Electrochemical reconstruction- and electronegativity-induced electron redistribution for activating sulfur atoms of MoO_xS_y-based materials toward efficient hydrogen evolution

Abstract

Amorphous MoO_xS_y compounds have attracted considerable attention in the field of hydrogen evolution reaction (HER) due to their unique structural advantages, which include the retention of high electrical conductivity of O-doped MoS₂ (O-MoS₂), and abundant potential active sites of amorphous MoS_x such as molybdenum and sulfur atoms. Nevertheless, the controllable preparation of MoO_xS_y compounds, precise regulation of the localized electron surroundings of the active sites, and elucidation of a possible electrocatalytic mechanism persist as significant challenges within the field of HER. Herein, we have proposed a novel and eco-friendly strategy for electrochemical reconstruction- and electronegativity-induced electron redistribution for activating sulfur atoms of MoO_xS_y compounds toward efficient HER. The anodic oxidation/cathodic reduction in situ reconstruction strategy has been employed to successfully transform crystalline O-MoS₂ into amorphous MoO_xS_y. A thorough investigation reveals that the number of oxygen atoms in MoO_xS_y is effectively modulated by extending the oxidation time. Our results demonstrate that electrochemical reconstruction induces electron redistribution among Mo, O and S atoms. Furthermore, the typical sample (Mo₃O₂S₆/CC) might establish the electron transfer pathways along the S1–Mo–S3–Mo–O route, where oxygen-mediated charge modulation activates inert sulfur atoms through optimizing hydrogen adsorption free energy change (ΔG_H*). Consequently, Mo₃O₂S₆/CC achieves exceptional HER activity with a low overpotential of about 237.8 mV at a current density of 200 mA cm⁻², surpassing most previously reported Mo-based catalysts. Additionally, the chronoamperometric test at 100 mA cm⁻² reveals that the corresponding overpotential remains almost constant during a 24.0 h operating period, with only a marginal increase from −0.40 to −0.45 V. This signifies excellent stability. Third, the electrode pair consisting of Mo₃O₂S₆/CC and RuO₂, at 200 mA cm⁻², exhibits satisfactory overall water splitting performance of about 1.72 V at 75 °C. This work opens up a novel concept for the design of MoO_xS_y-based electrocatalysts toward achieving efficient HER.

---

Green foundation

1. The utilization of renewable energy for water electrolysis, with the objective of producing green hydrogen, constitutes a pivotal strategy within the overarching context of decarbonization. Our work demonstrates a novel strategy for electrochemical reconstruction coupled with electronegativity-induced electron redistribution toward activating sulfur atoms of MoO_xS_y compounds, significantly enhancing hydrogen evolution reaction activity.

2. The as-developed anodic oxidation/cathodic reduction strategy successfully transforms crystalline O-doped MoS₂ into amorphous MoO_xS_y compounds. Water serves as the sole solvent throughout both the synthesis of MoO_xS_y and the production of H₂. This eliminates the use of hazardous organic solvents and avoids reliance on scarce resources, thereby aligning with green and sustainable principles.

3. Utilizing renewable energy to drive electrochemical reconstruction makes our strategy greener. Exploration of hydrogen production through water electrolysis under both pure water and seawater conditions represents a pivotal research direction.

1. Introduction

The rational design of highly efficient electrocatalysts is one of the pivotal technologies in the field of hydrogen evolution reaction (HER) by water electrolysis.^1–7 The fundamental principle about structural design is the construction of active sites within the molecular skeletons, and these sites should present a suitable free energy of hydrogen atom adsorption change (ΔG_H* ≈ 0).^8–11 Amorphous MoO_xS_y compounds have attracted considerable attention in the field of HER due to their unique structural advantages. The retention of both high electrical conductivity of O-doped MoS₂ (O-MoS₂) and abundant potential active sites such as molybdenum and sulfur atoms of amorphous molybdenum sulfide (MoS_x) is exhibited by MoO_xS_y compounds.¹² It is noteworthy that the highly electronegative Mo–O center has been shown to effectively regulate the electronic structure of other sites within the molecular skeletons of amorphous compounds, thereby promoting the adsorption and desorption of hydrogen atoms for efficient production of H₂.¹³ These findings suggest that O atoms should play an important role in modulating the electronic structure of the active sites of such compounds. However, the existing synthesis methods (e.g. sputtering, electrodeposition, hydrothermal method, freeze-drying, etc.)^12,14–16 make it difficult to precisely regulate the number of O atoms as well as their coordination environments, leading to the impossibility of establishing a suitable theoretical model based on experimental data to probe the mechanism of interactions among atoms within skeletons. This, in turn, severely imposes restrictions on the rational design of amorphous MoO_xS_y electrocatalysts with highly efficient HER.

The electrochemical reconstruction technology has been developed as one of the highly effective strategies for modulating the atomic coordination and electronic structure of electrocatalysts.¹⁷ It is evident that this strategy is frequently employed in diverse fields of electrocatalysis such as HER,¹⁸ hydrogen oxidation reaction (HOR),^19,20 oxygen evolution reaction (OER),^21–23 carbon dioxide reduction reaction (CO₂-RR),^24,25 methanol oxidation reaction (MOR)^26,27 and the oxidation of organic compounds.^28–30 In comparison with other strategies for enhancing MoS_x-based HER electrocatalysts, including phase transformation engineering³¹ and defect engineering,³² the electrochemical reconstruction strategy offers the following two major advantages: (1) it enables precise in situ modulation of key structures such as the vacancy concentration and heteroatom doping level directly within the electrochemical cell, thereby allowing direct intervention in the electrocatalyst's energy band and Fermi level position. For instance, Tsai et al. employed an electrochemical cathodic reduction strategy to selectively remove basal-plane sulfur atoms from MoS₂via protonation, thereby creating sulfur vacancies. A pivotal aspect of this strategy is that, by controlling the desulfurization potential, the concentration of the sulfur vacancy can be accurately tuned. This effectively tailors the electronic structure of MoS₂ to achieve an upward shift of electronic states near the Fermi level and a narrowing bandgap toward optimizing ΔG_H*.³³ (2) Electrochemical reconstruction can selectively transform crystalline catalysts into amorphous structures, thereby exposing a greater number of active sites. For example, Hu et al. employed in situ electrochemical transmission electron microscopy with electron energy loss spectroscopy (EELS) to directly monitor the replacement of lattice S atoms by O ones, forming an amorphous layer, during the pre-electrocatalysis step of (NiCo)S_1.33.²² This amorphous layer has been demonstrated to be effective in reducing the thermodynamic energy barrier for the process of reconstituting OER active substances, increasing the amount of lattice oxygen, and achieving the promotion of OER.

Inspired by the previously reported studies, it is pertinent to inquire whether the ‘via reconstruction toward electrocatalysis’ strategy, as observed in OER, can be applied to the synthesis of highly efficient amorphous MoO_xS_y as HER electrocatalysts. In order to achieve the aforementioned objective, it is imperative to undertake a systematic investigation with the aim of effecting a controllable transformation from crystalline MoS₂ to amorphous MoO_xS_y. The present study will entail the investigations of the effect of the oxidation potential and time, and the electrochemical reconstruction mode on MoO_xS_y compounds. This could solve the difficulty of controlling the microstructure of MoO_xS_y electrocatalysts with the existing synthesis methods. Moreover, it is both possible and indeed essential to address such challenges such as the precise regulation of the local electronic environment of the active sites and the disclosure of the possible electrocatalytic mechanisms.

This work has developed a novel and eco-friendly strategy for electrochemical reconstruction coupled with oxygen electronegativity-induced electron redistribution to activate sulfur atoms of MoO_xS_y compounds toward efficient HER. This strategy involves the in situ conversion of crystalline O-MoS₂ into amorphous MoO_xS_y through anodic oxidation followed by cathodic reduction. The synthesis of O-regulated molecular skeletons is achieved by extending the oxidation time. For instance, the synthesis of Mo₃O₂S₆ on carbon cloth (Mo₃O₂S₆/CC) is conducted in accordance with the typical experimental steps. However, extending the oxidation time can result in the selective formation of Mo₃O₃S₅ on carbon cloth (Mo₃O₃S₅/CC). Our results demonstrate that the electrochemical reconstruction process does indeed induce electron redistribution among the Mo, O and S atoms that constitute the molecular skeletons. Conversely, in contrast to the conventional MoO_xS_y-based materials that rely on charge transfer at heterogeneous interfaces to construct active sites,^12,16 the MoO_xS_y electrocatalysts designed in this work can establish efficient electron transfer pathways along the S1–Mo–S3–Mo–O route within the backbone. Of significance is that O-induced charge redistribution localizes partial electrons at the S atoms and optimizes the hydrogen atom adsorption/desorption capacity, thereby achieving satisfactory HER activity. The overpotential of Mo₃O₂S₆/CC is as low as approximately 237.8 mV at 200 mA cm⁻² in 0.5 M H₂SO₄, which is significantly lower than the overpotentials of most previously reported Mo-based materials. The related electrode pair consisting of Mo₃O₂S₆/CC and RuO₂ presents excellent overall water splitting performance. Remarkably, the MoO_xS_y HER electrocatalysts obtained are entirely free of precious metals (Pt, Pd, Ir, etc.). This deliberate design is intended to avoid reliance on scarce, expensive, and geopolitically sensitive resources, aligning with the principles of sustainability and resource efficiency. Therefore, the present work provides a novel and eco-friendly synthesis of oxygen-regulated molecular skeletons for amorphous transition metal sulfides, and also opens up a novel concept for designing MoO_xS_y-based electrocatalysts toward achieving efficient HER.

