from the intrinsic HMF affinity of Ni₃S₂.^21,35 Critically, Mo incorporation results in moderate adsorption versus Ni₃S₂/Ni(OH)_x. As per the Sabatier principle,^61,62 this optimized strength facilitates efficient HMF adsorption and product desorption, enabling high HMFOR activity. Mechanistically, DFT shows that Mo incorporation downshifts the d-band center (ε_d) of Mo-Ni₃S₂/Ni(OH)_x: spin-up states from −1.19 eV to −1.33 eV, with analogous spin-down shifts (Fig. 5e and f). This increased distance from the Fermi level lowers antibonding orbital energies via d-band theory, thereby weakening intermediate adsorption.⁶³

In summary, the strong HMF adsorption inhibition of external oxyanions explains their limited HMFOR enhancement. Conversely, self-generated oxyanions uniquely optimize HMF binding while accelerating dehydrogenation kinetics of both Ni(OH)₂ and HMF. We attribute this discrepancy to the dynamics of oxyanion migration and their impact on HMF adsorption.

3 Conclusions

In conclusion, we have successfully constructed Mo-Ni₃S₂/Ni(OH)_x nanosheets that undergo a specific surface reconstruction to enable the dual chemisorption of autogenously generated SO₄²⁻ and MoO₄²⁻ oxyanions. The unique structure of the Mo-Ni₃S₂/Ni(OH)_x electrocatalyst results in exceptional HMFOR performance: it requires only 1.46 V vs. RHE to deliver 100 mA cm⁻² and maintains complete HMF conversion (100%), high FDCA yield (98%), and FE (94%) after 8 cycles at 1.45 V vs. RHE. It also exhibits satisfactory practical applicability in the MEA electrolyzer at high HMF concentration. Through a combination of experimental and theoretical investigations, we elucidate the bifunctional synergy of dual oxyanions in accelerating the kinetics of the HMFOR. In the co-adsorbed system, SO₄²⁻ plays a prominent role in facilitating Ni(OH)₂ dehydrogenation kinetics and the formation of active Ni³⁺ species by lowering the energy barrier. Meanwhile, MoO₄²⁻ optimizes HMF adsorption through hydrogen bonding, thereby promoting the PCET process coupled with HMF dehydrogenation. Furthermore, we highlight a crucial distinction by demonstrating that autogenetic oxyanions provide a more pronounced promotional effect than electrolyte additives, as the latter suffer from competitive adsorption. This work not only presents a high-performance electrocatalyst but, more importantly, provides fundamental insights into the explicit regulatory role of similar oxyanions, paving the way for the rational design of advanced electrocatalyst systems for biomass valorization.

Author contributions

F. B. and Y. X. conceptualized the project. Y. X. developed the methodology and wrote the original draft. Y. X. and X. L. performed the investigation. Y. X., M. L., and J. L. performed data curation and formal analysis. Y. X., F. B., and F. Y. contributed to manuscript revision. F. B. was responsible for the supervision, project administration and funding acquisition. All authors have read and agreed to submit the manuscript.

Conflicts of interest

The authors declare that they have no conflict of interest.

Data availability

All the data supporting this article have been included in the main text and the supplementary information (SI). Supplementary information: materials and methods and supplementary data. See DOI: https://doi.org/10.1039/d6sc03912j.

Acknowledgements

The authors would like to acknowledge financial support from the foundation for the National Talent Plan Project (KZ6009). The authors would also like to thank Leyan Zhang (from Scientific Compass https://www.shiyanjia.com) for providing invaluable assistance with the XPS analysis and the characterization support provided by the multi-scale environmentally controlled nanocatalyst synthesis and characterization platform of Shihezi University.

