Constructing Double Heterojunctions on 1T/2H-MoS₂@Co₃S₄ Electrocatalysts for Regulating Li₂O₂ Formation in Lithium-Oxygen Batteries

Highlights

• 1T/2H-MoS₂@Co₃S₄ electrocatalysts were constructed by interfacial charge donation from Co to Mo atoms, resulting in formation of double heterojunctions including 1T-MoS₂@Co₃S₄ and 2H-MoS₂@Co₃S₄.

• Complementary effect from double heterojunctions not only triggered fast charge transport on Co–S–Mo couplings, but also enabled moderate e_g orbital occupancy to adsorb oxygen-containing intermediates for efficient oxygen electrocatalysis.

• Optimal adsorption energies for solution and surface dual reaction pathways were achieved, forming two kinds of discharge product morphologies during cycling to enhance performance of Li–O₂ batteries.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40820-025-01895-x.

Abstract

Co₃S₄ electrocatalysts with mixed valences of Co ions and excellent structural stability possess favorable oxygen evolution reaction (OER) activity, yet challenges remain in fabricating rechargeable lithium-oxygen batteries (LOBs) due to their poor OER performance, resulting from poor electrical conductivity and overly strong intermediate adsorption. In this work, fancy double heterojunctions on 1T/2H-MoS₂@Co₃S₄ (1T/2H-MCS) were constructed derived from the charge donation from Co to Mo ions, thus inducing the phase transformation of MoS₂ from 2H to 1T. The unique features of these double heterojunctions endow the 1T/2H-MCS with complementary catalysis during charging and discharging processes. It is worth noting that 1T-MoS₂@Co₃S₄ could provide fast Co–S–Mo electron transport channels to promote ORR/OER kinetics, and 2H-MoS₂@Co₃S₄ contributed to enabling moderate e_g orbital occupancy when adsorbed with oxygen-containing intermediates. On the basis, the Li₂O₂ nucleation route was changed to solution and surface dual pathways, improving reversible deposition and decomposition kinetics. As a result, 1T/2H-MCS cathodes exhibit an improved electrocatalytic performance compared with those of Co₃S₄ and MoS₂ cathodes. This innovative heterostructure design provides a reliable strategy to construct efficient transition metal sulfide catalysts by improving electrical conductivity and modulating adsorption toward oxygenated intermediates for LOBs.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40820-025-01895-x.

Introduction

Aprotic lithium-oxygen batteries (LOBs) are one of the most promising candidates to replace lithium-ion batteries (LIBs) for electric vehicles, robots, and large-scale grids, because of their ultrahigh theoretical energy density (3,500 W h kg⁻¹). Generally, the reaction mechanisms during cycling are based on Li⁺  + O₂ ↔ Li₂O₂ (E⁰ = 2.96 V vs. Li/Li⁺), which involves oxygen reduction reaction (ORR) for discharging and oxygen evolution reaction (OER) for charging [1, 2]. However, such a powerful system is greatly hindered for further applications by poor rate performance, limited cyclic life, and high overpotentials. These problems mainly result from their sluggish redox kinetics, triggering the accumulation of inactive discharge product (Li₂O₂) and even insulated side reaction product (LiOH and Li₂CO₃) on the cathode [3]. Considering that the cathodes should exposure the triple-phase contact interface regions (cathode/electrolyte/O₂) and enable discharge product storage, employing ideal cathode catalysts for fast electron transfer rate and reversible discharge product formation would greatly enhance the electrocatalytic performance and accelerate the commercial applications of LOBs.

Noble metals and their alloys exhibit high-efficiency catalytic properties in LOBs, while the high cost and scarcity on the Earth's crust seriously restrict their large-scale applications [4, 5]. Transition metal chalcogenides (TMCs) have been attracting extensive attention worldwide on account of their low cost, excellent chemically active electrocatalytic property, and diversity of structure [6]. Among them, Co₃S₄ possesses a typical spinel structure with a mixture valence state of Co²⁺ and Co³⁺, and this unique arrangement endows Co₃S₄ eye-catching recognition as the OER electrocatalyst, proved by density functional theory (DFT) calculations [7, 8]. Liu et al. [9] synthesized ultrathin Co₃S₄ nanosheets via a sonicated treatment, delivering high OER activities and reduced overpotentials in overall water splitting reactions. Wang et al. [10] present superior manner in OER performance, which can be realized by inducing sulfur vacancies in Co₃S₄. Furthermore, Yang et al. [11] demonstrate that the adsorption states of oxygen-containing intermediates could be tailored by Ni incorporation to largely influence OER activities. However, it is proposed that the monometallic sulfides generally fail to simultaneously exhibit the high electrocatalytic activities toward bifunctional electrocatalytic reactions (ORR and OER) [12]. In fact, when exposed to ambient environment or an oxygen-containing solution, Co₃S₄ tends to be extraordinarily reactive and unstable, which is prone to rapid oxidation due to the weak Co–S bonds and results in low electrical conductivity during ORR [13, 14]. Thus, it is necessary to construct heterostructures to offer conductive support for Co₃S₄ and further achieve bifunctional electrocatalytic abilities.