2. Experimental section

2.1. Synthesis of Mo₃O₂S₆/CC

Amorphization of O-MoS₂ was conducted in a standard three-electrode system on a CS310M electrochemical workstation (Wuhan Corrtest Instruments Corp., Ltd, China) at room temperature. A piece of O-MoS₂/CC was prepared as the working electrode (area: 2.0 × 2.0 cm²), a graphite rod as the counter electrode, an Ag/AgCl electrode (saturated KCl solution) as the reference electrode and the electrolyte was 80.0 mL of 0.5 M H₂SO₄ bubbled with N₂. In typical experiments, O-MoS₂/CC was electrochemically oxidized into MoO_xS_y-t3′/CC at a constant potential of 1.15 V for a period of 1200 s. Subsequently, the electrolyte was replaced with a new one. Mo₃O₂S₆/CC (or to be called MoO_xS_y-t3/CC) was obtained by the continuous 10 cyclic voltammetry (CV) scans of MoO_xS_y-t3′/CC at the scan rate of 100.0 mV s⁻¹ from −0.7 to 0.1 V by cathodic reduction until the CV curves were essentially unchanged. Finally, the as-obtained product (Mo₃O₂S₆/CC) was washed with deionized water and dried at 60 °C under vacuum for 6 h. Unless otherwise specified, all potentials of the electrochemical data in this work are converted to the potentials versus the reversible hydrogen electrode (RHE).

2.2. Synthesis of other MoO_xS_y compounds on CC (MoO_xS_y/CC)

Other MoO_xS_y/CC samples were synthesized as follows: the oxidation of O-MoS₂/CC with oxidation times of 20, 200 and 12 000 s results in the generation of related samples, designated as MoO_xS_y-t1′/CC, MoO_xS_y-t2′/CC and MoO_xS_y-t4′/CC, respectively. These samples were then subjected to the reduction at the cathode through 10 cycles of CV scans as well, forming MoO_xS_y-t1/CC, MoO_xS_y-t2/CC and Mo₃O₃S₅/CC (or to be called MoO_xS_y-t4/CC), respectively. In order to ensure that electrochemical reconstruction is completed whilst preventing excessive reduction from causing active material loss, the number of CV reduction cycles during the electrochemical reconstruction process will be optimized. As is demonstrated in Fig. S1, following 10 CV cycles, all samples exhibited profiles that are almost identical to their ninth cycle profile. This finding suggests that the microstructures of the samples are likely to remain almost unchanged at this step.

The synthesis of MoO_xS_y-v1/CC and MoO_xS_y-v2/CC can be achieved at oxidation potentials of 0.75 and 2.42 V, respectively, and all other conditions were the same as those of MoO_xS_y-t3/CC.

Remarkably, water serves as the sole solvent throughout both the synthesis of MoO_xS_y and the production of H₂. This eliminates the use of hazardous organic solvents, reduces waste generation, and minimizes environmental contamination.

Additionally, other experimental details about materials and chemicals, characterization, electrochemical measurements, faradaic efficiency, overall water splitting tests and DFT calculations are provided in the SI.

3. Results and discussion

3.1. Modelling and DFT calculations for the Mo₃O₂S₆ model

Six models of the Mo₃O₂S₆ compound are illustrated in Fig. 1a. Each model consists of two of the smallest units. Among these models, the purple, yellow, and red balls represent molybdenum (Mo), sulfur (S), and oxygen (O) atoms, respectively. Subsequently, the energies of these models have been calculated. It is determined that the first model exhibits the lowest energy (Fig. 1b) among them, indicating that it represents the most stable structure toward the Mo₃O₂S₆ compound. To ascertain the potential of the first model for electrocatalytic hydrogen production, we further predict the S1–S5 sites, as well as the Mo1–Mo3 and O1–O2 sites of the hydrogen atom adsorption free energy change (ΔG_H*), using density functional theory (DFT) calculations, which are illustrated in Fig. S2–4, Fig. 1c, d and Table S1. Of the sites under consideration, it is the S3 site which exhibits an optimum ΔG_H*, thus suggesting a potential active center for the HER.

Fig. 1 (a) Schematic illustration of six possible models of Mo₃O₂S₆ for DFT calculations; (b) energies of six models of Mo₃O₂S₆; ΔG_H* at (c) S1, S2, S3, S4 and S5 sites and (d) Mo1, Mo2, Mo3, O1 and O2 sites of Mo₃O₂S₆ model 1; (e) XRD patterns of O-MoS₂/CC, MoO_xS_y-t3′/CC and MoO_xS_y-t3/CC; HRTEM images of (f) O-MoS₂ and (g) MoO_xS_y-t3; SEM images of (h) O-MoS₂/CC, (i) MoO_xS_y-t3′/CC and (j) MoO_xS_y-t3/CC; (k) HAADF–STEM–EDX mapping of Mo, O and S elements over MoO_xS_y-t3. Insets of (f) and (g) are the corresponding FFT patterns. Insets of (h–j) are the low-magnified SEM images of the corresponding samples. Scale bars of (f and g), (h–j) and (k) are 5, 500, and 50 nm, respectively, while those of the insets of (h–j) are 2 μm.

3.2. Characterization of Mo₃O₂S₆/CC

Guided by the aforementioned theoretical predictions, we have developed a novel and environmentally friendly strategy involving the anodic oxidation of O-MoS₂/CC followed by cathodic reduction to synthesize a series of non-precious metal electrocatalysts (MoO_xS_y/CC) in a controlled manner toward efficient HER, as illustrated in Fig. S5. The initial investigation will focus on the voltage and time parameters during the anodic oxidation step. In addition, the optimal reduction conditions are those in which the CV curves are essentially superimposed and in which the ratio of unsaturated sulfur atoms no longer exhibits significant variation. The former has previously been delineated in the Experimental section, while the latter will be discussed in detail in Fig. 2b. The oxidation voltage of 1.15 V is selected based on the CV scans of O-MoS₂/CC in a 0.5 M H₂SO₄ electrolyte (Fig. S6). An insufficient oxidation voltage such as 0.75 V results in incomplete amorphization of O-MoS₂ according to the corresponding X-ray diffraction (XRD) pattern (Fig. S7a). Conversely, elevated oxidation voltages (for instance, 2.42 V) have been shown to promote the complete amorphization of O-MoS₂ (Fig. S7a). However, this condition leads to a substantial loss of electrochemically active material, as evidenced by the corresponding scanning electron microscopy (SEM) images (Fig. S7c).

Fig. 2 (a) Mo 3d, (b) S 2p and (c) O 1s XPS spectra of O-MoS₂/CC, MoO_xS_y-t3′/CC and MoO_xS_y-t3/CC; (d) energy changes from O-MoS₂ to O₂-MoS₂ at Mo-I and Mo-II sites; (e) Raman spectra of O-MoS₂/CC (the lower figure), MoO_xS_y-t3′/CC and MoO_xS_y-t3/CC (the upper figure); (f) normalized Mo K-edge XANES spectra of O-MoS₂/CC, MoO_xS_y-t3/CC and the selected reference materials; (g) Fourier transform k³-weighted Mo K-edge EXAFS spectra of MoS₂, MoO_xS_y-t3/CC and MoO₃; (h) the pCOHP analysis of the Mo–S bonds in MoS₂, O-MoS₂, Mo₃O₂S₆, and Mo₃O₃S₅; (i) the average bond energies calculated from the pCOHP analysis of the Mo–S bonds and total bonds of MoS₂, O-MoS₂, Mo₃O₂S₆, and Mo₃O₃S₅. As is illustrated in (d), the insets are in order from left to right as follows: O-MoS₂, H₂O adsorbed on the Mo-I site (Mo-I-H₂O), the dissociation of H₂O into H* and OH* on the Mo-I site (Mo-I-OH-H), and sulfur desorption from the Mo-I site (Mo-I-SD) and O₂-MoS₂ + H₂S. Additionally, the purple, yellow and red balls represent Mo, S and O atoms, respectively. The inset in (f) is the magnified image of (f) indicated in the black dotted frame. The models of MoS₂, O-MoS₂, and Mo₃O₃S₅ for COHP analysis are displayed in Fig. S22 and the first model in Fig. S20. In (h), the pCOHP curves of the Mo–S bond are presented in blue, while the total pCOHP curves consisting of the Mo–S and Mo–O bonds are displayed in red. Bonding states are indicated on the right side of the Fermi level (dotted line), and antibonding states are on the left.

On the other hand, at an oxidation voltage of 1.15 V, a careful investigation is carried out to further reveal the effect of the oxidation time (such as 20, 200, 1200 and 12 000 s) on the microstructures of MoO_xS_y/CC. As for O-MoS₂/CC, there are three diffraction peaks located at 2θ = 14.2, 32.8 and 57.8°, corresponding to the (002), (100) and (110) planes of MoS₂, respectively. This proves that the phase structure of the original sample belongs to 2H-MoS₂.³² The SEM image of O-MoS₂/CC (Fig. 1h) shows that lots of nanosheets are assembled on CC. Here, O-MoS₂, MoO_xS_y-t1, MoO_xS_y-t2, MoO_xS_y-t3 and MoO_xS_y-t4, peeled from O-MoS₂/CC, MoO_xS_y-t1/CC, MoO_xS_y-t2/CC, MoO_xS_y-t3/CC and MoO_xS_y-t4/CC, respectively, are further characterized for their phase structures using high-resolution transmission electron microscopy (HRTEM) images. The abundant lattice fringes are clearly observed in the HRTEM image of O-MoS₂ (Fig. 1f). Its fast Fourier transform (FFT) pattern (inset of Fig. 1f) exhibits (002), (100) and (110) diffraction rings corresponding to its XRD pattern (Fig. 1e).