Notes and references

1. L. Guo, X. Zhang, L. Gan, L. Pan, C. Shi, Z. F. Huang, X. Zhang and J. J. Zou, Adv. Sci., 2023, 10, 2205540,  DOI:10.1002/advs.202205540.
2. D. Chen, Y. Ding, X. Cao, L. Wang, H. Lee, G. Lin, W. Li, G. Ding and L. Sun, Angew Chem. Int. Ed. Engl., 2023, 62, e202309478,  DOI:10.1002/anie.202309478.
3. C. Lei, Z. Chen, T. Jiang, S. Wang, W. Du, S. Cha, Y. Hao, R. Wang, X. Cao and M. Gong, Angew Chem. Int. Ed. Engl., 2024, 63, e202319642,  DOI:10.1002/anie.202319642.
4. J. Woo, B. C. Moon, U. Lee, H.-S. Oh, K. H. Chae, Y. Jun, B. K. Min and D. K. Lee, ACS Catal., 2022, 12, 4078–4091,  DOI:10.1021/acscatal.1c05341.
5. M. Sajid, X. Zhao and D. Liu, Green Chem., 2018, 20, 5427–5453,  10.1039/c8gc02680g.
6. H. Kim, S. Baek and W. Won, Renewable Sustainable Energy Rev., 2022, 157, 112059,  DOI:10.1016/j.rser.2021.112059.
7. S. Li, S. Wang, Y. Wang, J. He, K. Li, Y. Xu, M. Wang, S. Zhao, X. Li, X. Zhong and J. Wang, Adv. Funct. Mater., 2023, 33, 2214488,  DOI:10.1002/adfm.202214488.
8. M. Zhang, Y. Liu, B. Liu, Z. Chen, H. Xu and K. Yan, ACS Catal., 2020, 10, 5179–5189,  DOI:10.1021/acscatal.0c00007.
9. Y. Wu, Z. Hou and C. Wang, Nanoscale, 2025, 17, 3114–3122,  10.1039/d4nr04389h.
10. X. Xia, J. Xu, X. Yu, J. Yang, A. Z. Li, K. Ji, L. Li, M. Ma, Q. Shao, R. Ge and H. Duan, Sci. Bull., 2024, 69, 2870–2880,  DOI:10.1016/j.scib.2024.06.015.
11. S. Yang, Y. Guo, P. Zhao, H. Jiang, H. Shen, Z. Chen, L. Jiang, X. Xue, Q. Zhang and H. Zhang, ACS Catal., 2023, 14, 449–462,  DOI:10.1021/acscatal.3c04977.
12. 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. Engl., 2021, 60, 22908–22914,  DOI:10.1002/anie.202109211.
13. X. Y. Zhang, S. S. Yu, J. J. Chen, K. Gao, H. Q. Yu and Y. Yu, Adv. Mater., 2025, 37, e2419050,  DOI:10.1002/adma.202419050.
14. N. Zhang, Y. Zou, L. Tao, W. Chen, L. Zhou, Z. Liu, B. Zhou, G. Huang, H. Lin and S. Wang, Angew Chem. Int. Ed. Engl., 2019, 58, 15895–15903,  DOI:10.1002/anie.201908722.
15. S. Barwe, J. Weidner, S. Cychy, D. M. Morales, S. Dieckhofer, D. Hiltrop, J. Masa, M. Muhler and W. Schuhmann, Angew Chem. Int. Ed. Engl., 2018, 57, 11460–11464,  DOI:10.1002/anie.201806298.
16. L. Chen, Z. Yang, C. Yan, Y. Yin, Z. Xue, Y. Yao, S. Wang, F. Sun and T. Mu, Chem. Sci., 2024, 15, 12047–12057,  10.1039/d4sc03470h.
17. J. Wang, W. Zhao, H. Yu, W. Wang, Y. Xu, L. L. Shen, G. R. Zhang and D. Mei, Appl. Catal., B, 2024, 353, 124086,  DOI:10.1016/j.apcatb.2024.124086.
18. W. Chen, C. Xie, Y. Wang, Y. Zou, C. L. Dong, Y. C. Huang, Z. Xiao, Z. Wei, S. Du, C. Chen, B. Zhou, J. Ma and S. Wang, Chem, 2020, 6, 2974–2993,  DOI:10.1016/j.chempr.2020.07.022.
19. J. Wang, W. Zhao, Z. Zhao, H. Yu, W. Wang, Y. Xu, X. Liu, L. L. Shen, G. R. Zhang and D. Mei, Appl. Catal., B, 2025, 371, 125229,  DOI:10.1016/j.apcatb.2025.125229.
20. G. Liu, T. Nie, Z. Song, X. Sun, T. Shen, S. Bai, T. Yu, L. Zheng and Y. F. Song, Adv. Funct. Mater., 2024, 34, 2411284,  DOI:10.1002/adfm.202411284.
21. C. Liu, X. R. Shi, K. Yue, P. Wang, K. Zhan, X. Wang, B. Y. Xia and Y. Yan, Adv. Mater., 2023, 35, 2211177,  DOI:10.1002/adma.202211177.