As another widely studied TMCs, MoS₂ catalysts have been widely applied as one of the key alternatives to noble metal catalysts in electrochemical fields [15, 16]. It is found that MoS₂ exhibits the layered S-Mo-S sandwich structure, connected by weak van der Waals forces along the c-axis. Besides, according to the stacking manner of S atoms, there are three different crystalline phases of MoS₂, including 1T (trigonal structure, octahedral coordination), 2H (hexagonal structure, trigonal prismatic coordination), and 3R (rhombohedral structure, trigonal prismatic coordination) [17]. Due to intrinsic properties, 3R-MoS₂ has been proven to be a promising material for optoelectronic and photonic devices [18, 19]. 2H-MoS₂ is dominant and more stable in nature, showing the characteristics of a semiconductor with a large bandgap between 1.29 and 1.90 eV. Researchers have calculated and proven that 2H-MoS₂ possesses ultra-low adsorption energy (E_ads) for reaction intermediates and hardly adsorbs the reaction intermediates, while the stack edge planes of 2H-MoS₂ can confine reaction species and restrict growth of the discharge product due to the ultrahigh E_ads. To increase electrical conductivity and E_ads toward oxygen-containing intermediates, a variety of strategies were carried out, including hybridization with carbon materials (carbon nanotubes [20], graphene oxide [21], and other carbon materials [22, 23]), doping with heteroatoms [24, 25], and constructing S vacancies [26, 27]. The bandgap of 1T-MoS₂ is close to zero with metallic property, exhibiting 10⁷ times higher electrical conductivity than that of 2H-MoS₂, which is pivotal for enhancing its catalytic performance [28, 29] Theoretical calculations confirm that the octahedral coordination favors the accommodation of these extra electrons in the d orbitals of Mo atoms, gifting 1T-MoS₂ with metallic property [30]. However, 1T-MoS₂ is considered to be thermodynamically unfavorable and hard to realize high-yield production, leading to spontaneous phase reversion to 2H-MoS₂ [31, 32]. It is reported that electron configuration modulation for 2H-MoS₂ could induce 1T-MoS₂ formation due to the electron injection to Mo 4d, and the existence of 2H-MoS₂ can contribute to stabilizing 1T-MoS₂ [33–35]. Therefore, heterostructures based on MoS₂ are believed to offer a new platform to explore this connection owing to the abundant charge donation on the heterointerfaces.

Herein, ZIF-67 was employed as template to synthesize 1T/2H-MoS₂@Co₃S₄ heterostructures (1T/2H-MCS) as bifunctional electrocatalysts for LOBs via one-pot hydrothermal method with two-step temperature-raising process. The original electron configuration of Mo 4d orbitals was changed via the electron injection by Co 3d, which triggered the phase transition from 2H to 1T to form double heterojunctions of 1T-MoS₂@Co₃S₄ and 2H-MoS₂@Co₃S₄ in 1T/2H-MCS. According to the theoretical calculations on Co₃S₄, 2H-MoS₂, 1T-MoS₂@Co₃S₄, and 2H-MoS₂@Co₃S₄, the complementary effects on 1T/2H-MCS were intensively investigated. On the one hand, owing to the increasing density of states (DOS) at the Fermi energy level, 1T-MoS₂@Co₃S₄ exhibits excellent electrical conductivity via the fast electron transport channels of Co–S–Mo bonds, while the low E_g orbital occupancy is unfavorable for adsorption of oxygen-containing intermediates, resulting from the large bandgap between d-band center (E_d) of Co and Fermi level. On the other hand, due to the appropriate d-p orbital hybridization of 2H-MoS₂@Co₃S₄ and LiO₂ with modulated E_d of Co, 2H-MoS₂@Co₃S₄ shows suitable E_ads, enabling the changed adsorption configuration of oxygen-containing intermediates to efficiently adsorb and desorb the discharge product. Nevertheless, the overall electrical conductivity of 2H-MoS₂@Co₃S₄ is poor due to the wide bandgap of 2H-MoS₂. Moreover, 1T/2H-MCS could completely maintain dodecahedral hollow structure from ZIF-67 and be decorated with MoS₂ nanosheets. This unique 3D architecture can effectively impede the agglomeration and restack of MoS₂, exposing more active sites and promoting mass diffusion for faster reaction kinetics. Together with these synergistic effects, 1T/2H-MCS cathodes exhibit significantly improved electrochemical activities for both OER and ORR, including a discharge specific capacity of 16,875 mAh g⁻¹ at 0.5 mA cm⁻² with a capacity retention of up to 96% and a steady cycle life of 375 cycles at 0.5 mA cm⁻² with a cutoff specific capacity of 1 mAh cm⁻², and these data demonstrate their practical application potential for next-generation high energy–density power sources.