Furthermore, in comparison with O-MoS₂, the phase structure and morphology of MoO_xS_y-t1/CC (with an oxidation time of 20 s) remain virtually unchanged, as evidenced by the XRD pattern (Fig. S8b) and HRTEM image (Fig. S9a). As the oxidation time is extended to 200 s, the diffraction peaks are found to be absent from the XRD pattern of MoO_xS_y-t2/CC (Fig. S8b). However, it has been found that the oxidation time of 200 s makes it difficult to achieve complete amorphization. This is evidenced by the presence of a lattice spacing of 0.62 nm related to the (002) plane of 2H-MoS₂ in the HRTEM images of MoO_xS_y-t2 (Fig. S9b).

However, at the oxidation time of 1200 s, after O-MoS₂/CC is anodized into MoO_xS_y-t3′/CC, the three diffraction peaks related to O-MoS₂ disappear completely, with the exception of the peaks located at 25.7 and 43.6°, which are attributed to the (002) and (102) planes of graphitic carbon (PDF# 26-1076). Meanwhile, the diffraction peaks attributed to 2H-MoS₂ persistently fail to reemerge by the following cathodic reduction of MoO_xS_y-t3′/CC into MoO_xS_y-t3/CC. As is evident in the HRTEM image of MoO_xS_y-t3 (Fig. 1g), the lattice fringes are also absent, and its FFT pattern manifestly displays amorphous characteristics (inset of Fig. 1g). With regard to the evolution of the sample morphology on CC, the surface of MoO_xS_y-t3′ exhibits a smooth appearance (Fig. 1i), suggesting that reconstruction has occurred during anodic oxidation. Subsequent cathodic reduction of MoO_xS_y-t3′ results in the second morphological change (Fig. 1j), giving rise to an irregular shape. When the oxidation time is further extended to 12 000 s, the amorphization of O-MoS₂ into MoO_xS_y compounds is achieved with consistent success (Fig. S8b and S9d). Furthermore, the surface morphology of MoO_xS_y-t4/CC (Fig. S10f) is characterized by a smoother texture than that of MoO_xS_y-t3/CC.

From the high-angle annular dark-field scanning transmission electron microscopy energy-dispersive X-ray (HAADF-STEM-EDX) mapping (Fig. 1k), a uniform distribution of Mo, S and O elements over MoO_xS_y-t3 can be observed. Moreover, O-MoS₂ or MoO_xS_y-t4 almost presents the same HAADF-STEM-EDX mapping (Fig. S11 and 12) as MoO_xS_y-t3. Consequently, the aforementioned findings demonstrate that O-MoS₂ on substrates does undergo a transition into an amorphous state using the electrochemically driven in situ reconstruction strategy. Subsequent investigations demonstrate the possible mechanism of electrochemical reconstruction in order to identify the molecular backbone. Moreover, this may provide a research basis for revealing the electron redistribution designed in this work, which is capable of optimizing the hydrogen atom adsorption/desorption capacity of the active sites.

The full X-ray photoelectron spectroscopy (XPS) spectra^34,35 (Fig. S13) provide unequivocal evidence that both O-MoS₂/CC and MoO_xS_y-t3/CC are composed of Mo, O, S and C elements. The Mo 3d spectrum of O-MoS₂/CC predominantly exhibits characteristic peaks of Mo⁴⁺ (229.0/232.2 eV), with only minor contributions from peaks of Mo⁵⁺ (230.2/233.4 eV) and Mo⁶⁺ (232.6/235.8 eV). The Mo⁶⁺ peaks have been shown to be closely related to the oxidation of the sample in air, which is difficult to avoid during the preparation of MoS₂.^36–38 During the process of oxidation, the electron transfer occurs from the interior of the O-MoS₂/CC crystal skeleton to the exterior, driven by the positive voltage of the external circuit. This results in an increase in the binding energies of the Mo⁴⁺ and Mo⁵⁺ 3d orbitals of MoO_xS_y-t3′/CC^13,16 by 0.3–0.4 eV (Fig. 2a). In addition, a proportion of Mo⁴⁺ is oxidized to Mo⁵⁺, thereby increasing the atomic content of Mo⁵⁺ from 18.9 to 31.3%. However, no significant characteristic peak for Mo⁶⁺ appears. This finding suggests that Mo⁶⁺ is not generated during the oxidation process. The reason for the absence of Mo⁶⁺ is attributable to the oxidation potential that is selected. It is evident that this potential can only oxidize Mo⁴⁺ to Mo⁵⁺ (Fig. S6). Consequently, no Mo⁶⁺ is available following oxidation. On the other hand, Mo⁶⁺ in the initial sample may have dissolved in the acidic electrolyte, as Mo⁶⁺ is known to dissolve in acidic solutions.³⁹ At the reduction step, external electrons instead enter into MoO_xS_y-t3′/CC, and thus, the Mo⁵⁺ binding energy of MoO_xS_y-t3/CC decreases by approximately 0.3 eV compared to that of MoO_xS_y-t3′/CC. Moreover, Mo⁴⁺ is found to be more stable than Mo⁵⁺,⁴⁰ and thereby the binding energy of Mo⁴⁺ remains essentially unchanged at this step. The relative contents of Mo⁴⁺ and Mo⁵⁺ in the MoO_xS_y-t3/CC samples are 61.0 and 31.2%, respectively. Following the reduction step, the Mo 3d spectrum of MoO_xS_y-t3/CC once again exhibits minor characteristic peaks related to Mo⁶⁺. This phenomenon is considered to be closely associated with the fact that amorphous molybdenum sulfide oxidizes more readily in air than crystalline molybdenum disulfide, as evidenced by the existing literature on the subject.⁴¹ Given that MoO_xS_y-t3′/CC can only undergo electron acquisition during electrochemical reduction without electron loss, Mo⁴⁺ or Mo⁵⁺ must not have been oxidized to Mo⁶⁺ during this step.

The binding energy of the 2s orbital of sulfur in O-MoS₂/CC has been determined to be 225.9 eV (ref. 42 and 43) in Fig. 2a. After electrochemical oxidation, a significant increase in the binding energy of the S 2s orbital is observed for MoO_xS_y-t3′/CC, reaching 227.7 eV. Following the reduction step, a decrease to 226.5 eV is recorded in Fig. 2a. These results indicate that the S element does undergo an electron transfer process during redox reconstruction. The XPS analysis of the S 2p orbitals^44,45 (Fig. 2b) further reveals that the atomic ratio of unsaturated S atoms to S ones within the molecular skeletons of O-MoS₂/CC is only ∼14.5%. After the oxidation step, a 0.4 eV rise in the S 2p orbital binding energy is observed as well, indicating the electron loss of the sulfur atom. Concurrently, the ratio of unsaturated S atoms of MoO_xS_y-t3′/CC reaches up to about 80.0%. This could be because the doped-O atoms of O-MoS₂ weaken the Mo–S bond,³² leading to the susceptibility of the Mo–S bond to be broken to produce additional unsaturated S atoms. It has been demonstrated that H⁺ attacks the unsaturated S atoms on the surface of the samples during the reduction process,⁴⁶ resulting in their loss in the form of hydrogen sulfide (H₂S). This hypothesis is supported by a lead acetate paper test (Fig. S14).⁴⁷ Consequently, the atomic ratio of the unsaturated S of MoO_xS_y-t3/CC reduces to 36.4% after the reduction step.

Here, we have further refined the electrochemical reduction conditions, drawing upon the findings of S 2p XPS analysis. As is mentioned above, the atomic ratio of unsaturated sulfur atoms of MoOxSy-t3′/CC is approximately 80.0%. After a single CV reduction cycle (MoO_xS_y-t3′′/CC), the atomic ratio of unsaturated sulfur atoms undergoes a rapid decline, reaching 57.2% (Fig. S15c). Following three reduction cycles (MoO_xS_y-t3′′′/CC), a further decrease is observed, with the ratio reducing to 54.9% (Fig. S15c). Subsequently, following eight cycles (MoO_xS_y-t3′′′′/CC), the ratio of unsaturated sulfur atoms decreases to 36.8% (Fig. S15c), which is essentially consistent with the ratio of unsaturated S atoms (36.4%) observed in MoO_xS_y-t3/CC (10 CV cycles, Fig. 2b). After careful investigations, the atomic ratios of unsaturated to saturated S atoms in MoO_xS_y-t3′/CC, MoO_xS_y-t3′′/CC, MoO_xS_y-t3′′′/CC, MoO_xS_y-t3′′′′/CC and MoO_xS_y-t3/CC are 4.00 : 1.00, 1.34 : 1.00, 1.22 : 1.00, 0.58 : 1.00 and 0.57 : 1.00, respectively. This finding serves to confirm that the ratios of unsaturated sulfur atoms to saturated sulfur atoms of these samples after 10 cycles of CV remains virtually unchanged. Therefore, the optimal reduction cycles are set to 10 continuous CV scans to ensure that electrochemical reconstruction is completed whilst preventing excessive reduction from causing active material loss.

The peak located at about 531.3 eV is ascribed to the Mo–O bond according to the XPS analysis of O 1s^12,48 (Fig. 2c), clearly confirming the O atom doping into the lattice. Therefore, the initial sample is indeed O-MoS₂. More importantly, an increase in the content of the Mo–O bond is observed during both oxidation and oxidation–reduction steps. The precise data are as follows: 12.0% for O-MoS₂, 14.1% for MoO_xS_y-t3′/CC and 18.1% for MoO_xS_y-t3/CC. This phenomenon can be attributed to the replacement of sulfur atoms in the lattice of the molecular backbone structure by O atoms in the H₂O molecule during the oxidation process.^18,22 After this, partial sulfur species are expelled as H₂S after the reduction step while the O atoms doped into the catalyst skeleton can stably exist within the structure. Therefore, the content of the Mo–O bond continues to increase.