22. M. Li, L. Liao, H. Li, Y. Shi, Z. Qie, K. Xiang, J. Zou and X. Z. Fu, Appl. Catal., B, 2026, 384, 126160,  DOI:10.1016/j.apcatb.2025.126160.
23. Y. F. Qi, K. Y. Wang, Y. Zhou, Y. Sun and C. Wang, Chem. Eng. J., 2023, 477, 146917,  DOI:10.1016/j.cej.2023.146917.
24. G. Liu, T. Nie, Z. Song, X. Sun, T. Shen, S. Bai, L. Zheng and Y. F. Song, Angew Chem. Int. Ed. Engl., 2023, 62, e202311696,  DOI:10.1002/anie.202311696.
25. B. Long, Q. Zhang, M. Yang, Y. Li, H. Liu, D. He, W. Li, Z. Ke and X. Xiao, ACS Appl. Mater. Interfaces, 2025, 17, 13794–13803,  DOI:10.1021/acsami.4c18860.
26. A. Mavrič, M. Fanetti, Y. Lin, M. Valant and C. Cui, ACS Catal., 2020, 10, 9451–9457,  DOI:10.1021/acscatal.0c01813.
27. Z. Yang, S. Wang, C. Wei, L. Chen, Z. Xue and T. Mu, Energy Environ. Sci., 2024, 17, 1603–1611,  10.1039/d3ee04543a.
28. X. Liu, J. Zhang, Y. Chen, G. Wang, J. Qiu, X. Sha, Y. Liu, M. Chang, J. Guo and D. Wang, Angew Chem. Int. Ed. Engl., 2025, 202509274,  DOI:10.1002/anie.202509274.
29. C. Wu, B. Liu, J. Wang, Y. Su, H. Yan, C. Ng, C. Li and J. Wei, Appl. Surf. Sci., 2018, 441, 1024–1033,  DOI:10.1016/j.apsusc.2018.02.076.
30. X. Luo, P. Ji, P. Wang, R. Cheng, D. Chen, C. Lin, J. Zhang, J. He, Z. Shi, N. Li, S. Xiao and S. Mu, Adv. Energy Mater., 2020, 10, 1903891,  DOI:10.1002/aenm.201903891.
31. Y. Yan, Z. Ran, T. Zeng, X. Wen, H. Xu, R. Li, C. Zhao and C. Shu, Small, 2022, 18, 2106707,  DOI:10.1002/smll.202106707.
32. Z. Wu, Y. Zhou, C. Hai, J. Zeng, Y. Sun, X. Ren, Y. Shen, X. Li and G. Zhang, Appl. Surf. Sci., 2023, 619, 156379,  DOI:10.1016/j.apsusc.2023.156379.
33. C. Yan, E. Han, X. Yang, K. Hu, H. Xu, Y. Li, Y. He and S. Lu, Ceram. Int., 2023, 49, 14155–14165,  DOI:10.1016/j.ceramint.2023.01.002.
34. Y. Zhou, Q. Li, Q. Han, L. Zhao, Y. Liu, Y. Wang, Z. Li, C. Dong, X. Sun, J. Yang, X. Zhang and F. Jiang, Small, 2023, 19, e2303742,  DOI:10.1002/smll.202303742.
35. D. Xiao, X. Bao, D. Dai, Y. Gao, S. Si, Z. Wang, Y. Liu, P. Wang, Z. Zheng, H. Cheng, Y. Dai and B. Huang, Adv. Mater., 2023, 35, e2304133,  DOI:10.1002/adma.202304133.
36. Y. Huang, Y. Chen, K. Zhu, P. Li, X. Wu and K. Yan, Results Eng., 2024, 23, 102453,  DOI:10.1016/j.rineng.2024.102453.
37. L. Chen, Z. Yang, Q. Hu, C. Yan, Y. Yao, Y. Bao, Z. Pei, T. Mu and Z. Xue, Angew Chem. Int. Ed. Engl., 2025, 64, e202511868,  DOI:10.1002/anie.202511868.
38. Z. Yang, L. Chen, Y. Yin, C. Wei, Z. Xue and T. Mu, Energy Environ. Sci., 2024, 17, 8801–8809,  10.1039/d4ee03482a.
39. H. Zhou, Y. Ren, B. Yao, Z. Li, M. Xu, L. Ma, X. Kong, L. Zheng, M. Shao and H. Duan, Nat. Commun., 2023, 14, 5621,  DOI:10.1038/s41467-023-41497-y.
40. M. L. Krebs, A. Bodach, C. Wang and F. Schüth, Green Chem., 2023, 25, 1797–1802,  10.1039/d2gc04732b.
41. J. N. Hausmann and P. W. Menezes, Angew. Chem., Int. Ed., 2022, 61, e202207279,  DOI:10.1002/anie.202207279.
42. L. Zhou, C. Yang, W. Zhu, R. Li, X. Pang, Y. Zhen, C. Wang, L. Gao, F. Fu, Z. Gao and Y. Liang, Adv. Energy Mater., 2022, 12, 2202367,  DOI:10.1002/aenm.202202367.