Experimental Section

Synthesis of ZIF-67

The typical reaction process for the preparation of ZIF-67 template is as follows. 3 mmol cobalt nitrate hexahydrate and 12 mmol 2-methylimidazole were dissolved in 30  and 10 mL of methanol , respectively, which were then mixed and aged at room temperature for 24 h. After complete centrifugal washing with methanol, ZIF-67 product was dried at 80 °C overnight.

Synthesis of 1T/2H-MoS₂@Co₃S₄ (1T/2H-MCS)

50 mg of as-prepared ZIF-67 templates were added into 10 mL ethanol under sonication for 10 min. Consequently, 10 mL aqueous solution containing Na₂MoO₄⋅2H₂O (0.033 M) and thioacetamide (TAA, 0.1 M) was added slowly into the above solution under sonication for another 10 min. Afterward, the uniform mixture was placed into a 40-mL Teflon-lined stainless-steel autoclave, which was maintained at 120 °C for 4 h and then raised to 200 °C for 8 h. After being cooled down naturally to room temperature, final product was centrifuged and washed with ethanol and deionized water, followed by drying at 80 °C overnight.

Material Characterizations

X-ray diffraction (XRD, Rigaku D/Max-r B) was used to characterize the crystal phases. Field-emission scanning electron microscopy (FESEM, JSM-7610F) and transmission electron microscopy (TEM, JEOL JEM-F200) with the acceleration voltage of 200 kV were applied to display the microstructures. Electrical conductivity was measured at room temperature by a ST-2258A digital four-point probe test system. Before the measurements, sample powders were compressed into a wafer with a thickness of 0.2 mm and a diameter of 13 mm by an oil pressure machine under a pressure of 30 MPa. Thickness of the nanosheets was measured by using an atomic force microscope (AFM) on a Bruker DI MultiMode-8 system. X-ray photoelectron spectroscopy (XPS, ESCALAB 250) was conducted to investigate surface chemical states. Raman spectra (LabRAM HR Evolution) with a 532-nm laser were used to further detect the material structures. Electron paramagnetic resonance (EPR, EMXnano) tests were performed at 300 K to confirm the defects. N₂ adsorption–desorption measurements (ASAP 2020) were implemented to detect the surface areas and pore sizes. Inductively coupled plasma spectrometer (ICP-OES, Agilent 5110) was used to determine element contents. The gas evolution rates of O₂, H₂O, and CO₂ were tested by in situ DEMS (i-DEMS 100). X-ray absorption fine structure (XAFS) spectra of Co K-edge were collected in transmission mode on a commercial Laboratory-Based XAFS spectrometer (Table XAFS-500A, Specreation Instruments Co., Ltd.). An X-ray tube was used to generate X-ray, and the voltage and current were set to 20 kV and 20 mA. The R250-mm Rowland circle was used to provide monochromatized X-ray beam.