Here, the XPS data presented above are utilized to enumerate the contents of Mo (including Mo⁴⁺ and Mo⁵⁺), O and S elements within the sample's skeletal structure. This process reveals that the ratio of Mo, O and S of MoO_xS_y-t3/CC is approximately 3.00 : 1.92 : 5.64. This atomic ratio is more in line with Mo₃O₂S₆ (Fig. 1a), which provides an important experimental basis for the chemical structure of MoO_xS_y-t3/CC, that is, Mo₃O₂S₆/CC.

3.3. Possible mechanism of electrochemical reconstruction into the amorphous MoO_xS_y compounds

The present work employs theoretical calculations to offer a comprehensive analysis of the mechanism of O atom substitution for S ones (Fig. 2d, Fig. S16 and Tables S2 and 3).^18,22Fig. 2d illustrates energy changes from O-MoS₂ to O₂-MoS₂ at Mo-I and Mo-II sites in the electrolyte. Here, the Mo-I site can be used as a case to illustrate the interactions of H₂O molecules with Mo sites, which can be broadly divided into four steps. (1) The H₂O molecule exhibits a process of adsorption at the Mo-I site (Mo-I-H₂O), accompanied by an energy change of ΔE₁ = −0.70 eV. (2) The dissociation of H₂O into OH* and H* (Mo-I-OH-H) is accompanied by an increase in energy of +0.06 eV (ΔE₂). (3) The subsequent transfer reaction of the hydrogen atom within the OH* intermediate is conducted to be synergistic with the H* intermediate and the sulfur atom within the lattice, resulting in the production of H₂S gas desorption. This step constitutes sulfur desorption from the Mo-I site (to be denoted as Mo-I-SD), with ΔE₃ = +1.85 eV. (4) It is evident that the O atom occupies the vacancy created by the desorbed S atom with minimal energy requirement (to be called the doping of O atoms into the O-MoS₂ lattice). This process is accompanied by a negative energy change of −0.23 eV, also known as ΔE₄. A comparison of the energy changes across these four steps reveals that the S atom desorption step exhibits the highest energy barrier, and therefore should be the rate-determining step (RDS). For the Mo-II site, the energy change of its RDS (approximately 2.02 eV) is higher than ΔE₃ at the Mo-I site, indicating that interactions of H₂O molecules with the Mo-I site are thermodynamically more favorable. It is noteworthy that the sites of O doping determined by theoretical calculations are in good agreement with the arrangement of O atoms in the most stable model (the first model) presented in Fig. 1a, thereby providing strong support for the rationality of this model.

Raman spectral analysis is utilized to further investigate the atomic coordination environment within the as-synthesized samples (Fig. 2e). The initial sample exhibits the characteristic peaks of MoS₂ at 370 cm⁻¹ (E¹_2g), 402 cm⁻¹ (A_1g) and 817 cm⁻¹ (Mo–O–Mo).⁴⁹ However, in the case of MoO_xS_y-t3′/CC (the oxidized sample), the E¹_2g and A_1g characteristic peaks corresponding to O-MoS₂ are no longer observed, while the vibrational peaks of the Mo–S and Mo–S–Mo bonds of the amorphous state appear at 240–370 and 382–487 cm⁻¹,^34,50,51 respectively. Furthermore, the peaks appearing at 790–914 cm⁻¹ in the Raman spectrum of MoO_xS_y-t3/CC is attributed to Mo–O–Mo stretching vibrations.^52,53 The aforementioned results provide conclusive spectral evidence for the electrochemical reconstruction of O-MoS₂ into MoO_xS_y compounds, Mo₃O₂S₆.

Subsequently, the Mo K-edge X-ray absorption near-edge structure (XANES) spectra of MoS₂, MoO₃ and MoO_xS_y-t3/CC are tested in Fig. 2f. Their extended X-ray absorption fine structure (EXAFS) spectra, and the fitted results are displayed in Fig. 2g, Fig. S17 and Table S4. As for MoS₂, its peak located at 1.97 Å is ascribed to the Mo–S bond.⁵⁴ The two main peaks at 1.23 and 1.66 Å of MoO₃ correspond to the Mo–O_I and Mo–O_II bonds,^16,55 respectively. The multi-shell fitting with the first model in Fig. 1a is performed within the range of 1.00 to 3.00 Å, with the detailed data listed in Table S4. The peaks located at approximately 1.24 and 1.80 Å are unambiguously identified as the distances to the first Mo–O and Mo–S shells from the Mo R space curves, respectively. This finding serves to reinforce the validity of employing the first model of Mo₃O₂S₆ as a physical model for MoO_xS_y-t3. Furthermore, it can be posited that Mo₃O₃S₅ can be regarded as a model of MoO_xS_y-t4, in accordance with its microstructural characterization as depicted in Fig. S18–21. These experimental data provide a significant insight: the number of O atoms in amorphous MoO_xS_y compounds can be effectively modulated by extending the oxidation time.

Here, the projected crystal orbital Hamilton population (pCOHP)⁵⁶ is further calculated to evaluate the strength of the Mo–S bond for MoS₂, O-MoS₂, Mo₃O₂S₆ and Mo₃O₃S₅ (Fig. 2h). The related models are shown in Fig. S22, and the first models in Fig. 1a (Model 1) and Fig. S20 (Model 1′). The results demonstrate that the orbital interactions between Mo and S/O atoms around the Fermi energy level generate bonded and anti-bonded states. As is shown in Fig. 2h, the total pCOHP curves of O-MoS₂, Mo₃O₂S₆ and Mo₃O₃S₅ consisting of the Mo–S and Mo–O bonds are displayed in red. The pCOHP curves of the Mo–S bond of all models are shown in blue in this figure. The bond strength such as the total bonds and the Mo–S bond can be assessed by integral COHP (ICOHP) coupled with normalizing (Fig. 2i). The computational results indicate that the ICOHP values are 3.48 eV for MoS₂ and 3.24 eV for O-MoS₂. This further suggests that O doping indeed weakens the strength of the Mo–S bond³² and facilitates the structural reconstruction into MoO_xS_y compounds. The total ICOHP values of Mo₃O₂S₆ and Mo₃O₃S₅ are found to be elevated to approximately 4.49 and 4.60 eV, respectively, suggesting that the structural stability of Mo₃O₂S₆ or Mo₃O₃S₅ is significantly enhanced compared to that of O-MoS₂. Moreover, the trend of ICOHP values of the Mo–S bonds is consistent with that of the total ones. These results further validate the possibility that O-MoS₂ rearrangement can generate MoO_xS_y compounds, e.g., Mo₃O₂S₆ and Mo₃O₃S₅.

Here, the possible electrochemical reconstruction mechanism has been summarized as follows: the S atoms in O-MoS₂ are partially substituted by O atoms from water during the oxidation step. Furthermore, the O atoms in the molecular backbone have been shown to effectively weaken the Mo–S bond. This provides a robust material basis for breaking the Mo–S bonds. It can be posited that during the oxidation step, when the applied potential drives electron transfer to the external circuit, the breaking of molybdenum–sulfur bonds concomitantly generates unsaturated sulfur atoms. In the subsequent reduction step, these unsaturated S atoms are attacked by numerous H⁺ ions in the electrolyte, accompanied by the release of H₂S gas. This structural reconstruction will efficiently induce electron redistribution among the Mo, O and S atoms in the molecular skeleton, as shown in Fig. 5a. In accordance with this mechanism, the synthesis of O-regulated molecular skeletons is achieved by extending the oxidation time. Specifically, the oxidation times are set to 1200 and 12 000 s, resulting in the selective synthesis of Mo₃O₂S₆/CC and Mo₃O₃S₅/CC.

3.4. HER activity of the amorphous MoO_xS_y/CC compounds

In order to evaluate the electrocatalytic HER activity of all samples, a series of electrochemical tests are performed using a three-electrode system in 0.5 M H₂SO₄ solution. MoO_xS_y-t3/CC, that is, Mo₃O₂S₆/CC, exhibits the most excellent HER activity among all samples in terms of Fig. 3a (polarization curves normalized by geometric area). For instance, the overpotential of Mo₃O₂S₆/CC is recorded at approximately 237.8 mV at a current density of 200 mA cm⁻², which is higher than the overpotentials of other previously reported analogous electrocatalysts (Table 1). Furthermore, Mo₃O₂S₆/CC demonstrates the fastest electrode kinetics due to its low Tafel slope of ∼47.6 mV dec⁻¹ (Fig. 3b). As for MoO_xS_y-t4/CC (that is, Mo₃O₃S₅/CC), the performances are slightly lower than those of Mo₃O₂S₆/CC, for example, about 263.4 mV at the same current density and 85.1 mV dec⁻¹ (Fig. 3a and b).

Fig. 3 (a) Polarization curves, (b) Tafel plots, (c) C_dl at different scan rates, (d) ECSA-normalized polarization curves and (e) Nyquist plots of O-MoS₂/CC, MoO_xS_y-t1/CC, MoO_xS_y-t2/CC, Mo₃O₂S₆/CC and Mo₃O₃S₅/CC; (f) faradaic efficiency and the volume of producing H₂ of Mo₃O₂S₆/CC at a current density of 100 mA cm⁻²; (g) time-dependent potential curve of Mo₃O₂S₆/CC at a current density of 100 mA cm⁻² for 24.0 h; (h) XRD pattern of Mo₃O₂S₆/CC after the E–t test for 24.0 h; (i) ICP-determined Mo and S contents of Mo₃O₂S₆/CC before and after the E–t test for 24.0 h. The inset in (g) is polarization curves of Mo₃O₂S₆/CC before and after 2000 cycles. All the electrochemical experiments are performed in N₂-saturated 0.5 mol L⁻¹ H₂SO₄ solution at room temperature.