43. Y. Sun, J. Wang, Y. Qi, W. Li and C. Wang, Adv. Sci., 2022, 9, e2200957,  DOI:10.1002/advs.202200957.
44. S. Klaus, Y. Cai, M. W. Louie, L. Trotochaud and A. T. Bell, J. Phys. Chem. C, 2015, 119, 7243–7254,  DOI:10.1021/acs.jpcc.5b00105.
45. Q. Wen, Y. Lin, Y. Yang, R. Gao, N. Ouyang, D. Ding, Y. Liu and T. Zhai, ACS Nano, 2022, 16, 9572–9582,  DOI:10.1021/acsnano.2c02838.
46. W. Du, Y. Shi, W. Zhou, Y. Yu and B. Zhang, Angew Chem. Int. Ed. Engl., 2021, 60, 7051–7055,  DOI:10.1002/anie.202015723.
47. F. Ma, C. Zhang, W. Li, R. Hu, Z. Wang, J. Wang, J. Li, Y. Nie, Z. Zheng and X. Jiang, ACS Catal., 2025, 15, 4759–4769,  DOI:10.1021/acscatal.4c07644.
48. Z. Yang, B. Zhang, C. Yan, Z. Xue and T. Mu, Appl. Catal., B, 2023, 330, 122590,  DOI:10.1016/j.apcatb.2023.122590.
49. W. Chen, Y. Wang, B. Wu, J. Shi, Y. Li, L. Xu, C. Xie, W. Zhou, Y. C. Huang, T. Wang, S. Du, M. Song, D. Wang, C. Chen, J. Zheng, J. Liu, C. L. Dong, Y. Zou, J. Chen and S. Wang, Adv. Mater., 2022, 34, e2105320,  DOI:10.1002/adma.202105320.
50. R. Gao, T. Y. Dai, Z. Meng, X. F. Sun, D. X. Liu, M. M. Shi, H. R. Li, X. Kang, B. Bi, Y. T. Zhang, T. W. Xu, J. M. Yan and Q. Jiang, Adv. Mater., 2023, 35, e2303455,  DOI:10.1002/adma.202303455.
51. Z. Xu, W. Zuo, Y. Yu, J. Liu, G. Cheng and P. Zhao, Adv. Sci., 2024, 11, e2306758,  DOI:10.1002/advs.202306758.
52. H. M. Zhang, Nano Energy, 2024, 129, 110089,  DOI:10.1016/j.nanoen.2024.110089.
53. W. Bao, M. Yang, T. Ai, J. Han, Z. Deng, X. Zou, P. Jiang and J. Zhang, Inorg. Chem. Front., 2024, 11, 8824–8836,  10.1039/d4qi02333a.
54. Y. Shi, W. Du, W. Zhou, C. Wang, S. Lu, S. Lu and B. Zhang, Angew. Chem., Int. Ed., 2020, 59, 22470–22474,  DOI:10.1002/anie.202011097.
55. H. Liao, X. Zhang, S. Niu, P. Tan, K. Chen, Y. Liu, G. Wang, M. Liu and J. Pan, Appl. Catal., B, 2022, 307, 121150,  DOI:10.1016/j.apcatb.2022.121150.
56. N. R. Song, J. H. Moon, J. Choi, E. J. Jun, Y. Kim, S.-J. Kim, J. Y. Lee and J. Yoon, Chem. Sci., 2013, 4, 1765,  10.1039/c3sc22201b.
57. L. Zhang, L. Li, J. Liang, X. Fan, X. He, J. Chen, J. Li, Z. Li, Z. Cai, S. Sun, D. Zheng, Y. Luo, H. Yan, Q. Liu, A. A. Alshehri, X. Guo, X. Sun and B. Ying, Inorg. Chem. Front., 2023, 10, 2766–2775,  10.1039/d3qi00341h.
58. J. Wu, Z. Kong, Y. Li, Y. Lu, P. Zhou, H. Wang, L. Xu, S. Wang and Y. Zou, ACS Nano, 2022, 16, 21518–21526,  DOI:10.1021/acsnano.2c10327.
59. N. Heidary and N. Kornienko, Chem. Sci., 2020, 11, 1798–1806,  10.1039/d0sc00136h.
60. Y. Shi, W. Du, W. Zhou, C. Wang, S. Lu, S. Lu and B. Zhang, Angew Chem. Int. Ed. Engl., 2020, 59, 22470–22474,  DOI:10.1002/anie.202011097.
61. Z. W. Chen, J. Li, P. Ou, J. E. Huang, Z. Wen, L. Chen, X. Yao, G. Cai, C. C. Yang, C. V. Singh and Q. Jiang, Nat. Commun., 2024, 15, 359,  DOI:10.1038/s41467-023-44261-4.
62. J. Tian, Y. Rao, W. Shi, J. Yang, W. Ning, H. Li, Y. Yao, H. Zhou and S. Guo, Angew Chem. Int. Ed. Engl., 2023, 62, 202310894,  DOI:10.1002/anie.202310894.
63. J. K. Norskov, F. Abild-Pedersen, F. Studt and T. Bligaard, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 937–943,  DOI:10.1073/pnas.1006652108.

---