Electrochemical Measurements

LOBs were assembled using CR 2032 coin-type cells with holes in their positive top covers for electrochemical tests. Electrolyte was composed of 1 M lithium bis(trifluoromethanesulfonyl)imide dissolved in trimethylene glycol dimethyl ether (LiTFSI/TEGDME). Active material, Ketjen black, and polytetrafluoroethylene (PTFE) polymer binder were mixed in 5 mL isopropanol solvent with a mass ratio of 4:4:2 to obtain the slurry, which was coated on the carbon paper on a heating base at 80 °C to fabricate the cathode. Lithium foil and glass fiber membrane were served as anode and separator, respectively. LOBs were assembled in an argon-filled glove box (H₂O < 0.1 ppm, O₂ < 0.1 ppm) and placed in a chamber purged with O₂ to investigate galvanostatic discharge/charge performance on a multichannel battery system (LAND CT 2001A). Cyclic voltammetry (CV) testing at 0.15 mV s⁻¹ within the potential range of 2.35–4.50 V (vs Li/Li⁺) as well as electrochemical impedance spectroscopy (EIS) measurement with a limited frequency range of 10⁵ to 10^–2 Hz and a 10 mV oscillation amplitude was carried out on the electrochemical workstation (CHI760E).

In a pouch-type LOB, lithium foil (4 × 4 cm²), 1T/2H-MCS-coated carbon cloth (4 × 4 cm²), and 1 M LiTFSI/TEGDME-soaked glass fiber membrane (4.5 × 4.5 cm²) were, respectively, used as the anode, the cathode, and the separator, which were encapsulated with perforated aluminum plastic film and sealed by a heat press sealer. The mass loading of 1T/2H-MCS is 16 mg for each cathode.

Theoretical Calculations

First-principles computations were employed by the Vienna Ab initio Simulation Package (VASP) based on DFT. The exchange–correlation energy of electrons was described using generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) parametrization. The core valence interaction was performed using the projector augmented wave (PAW) method. The cutoff energy was 450 eV and the k-point mesh was set to be 1 × 1 × 1 for all slab calculations. The energy convergence criteria were set to 10^–5 eV. A vacuum layer of 15 Å along the z-axis was applied to all slabs to avoid the periodic structural interactions. The maximum Hellmann–Feynman stress on each atom was set to be less than 0.02 eV Å⁻¹. All atoms were fully relaxed for the heterojunction structure optimizations, while the lower four layers of atoms were fixed for the structure optimization of all the catalytic intermediates.

Results and Discussion

Theoretical Prediction on Double Complementary Effect in 1T/2H-MoS₂@Co₃S₄

DFT calculations were conducted to investigate the phase transition mechanisms of MoS₂ after combined with Co₃S₄, and Fig. S1 displays the optimal structural models of the Co₃S₄, 2H-MoS₂, 2H-MoS₂@Co₃S₄, and 1T-MoS₂@Co₃S₄ heterostructures. In our previous research, DFT calculations of 2H-MoS₂ have been carried out by the same way as in this work [36]. By Bader charge analysis, it is revealed that the electron transfer occurred from Co₃S₄ into 2H-MoS₂ with an amount of 0.638 e, inducing phase transition from 2H-MoS₂ to 1T-MoS₂ (Figs. 1a and S2). Charge density difference and electron localization function (ELF) further prove the electron transfer from Co₃S₄ to 2H-MoS₂ on heterogeneous interfaces, as shown in Fig. 1b. Therefore, the electron configuration of Mo 4d orbitals was rearranged. Based on crystal-field theory, Mo 4d orbitals of trigonal prismatically coordinated 2H-MoS₂ cleavage into three energy levels: d_z2 (a₁′), d_x2-y2/d_xy (e′), and d_xz/d_yz (e′′) orbitals. Two Mo d electrons could completely be filled to the lower-lying a₁′ orbital, endowing the 2H-MoS₂ semiconducting characteristic with high structural stability. In contrast, Mo 4d orbitals of octahedrally coordinated 1T-MoS₂ cleavage into d_xy/d_xz/d_yz (t_2g) and d_z2/d_x2-y2 (e_g) orbitals. Because of the incomplete occupation of the lower-lying t_2g orbitals, 1T-MoS₂ shows metallic electronic property but poor structural stability [37].

Fig. 1.

a Bader charge transfer, b ELF, and charge density difference of 2H-MoS₂@Co₃S₄, in which charge accumulation and depletion are displayed by yellow and cyan, respectively. c Schematic diagram of proposed electron transfer processes between Co₃S₄ and 2H-MoS₂. d TDOS and e PDOS of different samples. f, g d-band and p-band centers of Co 3d, Mo 4d, and S 2p. h Schematic diagram of double heterojunction effects on 1T/2H-MCS for LOBs