Table 1 Comparison of the HER activity of Mo₃O₂S₆/CC with other Mo-based electrocatalysts reported in the literature

Catalysts	Current density (mA cm^−2)	Overpotential (mV)	Ref.
a Co and Pd co-doped MoS2. b PH3-annealed MoS2 thin films. c 3D 1T-MoS2/CoS2 heterostructure. d W-doped MoS2 with a high concentration 1T phase. e Cu single atom substituted the Mo atom in 1T MoS2. f Oxygen and phosphorus dual-doped MoS2 nanosheets. g N- and S-doped graphene on MoOx. h Amorphous molybdenum sulfide monolayer nanosheets deposited on CC. i Amorphous molybdenum tungsten sulfide/nitrogen-doped reduced graphene oxide nanocomposites. j Carbon-supported MoSx nanocomposites with an amorphous Mo3S7Cly-like structure.
Co-Pd-MoS2^a	300	>300	62
MoS2-10 nm-250 mg^b	100	>350	63
MoS2/CoS2^c	200	∼230	64
MWS-3T^d	100	283	65
Cusub@MoS2^e	100	∼320	66
O, P-MoS2^f	50	277	67
NSGr@MoOx^g	10	367	68
ML-a-MoSx/CNT^h	300	∼220	53
a-MoWSx/N-RGO@1 : 1^i	300	∼250	69
MS-200^j	100	>270	70
	10	185
Mo3O2S6/CC	100	226	This work
	300	251

---

The performance of a catalyst is typically contingent on both its number of active sites available and its intrinsic activity. The total numbers of surface sites (n) of O-MoS₂/CC, Mo₃O₂S₆/CC and Mo₃O₃S₅/CC are confirmed by using CV in phosphate buffer (pH = 7.0) at a scan rate of 50.0 mV s⁻¹ (Fig. S23a).³⁴ The findings demonstrate that n decreases with an increase in the oxidation time (Fig. S23b). In this work, the related equation for the calculation of the number of active sites (n_as) is as follows:

n_as = n × R,

where R represents the atomic ratio of the active sites of each electrocatalyst. R is the ratio of unsaturated S atoms to all atoms within the molecular skeletons of O-MoS₂/CC, approximately 8.2% (Fig. 2b). The R values of Mo₃O₂S₆/CC and Mo₃O₃S₅/CC are defined as the atomic ratios of the S3 site in Mo₃O₂S₆ model 1 and the S2 site in the first model of Mo₃O₃S₅ (Mo₃O₃S₅ model 1′), respectively (Fig. 4e and f). Their data are both 9.1%. The corresponding results are displayed in Fig. S24a, which demonstrates that the n_as values of O-MoS₂/CC, Mo₃O₂S₆/CC and Mo₃O₃S₅/CC gradually decrease as well.

Fig. 4 (a) Polarization curves of Mo₃O₂S₆/CC and Mo₃O₃S₅/CC before and after poisoning the Mo and S sites in 0.5 mol L⁻¹ H₂SO₄ solution at room temperature; ΔG_H* of (b) S1, S2 and S3 sites of the model of Mo₃S₈, and (c) S1, S2, S3, S4 and S5 sites of Mo₃O₃S₅ model 1′; The q_Bader of the related atoms of (d) the Mo₃S₈ model, the first models of (e) Mo₃O₂S₆ and (f) Mo₃O₃S₅; LDOS of the S2 sites with the hydrogen atoms in (g) the Mo₃S₈ model, the first models of (h) Mo₃O₂S₆ and (i) Mo₃O₃S₅; LDOS of the S3 sites with the hydrogen atoms in (j) the Mo₃S₈ model, the first models of (k) Mo₃O₂S₆ and (l) Mo₃O₃S₅.

On the other hand, the electrochemically active area (ECSA) is a crucial metric that reflects the active area involved in electrocatalytic reactions in the domain of electrocatalysis research. Subsequently, the CV curves of O-MoS₂/CC, MoO_xS_y-t1/CC, MoO_xS_y-t2/CC, Mo₃O₂S₆/CC and Mo₃O₃S₅/CC are measured at different sweep speeds (Fig. S25) and their double electric layer capacitances (C_dl) are obtained in Fig. 3c. The ECSA data of all samples are further calculated according to the following equation of ECSA = C_dlC_s⁻¹ (Table S5).⁵⁷ Here, C_s is the specific capacity, 0.035 mF cm⁻². The related results present a markedly decreased trend with extended oxidation time. Here, we can rationally deduce that the decline in n_ac or ECSA is closely associated with the decrease in the loading on CC, in addition to the evolution of the surface morphology of the electrochemically reconstructed samples.

As is seen from Fig. 3d, the polarization curve normalized by the ECSA of each electrochemically reconstructed sample is higher than that of O-MoS₂/CC. Remarkably, Mo₃O₂S₆/CC presents lower HER activity normalized by ECSA in comparison with Mo₃O₃S₅/CC. To further investigate the intrinsic activities of MoO_xS_y/CC samples, the turnover frequency (TOF) will be calculated with the following equation:³⁴

TOF = j/(2 × n_as × F),

where j and F represent the current density at −0.25 V in the polarization curves and the Faraday constant, respectively. This finding serves to further substantiate the hypothesis that Mo₃O₃S₅/CC exhibits the highest intrinsic activity among all samples (Fig. S24b). As a consequence, we conclude that Mo₃O₂S₆/CC presents higher HER activity (Fig. 3a) than Mo₃O₃S₅/CC, which is attributed to the more abundant active sites instead of higher intrinsic activity.

O-MoS₂/CC, MoO_xS_y-t1/CC, MoO_xS_y-t2/CC, Mo₃O₂S₆/CC and Mo₃O₃S₅/CC exhibit satisfactory electrical conductivity. Nevertheless, the related data of charge transfer impedance (R_ct) display an increase in conjunction with the extension of the oxidation time (Fig. 3e and Table S5). In our viewpoint, the diminished efficiency of charge transfer during the HER process originates from the decreased n_ac and ECSA, as well as the evolution of the surface morphology of these samples.^58,59 To assess the faradaic efficiency (FE) of Mo₃O₂S₆/CC, the hydrogen (H₂) produced at constant potential is collected through the drainage method (Fig. S26) and the experimentally generated H₂ moles are calculated.³⁵ The result shows that the FE of Mo₃O₂S₆/CC is as high as ∼97.2% (Fig. 3f).

Besides the electrocatalytic activity, the stability is also a significant factor in evaluating the performance of an electrocatalyst. As is demonstrated in Fig. 3g, the chronoamperometric test at a constant current density of 100 mA cm⁻² reveals that the corresponding overpotential remains almost constant during a 24.0 h operating period, with only a marginal increase from −0.40 to −0.45 V. The polarization curves (inset of Fig. 3g) after conducting 2000 CV tests show that the overpotential increases by only 23 mV at 300 mA cm⁻². These results signify that Mo₃O₂S₆/CC exhibits excellent electrochemical durability. The microstructural characterization of Mo₃O₂S₆/CC following the stability test is as follows: the XRD pattern (Fig. 3h) demonstrates that Mo₃O₂S₆/CC retains its original amorphous characteristics; the SEM images (Fig. S27) show that its surface morphology does not undergo significant change over the duration of the testing period. The related inductively coupled plasma-optical emission spectroscopy (ICP-OES) results are displayed in Fig. 3i. After stability testing, the Mo content in Mo₃O₂S₆/CC decreases from 70.4 to 46.6 μmol, and the S content decreases from 125.7 to 85.1 μmol, indicating that it undergoes a certain degree of dissolution in the acidic electrolyte.

In view of the fact that the working temperature of the electrode will be increased for industrial applications, the polarization curves of Mo₃O₂S₆/CC are tested at 50 and 75 °C (Fig. S28a). The corresponding results demonstrate that Mo₃O₂S₆/CC still exhibits excellent HER activity. In addition, Mo₃O₂S₆/CC is integrated with the RuO₂ electrocatalyst (a loading of 10 mg cm⁻²) into a Membrane Electrode Assembly (MEA) to evaluate the overall water splitting performance (Fig. S28b, see the inset of Fig. S28b for the related experimental configuration). The experimental findings demonstrate that the electrode pair consisting of Mo₃O₂S₆/CC and RuO₂ (Mo₃O₂S₆/CC||RuO₂), operating at a current density of 200 mA cm⁻², exhibits satisfactory performances of about 1.80 and 1.72 V at room temperature and 75 °C, respectively, thereby suggesting the potential application of Mo₃O₂S₆/CC in industrial-scale electrolysis systems.

3.5. Possible mechanism of electrochemical reconstruction- and electronegativity-induced electron redistribution for activating sulfur atoms

The active sites of Mo₃O₂S₆/CC and Mo₃O₃S₅/CC are determined by poisoning experiments (Fig. 4a) in order to investigate their mechanism of producing H₂. It has been indicated that SCN⁻ ions form coordination complexes with molybdenum ions in acidic environments, thereby poisoning the Mo sites.^60,61 On the other hand, dodecanethiols have been shown to be adsorbed on the surface of the electrocatalyst to promote the dissociation of the Mo–S bonds, thus inhibiting the HER activity of S atoms.^60,61 Subsequent to the poisoning of the Mo site experiments, there is a slight decline in the HER activities of Mo₃O₂S₆/CC or Mo₃O₃S₅/CC, whereas their activities exhibit a significant decrease after the poisoning of the S sites. This finding thus confirms that the active sites of Mo₃O₂S₆/CC and Mo₃O₃S₅/CC should originate from the S sites rather than the Mo ones.

Subsequently, we calculate ΔG_H* at different sites for the Mo₃S₈ model and Mo₃O₃S₅ model 1′, as illustrated in Fig. 4b, c, and Fig. S29–35. In the majority of cases, the basal plane of O-MoS₂ remains inert due to a ΔG_H* value of up to 2.15 eV.³² The Mo₃S₈ model may exhibit an electrocatalytic center at the S3 site, as its ΔG_H* is decreased to 0.28 eV (Fig. 4b). The S3 site continues to demonstrate the highest catalytic activity in comparison with other sites within the first model of Mo₃O₂S₆ (Mo₃O₂S₆ model 1) due to its ΔG_H* further decreasing to approximately −0.24 eV (Fig. 1c). However, the S2 site at Mo₃O₃S₅ model 1′ (Fig. 4c) exhibits a ΔG_H* value of 0.01 eV, approaching zero. This suggests that the S2 site may in fact exhibit superior HER activity. It is evident that this finding provides a reasonable explanation for the higher intrinsic activity of Mo₃O₃S₅/CC compared to Mo₃O₂S₆/CC.

It is further hypothesized that the intrinsic activity of MoO_xS_y/CC compounds should be closely related to electrochemical reconstruction and electronegativity-induced electron redistribution. Therefore, in order to ascertain the underlying reasons for the synergistic regulation of electron redistribution towards achieving excellent intrinsic HER activity by electrochemical reconstruction and the electronegativity of O atoms, the Bader charges (q_Bader) of the atoms concerned in the above models are further calculated. The q_Bader values of O-MoS₂ before and after electrochemical reconstruction (Mo₃O₂S₆ model 1) are shown in Fig. S36 and Fig. 4e, respectively. In comparison with the related data of O-MoS₂, the Mo and O atoms of Mo₃O₂S₆ exhibit increased q_Bader values, approximately 0.35 and 0.08e, respectively, while the increase in its S atoms ranges from 0.00 to 0.32e. The aforementioned data unequivocally demonstrates that the electrochemical reconstruction can efficiently induce the electron redistribution among these atoms. In this study, we conducted a detailed analysis of the q_Bader value of Mo₃S₈, Mo₃O₂S₆ model 1 and Mo₃O₃S₅ model 1′, with the objective of investigating the impact of the number of O atoms in MoO_xS_y on the electron redistribution. As demonstrated in Fig. 4d–f, the electron loss (∼0.11e) at the S1 site on Mo₃O₂S₆ model 1 or Mo₃O₃S₅ model 1′ is observed in comparison with the S1 site on the Mo₃S₈ model. Consequently, the adsorption capacity of these S1 sites for hydrogen atoms is diminished, resulting in the related ΔG_H* data of −0.26 eV for Mo₃O₂S₆ model 1 and −0.18 eV for Mo₃O₃S₅ model 1′. A similar trend is exhibited by the S2 sites in Mo₃O₂S₆ and Mo₃O₃S₅ models, with these sites also losing electrons and exhibiting the corresponding rise in their ΔG_H* to 0.31 and 0.01 eV, respectively. Conversely, the S3 sites in the Mo₃O₂S₆ and Mo₃O₃S₅ models are both gaining electrons, indicating that their adsorption capacity for hydrogen atoms is enhanced (−0.24 and −0.26 eV). In conclusion, the electronic states of the S sites in the molecular backbone structure have been significantly modulated through the effect of electron redistribution.

The S1 and S2 sites in the Mo₃O₂S₆ and Mo₃O₃S₅ models undergo a positive shift in ΔG_H* as a consequence of electron loss, while the S3 sites shifted in a negative direction in ΔG_H* due to the electron enrichment. In our viewpoint, the electron redistribution of this nature may be realized on the basis of the O–Mo–S electron transport pathway within the backbone. It is evident that the high electronegativity of the O atom is another primary factor that governs the migration of electrons. The electrons migrate along the O–Mo–S1 or O–Mo–S2 pathway from the S1 and S2 sites (Fig. 4e and f). Conversely, partial electrons can be localized at the S3 sites through electron transport along the O–Mo–S3–Mo–S1 routes (Fig. 4e and f), thereby leading to electron enrichment. It is precisely this electron redistribution effect, based on the O–Mo–S transport routes, that optimizes the adsorption capacity of the S3 site of Mo₃O₂S₆ and the S2 site of Mo₃O₃S₅ for hydrogen atoms, thereby endowing them with excellent HER activities.

A comprehensive investigation is conducted on the adsorption capacity of hydrogen atoms at the S2 and S3 sites in the Mo₃S₈, Mo₃O₂S₆ and Mo₃O₃S₅ models using the local density of states (LDOS) (Fig. 4g–l). The LDOS calculations demonstrate that the 1s orbitals of the hydrogen atoms can overlap with the S-3p_y and S-3p_z ones of the S2 site at the Mo₃S₈ model, respectively (Fig. 4g); similar outcomes can be obtained for its S3 site (Fig. 4j). As demonstrated in Fig. 4g, the overlap between the H1s and S-3p_y, S-3p_z orbitals of the S2 site is observed at −5.90 eV. In contrast, the overlap for the S3 site negatively shifts to approximately −6.40 eV, as illustrated in Fig. 4j. This finding suggests that the S3 site has stronger affinity for hydrogen atoms, which further implies that the S3 site should be the possible catalytic center in the Mo₃S₈ model. These results are in good agreement with the ΔG_H* values in Fig. 4a (−0.90 eV for the S2 site and 0.28 eV for the S3 site). In the Mo₃O₂S₆ model, the S2 site also exhibits S-3p_y, S-3p_z orbitals overlapping with H 1s (Fig. 4h), with their overlapped peaks occurring at approximately −6.78 eV. The S3 site in this model exhibits analogous orbital selectivity, with its S-3p_y, S-3p_z orbitals overlapping with the H 1s orbitals at −6.09 eV or so (Fig. 4k). This finding serves to further substantiate the hypothesis that the S3 site exhibits superior adsorption capacity for hydrogen atoms in comparison with its S2 site, thereby manifesting enhanced HER activity (Fig. 1c). As is demonstrated in Fig. 4c, the S2 site in the Mo₃O₃S₅ model exhibits optimal ΔG_H*. The S-3p_y and S-3p_z orbitals of this site present an overlap with H 1s at −6.20 eV, as illustrated in Fig. 4i. It is noteworthy that the S3 site exhibits an excessive adsorption capacity for hydrogen atoms, which can be attributed to the upward shift of its overlap peak to −5.92 eV (Fig. 4l). This phenomenon contributes to the complexity of desorbing hydrogen atoms at the S3 site. Consequently, the S2 site exhibits superior HER activity.

The possible catalytic mechanism may be outlined as follows: within the Mo₃S₈ model, the migration of electrons from Mo atoms to the S atoms can be ascribed to the high electronegativity of the S element. Furthermore, due to the absence of electron redistribution driven by the action of other elements with higher electronegativity, the q_Bader value of the sulfur atoms within Mo₃S₈ is either too high or too low to easily achieve a suitable ΔG_H* value, as demonstrated in Fig. 5b. As for Mo₃O₂S₆ model 1, the O-induced charge redistribution localizes partial electrons at the S3 site (to be termed ‘the rational electron localization’). This is due to the electron transport pathways along the O–Mo–S3–Mo–S1 route being established within Mo₃O₂S₆ model 1. It is imperative to acknowledge that this optimized electronic structure significantly enhances the hydrogen atom adsorption/desorption capacity from the S3 site (Fig. 5c), thereby realizing satisfactory HER activity of Mo₃O₂S₆/CC. In contrast, the S2 site exhibits excessive electron loss along the O–Mo–S2 electron transport pathways due to the attraction of O atoms from the two structural units. Nevertheless, the effect of such O–Mo–S electron transport pathways on charge redistribution is not observed in the Mo₃S₈ model due to the absence of an O atom. This catalytic mechanism is basically applicable to Mo₃O₃S₅ model 1′ as well. Differently, the presence of more O atoms induces electron over-localization at the S3 site, thereby resulting in heightened adsorption affinity for hydrogen atoms. Interestingly, the S2 site in this model is activated through rational charge modulation based on the O–Mo–S2 electron transport pathways (Fig. 5d).

Fig. 5 (a) Schematic illustration of the electrochemical reconstruction-induced electron redistribution; schematic illustration of the possible catalytic mechanism of (b) the Mo₃S₈ model; the possible catalytic mechanism based on the electron redistribution via various O–Mo–S electron transport pathways for activating the different sulfur sites in (c) the Mo₃O₂S₆ and (d) Mo₃O₃S₅ models toward efficient HER.

One more thing, the synthesis route of MoO_xS_y/CC possesses four main characteristics: decarbonization, environmental friendliness, cost-effectiveness and sustainability compared to preparation methods of the recently reported transition metal-based counterparts (Table S6). They are primarily manifested as follows: (1) the utilization of renewable energy for water electrolysis, with the objective of producing green H₂, constitutes a pivotal strategy within the overarching context of decarbonization; (2) water serves as the sole solvent throughout both the synthesis of MoO_xS_y and the production of H₂. This eliminates the use of hazardous organic solvents, reduces waste generation, and minimizes environmental contamination; (3) this deliberate design is intended to avoid reliance on scarce, expensive, and geopolitically sensitive resources, aligning with the principles of sustainability and resource efficiency.

4. Conclusion

In summary, the present work focuses on the development of a novel and eco-friendly strategy for electrochemical reconstruction coupled with electronegativity to synergistically induce electron redistribution for the activation of sulfur atoms in MoO_xS_y-based materials toward efficient HER. In this work, phase transformation from crystalline O-MoS₂ to amorphous MoO_xS_y compounds is successfully achieved by an innovative anodic oxidation/cathodic reduction in situ reconstruction strategy. The results obtained unequivocally demonstrate that electrochemical reconstruction can efficiently induce electron redistribution among Mo, O and S atoms. It is noteworthy that the Bader charge analysis demonstrates that efficient electron transfer pathways along the O–Mo–S3–Mo–S1 route can be established within Mo₃O₂S₆. The O-induced charge redistribution localizes partial electrons at the S3 site and optimizes their ΔG_H* to achieve fast HER dynamics. Conversely, the presence of additional O atoms has been observed to induce electron over-localization at the S3 site within Mo₃O₃S₅, thereby resulting in excessively strong adsorption for hydrogen atoms. Interestingly, the S2 site is activated through rational charge modulation based on the O–Mo–S2 electron transport pathways. The overpotential of the as-synthesized MoO_xS_y-based electrocatalyst is as low as approximately 237.8 mV at a current density of 200 mA cm⁻² in 0.5 M H₂SO₄, as demonstrated by the aforementioned benefits. Furthermore, the electrode pair consisting of Mo₃O₂S₆/CC||RuO₂ exhibits satisfactory overall water splitting performance of about 1.72 V at the same current density at 75 °C. The present study proposes a strategy for the construction of O–Mo–S electron transport pathways, with the aim of achieving efficient HER through the rational design of amorphous MoO_xS_y electrocatalysts. This strategy promises important applications in the fields of electrocatalysis and energy conversion.

Author contributions

Xiao-Yu Yang: conceptualization, data curation, investigation, software, writing – original draft, and writing – review & editing. Rui-Yuan Li: investigation. Zhan Liu: investigation. Chun-Mu Guo: investigation. Xiao-Yun Li: investigation. Zhao Deng: investigation. Jia-Min Lyu: investigation. Ming-Hui Sun: investigation. Shen Yu: investigation. Yu Li: investigation. Yi-Yong Huang: investigation. Li-Hua Chen: conceptualization, funding, acquisition, project administration, supervision, writing – original draft, and writing – review & editing. Kai Liu: conceptualization, supervision, writing – original draft, and writing – review & editing. Bao-Lian Su: funding acquisition and investigation. Yi-Long Wang: conceptualization, investigation, project administration, supervision, writing – original draft, and writing – review & editing.

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary information (SI). Additional experimental sections and computation details, characterization and electrochemical measurements; CV curves; schematic illustration of the first model of Mo₃O₂S₆ for the adsorption of the H atoms at different sites; Schematic illustration of the synthesis; XRD patterns; SEM images; HRTEM images; FFT patterns; HAADF-STEM-EDX mapping; XPS full spectra; Photographs of lead acetate test paper detection during the reduction step of reconstitution; schematic illustration of Mo-II-H₂O, Mo-II-OH-H and transition state; k space curve; Mo 3d, S 2p and O 1s XPS spectra; Raman spectra; schematic illustration of six possible models of Mo₃O₃S₅; energies of six models of Mo₃O₃S₅; schematic illustration of models of MoS₂ and O-MoS₂ for pCOHP; the total number of surface sites; the number of active sites; photographs of the working electrode of Mo₃O₂S₆/CC for testing Faraday efficiency; polarization curves; overall water splitting performance; schematic illustration of the Mo₃S₈ model; DGH* at Mo1 and Mo2 sites of Mo₃S₈; schematic illustration of the Mo₃S₈ model for the adsorption of the H atoms at different sites; DGH* at different sites of the first model of Mo₃O₃S₅; schematic illustration of the first model of Mo₃O₃S₅ for the adsorption of the H atoms at different sites; Bader charge of the related atoms of O-MoS₂; the calculated adsorption energy ΔE(ads) of H₂O on Mo-I and Mo-II sites of O-MoS₂; the calculated energy changes (ΔE) from O-MoS₂ to O₂-MoS₂ at Mo-I and Mo-II sites; EXAFS fitting results for the local structure parameters around Mo; The Cdl, ECSA and Rct (PDF). See DOI: https://doi.org/10.1039/d5gc03386a.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (22293022 and U22B6011) and the Program of Introducing Talents of Discipline to Universities-Plan 111 (Grant No. B20002) from the Ministry of Science and Technology and the Ministry of Education of China. This research was also supported by the European Commission Interreg V France-Wallonie-Vlaanderen project “DepollutAir”, the Program Win2Wal (TCHARBONACTIF: 2110120), the Wallonia Region of Belgium and the National Key R&D Program Intergovernmental Technological Innovation Special Cooperation Project Wallonia-Brussels/China (MOST) (SUB/2021/IND493971/524448). B.-L. Su acknowledges a Clare Hall Life Membership at the University of Cambridge. L.-H. Chen acknowledges the Hubei Provincial Department of Education for the “Chutian Scholar” program.

References

1. Y. Liu, G. Yu, G. Li, Y. Sun, T. Asefa, W. Chen and X. Zou, Angew. Chem., 2015, 127, 10902–10907.
2. W. Liu, X. Niu, J. Tang, Q. Liu, J. Luo, X. Liu and Y. Zhou, Chem. Synth., 2023, 3, 44.
3. M. Fan, J. Guo, G. Fang, H. Tian, Y. You, Z. Huang, J. Huang, H. Jiang, W. Xu and J. Wan, Chem. Synth., 2024, 4, 36.
4. L. Zhang, J. Zhu, X. Li, S. Mu, F. Verpoort, J. Xue, Z. Kou and J. Wang, Interdiscip. Mater., 2022, 1, 51–87.
5. L. Tang, X. Meng, D. Deng and X. Bao, Adv. Mater., 2019, 31, 1901996.
6. H. Zhao, X. Zhang, C. Yu, W. Gao, X. Chen, H. Lan, X. Xin, K. Liu and D. Ma, Green Chem., 2025, 27, 8569–8575.
7. X. Huang, M. Leng, W. Xiao, M. Li, J. Ding, T. L. Tan, W. S. V. Lee and J. Xue, Adv. Funct. Mater., 2017, 27, 1604943.
8. Y.-C. Zhang, M. Zhao, J. Wu, Y. Wang, L. Zheng, F. Gu, J.-J. Zou, J. Gao and X.-D. Zhu, ACS Catal., 2024, 14, 7867–7876.
9. X. Liang, L. Shi, Y. Liu, H. Chen, R. Si, W. Yan, Q. Zhang, G.-D. Li, L. Yang and X. Zou, Angew. Chem., Int. Ed., 2019, 23, 7631–7635.
10. R. Paul, Q. Zhai, A. K. Roy and L. Dai, Interdiscip. Mater., 2022, 1, 28–50.
11. X. Chen, X. Wei, X. Zhang, J. Wang, Z. Wang, S. Wei, S. Liu, B. Liao, Z. Sun and X. Lu, Green Chem., 2025, 27, 4959–4985.
12. Y. Zhou, W. Hao, X. Zhao, J. Zhou, H. Yu, B. Lin, Z. Liu, S. J. Pennycook, S. Li and H. J. Fan, Adv. Mater., 2022, 34, 2100537.
13. B. Seo, G. Y. Jung, S. J. Lee, D. S. Baek, Y. J. Sa, H. W. Ban, J. S. Son, K. Park, S. K. Kwak and S. H. Joo, ACS Catal., 2020, 10, 652–662.
14. A. Levasseur, E. Schmidt, G. Meunier, D. Gonbeau, L. Benoist and G. Pfister-Guillouzo, J. Power Sources, 1995, 54, 352–355.
15. X.-Q. Bao, D. Y. Petrovykh, P. Alpuim, D. G. Stroppa, N. Guldris, H. Fonseca, M. Costa, J. Gaspar, C. Jin and L. Liu, Nano Energy, 2015, 16, 130–142.
16. Z. Yu, H. Yao, Y. Yang, M. Yuan, C. Li, H. He, T.-S. Chan, D. Yan, S. Ma, P. Zapol and M. G. Kanatzidis, Chem. Mater., 2022, 34, 798–808.
17. S. Li, D. Tang and X. Jing, Chem. Synth., 2024, 4, 70.
18. Y. Deng, Z. Liu, A. Wang, D. Sun, Y. Chen, L. Yang, J. Pang, H. Li, H. Li, H. Liu and W. Zhou, Nano Energy, 2019, 62, 338–347.
19. J. Yue, Y. Li, C. Yang and W. Luo, Angew. Chem., Int. Ed., 2025, 64, e202415447.
20. G. Meng, H. Yao, H. Tian, F. Kong, X. Cui, S. Cao, Y. Chen, Z. Chang, C. Chen and J. Shi, J. Mater. Chem. A, 2022, 10, 622–631.
21. T. Zhao, B. Gong, G. Xu, J. Jiang and L. Zhang, Chin. J. Catal., 2024, 61, 269–280.
22. Y. Hu, Y. Zheng, J. Jin, Y. Wang, Y. Peng, J. Yin, W. Shen, Y. Hou, L. Zhu, L. An, M. Lu, P. Xi and C.-H. Yan, Nat. Commun., 2023, 14, 1949.
23. J. Li, L. Wang, H. He, Y. Chen, Z. Gao, N. Ma, B. Wang, L. Zheng, R. Li, Y. Wei, J. Xu, Y. Xu, B. Cheng, Z. Yin and D. Ma, Nano Res., 2022, 15, 4986–4995.
24. S. H. Lee, J. C. Lin, M. Farmand, A. T. Landers, J. T. Feaster, J. E. Avilés Acosta, J. W. Beeman, Y. Ye, J. Yano, A. Mehta, R. C. Davis, T. F. Jaramillo, C. Hahn and W. S. Drisdell, J. Am. Chem. Soc., 2021, 143, 588–592.
25. M.-N. Zhu, B.-W. Zhang, M.-R. Gao, P.-F. Sui, C. Xu, L. Gong, H. Zeng, K. Shankar, S. Bergens and J.-L. Luo, Appl. Catal., B, 2022, 306, 121101.
26. X. Wang, M. Xie, F. Lyu, Y.-M. Yiu, Z. Wang, J. Chen, L.-Y. Chang, Y. Xia, Q. Zhong, M. Chu, H. Yang, T. Cheng, T.-K. Sham and Q. Zhang, Nano Lett., 2020, 20, 7751–7759.
27. L. Zhao, Q. Sun, M. Li, Y. Zhong, P. Shen, Y. Lin and K. Xu, Sci. China Mater., 2023, 66, 1820–1828.
28. H. Zhou, Z. Li, S. Xu, L. Lu, M. Xu, K. Ji, R. Ge, Y. Yan, L. Ma, X. Kong, L. Zheng and H. Duan, Angew. Chem., Int. Ed., 2021, 60, 8976–8982.
29. R. Luo, Y. Li, L. Xing, N. Wang, R. Zhong, Z. Qian, C. Du, G. Yin, Y. Wang and L. Du, Appl. Catal., B, 2022, 311, 121357.
30. L. Wang, Y. Yan, R. Li, X. Han, J. Li, T. Ran, J. Li, B. Xiong, X. Song, Z. Yin, H. Wang, Q. Zhu, B. Cheng and Z. Yin, Chin. Chem. Lett., 2024, 35, 110011.
31. X. Guo, E. Song, W. Zhao, S. Xu, W. Zhao, Y. Lei, Y. Fang, J. Liu and F. Huang, Nat. Commun., 2022, 13, 5954.
32. T. Lian, X. Li, Y. Wang, S. Zhu, X. Yang, Z. Liu, C. Ye, J. Liu, Y. Li, B. Su and L. Chen, ACS Appl. Mater. Interfaces, 2022, 14, 30746–30759.
33. C. Tsai, H. Li, S. Park, J. Park, H. S. Han, J. K. Nørskov, X. Zheng and F. Abild-Pedersen, Nat. Commun., 2017, 8, 15113.
34. Y.-X. Guo, M. Yang, Z. Liu, X.-Y. Yang, Y. Xiao, X.-Y. Li, C.-F. Ye, Y. Li, J.-P. Liu, B.-L. Su, L.-H. Chen and Y.-L. Wang, Electrochim. Acta, 2023, 449, 142191.
35. R.-Y. Li, Y.-X. Guo, X.-Y. Yang, Z. Liu, X.-Y. Li, Y. Li, Y.-Y. Huang, J.-P. Liu, S. Yu, M.-H. Sun, L.-H. Chen, B.-L. Su and Y.-L. Wang, J. Mater. Chem. A, 2024, 12, 29615–29629.
36. Y. Zhan, Y. Li, Z. Yang, X. Wu, M. Ge, X. Zhou, J. Hou, X. Zheng, Y. Lai, R. Pang, H. Duan, X. Chen, H. Nie and S. Huang, Adv. Sci., 2019, 6, 1801663.
37. Y. Li, Y. Wang, B. Pattengale, J. Yin, L. An, F. Cheng, Y. Li, J. Huang and P. Xi, Nanoscale, 2017, 9, 9230–9237.
38. Y.-H. Choi, J. Cho, A. M. Lunsford, M. Al-Hashimi, L. Fang and S. Banerjee, J. Mater. Chem. A, 2017, 5, 5129–5141.
39. D. Escalera-López, Z. Lou and N. V. Rees, Adv. Energy Mater., 2019, 9, 1802614.
40. C. R. Clayton and Y. C. Lu, Surf. Interface Anal., 1989, 14, 66–70.
41. M. R. Vazirisereshk, A. Martini, D. A. Strubbe and M. Z. Baykara, Lubricants, 2019, 7, 57.
42. L. R. L. Ting, Y. Deng, L. Ma, Y.-J. Zhang, A. A. Peterson and B. S. Yeo, ACS Catal., 2016, 6, 861–867.
43. J. Kibsgaard, T. F. Jaramillo and F. Besenbacher, Nat. Chem., 2014, 6, 248–253.
44. P. Deng, P. Wang, X. Wang, F. Chen and H. Yu, Nano Res., 2023, 16, 8977–8986.
45. L. Benoist, D. Gonbeau, G. Pfister-Guillouzo, E. Schmidt, G. Meunier and A. Levasseur, Thin Solid Films, 1995, 258, 110–114.
46. F. Xi, P. Bogdanoff, K. Harbauer, P. Plate, C. Höhn, J. Rappich, B. Wang, X. Han, R. Van De Krol and S. Fiechter, ACS Catal., 2019, 9, 2368–2380.
47. N. Abidi, A. Sahu, P. Raybaud and S. N. Steinmann, ACS Catal., 2023, 13, 15290–15300.
48. Z. Luo, R. Miao, T. D. Huan, I. M. Mosa, A. S. Poyraz, W. Zhong, J. E. Cloud, D. A. Kriz, S. Thanneeru, J. He, Y. Zhang, R. Ramprasad and S. L. Suib, Adv. Energy Mater., 2016, 6, 1600528.
49. L. Gan, J. Liu, Y. Ren, Z. Xiong, K. Chen and Y. Zhao, Carbon Lett., 2023, 33, 1367–1380.
50. C.-H. Lee, S. Lee, G.-S. Kang, Y.-K. Lee, G. G. Park, D. C. Lee and H.-I. Joh, Appl. Catal., B, 2019, 258, 117995.
51. J. A. Franz, J. C. Birnbaum, D. S. Kolwaite, J. C. Linehan, D. M. Camaioni and M. Dupuis, J. Am. Chem. Soc., 2004, 126, 6680–6691.
52. M. Vasilopoulou, A. M. Douvas, D. G. Georgiadou, L. C. Palilis, S. Kennou, L. Sygellou, A. Soultati, I. Kostis, G. Papadimitropoulos, D. Davazoglou and P. Argitis, J. Am. Chem. Soc., 2012, 134, 16178–16187.
53. D. Wang, H. Li, N. Du and W. Hou, Chem. Eng. J., 2020, 398, 125685.
54. B. Lassalle-Kaiser, D. Merki, H. Vrubel, S. Gul, V. K. Yachandra, X. Hu and J. Yano, J. Am. Chem. Soc., 2015, 137, 314–321.
55. J. Mao, W. Chen, D. He, J. Wan, J. Pei, J. Dong, Y. Wang, P. An, Z. Jin, W. Xing, H. Tang, Z. Zhuang, X. Liang, Y. Huang, G. Zhou, L. Wang, D. Wang and Y. Li, Sci. Adv., 2017, 3, e1603068.
56. G. Cai, C. Hua, H. Ren, R. Yu, D. Xu, M. A. Khan, J. Guo, Y. Sun, Y. Tang, H. Qian, Z. Xia, D. Ye, J. Zhang and H. Zhao, J. Colloid Interface Sci., 2025, 678, 599–608.
57. X. Zhao, Y. Li, C. Zhao and Z. Liu, Small, 2020, 16, 2004973.
58. Z. W. Seh, J. Kibsgaard, C. F. Dickens, I. Chorkendorff, J. K. Nørskov and T. F. Jaramillo, Science, 2017, 355, eaad4998.
59. Y. Li, H. Wang, L. Xie, Y. Liang, G. Hong and H. Dai, J. Am. Chem. Soc., 2011, 133, 7296–7299.
60. P. Zeng, Y. Meng, Z. Liu, G. Sun, X. Li, X. Yang, C. Ye, Y. Li, J. Liu, L. Chen, B. Su and Y. Wang, Small, 2023, 19, 2301279.
61. J. Yin, Q. Fan, Y. Li, F. Cheng, P. Zhou, P. Xi and S. Sun, J. Am. Chem. Soc., 2016, 138, 14546–14549.
62. W. Yang, S. Zhang, Q. Chen, C. Zhang, Y. Wei, H. Jiang, Y. Lin, M. Zhao, Q. He, X. Wang, Y. Du, L. Song, S. Yang, A. Nie, X. Zou and Y. Gong, Adv. Mater., 2020, 32, 2001167.
63. Y. Zhou, C. Li, Y. Zhang, L. Wang, X. Fan, L. Zou, Z. Cai, J. Jiang, S. Zhou, B. Zhang, H. Zhang, W. Li and Z. Chen, Adv. Funct. Mater., 2023, 33, 2304302.
64. Y. Feng, T. Zhang, J. Zhang, H. Fan, C. He and J. Song, Small, 2020, 16, 2002850.
65. C. Li, L. Zhu, Z. Wu, Q. Chen, R. Zheng, J. Huan, Y. Huang, X. Zhu and Y. Sun, Small, 2023, 19, 2303646.
66. Z. Li, X. Yan, D. He, W. Hu, S. Younan, Z. Ke, M. Patrick, X. Xiao, J. Huang, H. Wu, X. Pan and J. Gu, ACS Catal., 2022, 12, 7687–7695.
67. J. Liu, Z. Wang, J. Li, L. Cao, Z. Lu and D. Zhu, Small, 2020, 16, 1905738.
68. K. Aydin, S. Woo, V. K. Kanade, S. Choi, C. Ahn, B. Lim and T. Kim, Carbon Energy, 2023, 5, e340.
69. D. Zhang, F. Wang, W. Zhao, M. Cui, X. Fan, R. Liang, Q. Ou and S. Zhang, Adv. Sci., 2022, 9, 2202445.
70. Q. Liu, F. Nichols, A. Bhuller, K. Singewald, H.-L. Kuo, J.-Q. Lu, G. L. Millhauser, F. Bridges, Q. Ge and S. Chen, Appl. Catal., B, 2024, 342, 123399.

---