To explain the interaction between Mo ions and the adjacent Co ions after combination, the partial electron transfer process from Co to Mo ions through the S ligand is illustrated in terms of their electronic configuration, as given in Figs. S3 and S4. As mentioned above, the a₁′ orbital of Mo⁴⁺ with 4d² trigonal prismatically coordinated geometry is fully occupied. In this condition, the major electronic interaction between the bridging S²⁻ and Mo⁴⁺ is e⁻-e⁻ repulsion. Similarly, trivalent Co ions present the valence electron configuration of 3d⁶ and exhibit low-spin octahedral geometry with fully occupied t_2g orbitals. After formation of heterostructures, the d electrons of Co 3d move to Mo 4d. Due to the relative higher electronegativity of Mo ions, it thus can be claimed that the e⁻–e⁻ repulsion of Co–S bonds aided in intensifying the π-donation interaction between Mo⁴⁺ and S²⁻, which trigger partial electron transfer from Co to Mo ions, accompanying the changes of coordination with surrounding S²⁻ [38]. As depicted in Fig. 1c, when 2H-MoS₂ was combined with Co₃S₄, the extra electrons donated by Co₃S₄ firstly occupied the e′ orbitals of 2H-MoS₂, destabilizing the structure and inducing the Mo 4d orbital reorganization to generate MoS₂ transformation from 2H to 1T phase. As a result, the intrinsic electrical conductivity of MoS₂ was largely improved, which could serve as conductive phase for Co₃S₄. Total density of states (TDOS) around the Fermi level in Fig. 1d exhibits stronger metallic properties in 1T-MoS₂@Co₃S₄. Particularly, the orbitals of Co 3d, S 2p, and Mo 4d in Fig. 1e show substantial overlaps, implying strong electron interaction among them to accelerate electron transfer in the Co–S–Mo channels. Furthermore, after the combination of MoS₂ and Co₃S₄, the electron transferred from Co to Mo ions induced the downshift of the E_d of Co and the opposite trend of Mo (Fig. 1f), respectively, narrowing the energy differences between Co 3d and S 2p band centers and Mo 4d and S 2p band centers and thereby leading to the strong hybridization among Co, S, and Mo (Fig. 1g). The strong hybridization could improve the covalency Co–S–Mo bonds, which can directly facilitate the electron transport between Co ions and oxygen adsorbates via Co–S–Mo channels and activate the Co sites during the ORR process. Besides, the E_ads of LiO₂ on Co₃S₄ surfaces is extraordinarily strong, and it led to the high charge overpotentials, due to the fact that the decomposition of the discharge product needs to cross a higher energy barrier. However, when combined with 1T-MoS₂, the E_ads of 1T-MoS₂@Co₃S₄ could be reduced by modulated E_d of Co 3d, which will be discussed in the following parts.

Based on the above analysis, Fig. 1h systematically demonstrates the synergistic effects of double complementary heterojunctions derived from charge donation from Co to Mo ions in 1T/2H-MCS. Specifically, according to theoretical calculations, 2H-MoS₂@Co₃S₄ possesses suitable E_ads to efficiently adsorb and desorb the discharge product, but the overall electrical conductivity is poor due to the wide bandgap of 2H-MoS₂. On the contrary, owing to the increasing DOS at the Fermi energy level, 1T-MoS₂@Co₃S₄ exhibits excellent electrical conductivity via fast electron transport channels of Co–S–Mo bonds, while low antibonding orbital occupancy is unfavorable for adsorption of oxygen-containing intermediates, resulting from the large bandgap between E_d of Co and Fermi level. Therefore, 1T/2H-MCS containing complementary double heterojunctions is expected to enhance the catalytic activities of Co₃S₄.

Morphological and Structural Characterization

In a typical synthesis, ZIF-67 was applied as a template to obtain 1T/2H-MCS through a facile hydrothermal process (Fig. 2a), which could be divided into two steps. Under lower temperature, ZIF-67 was transformed into Co₃S₄ via a ligand exchange reaction between S₂⁻ and 2-methylimidazole, still maintaining rhombic dodecahedral shape. After higher temperature, disordered MoS₂ nanosheets grew on the surfaces of hollow Co₃S₄ polyhedrons to form 1T/2H-MCS. The digital photograph illustration of the synthesis path is given in Fig. S5.

Fig. 2.

a Schematic illustration of synthetic process for 1T/2H-MCS with the proposed phase transformation mechanism. b XRD pattern, c high-resolution Co 2p XPS spectra, e Co K-edge XANES spectra, f Fourier transform of Co K-edge EXAFS spectra, and g wavelet transform analysis of different samples. d 1T and 2H phase ratios, h FESEM image, i HAADF image, and j HRTEM image with enlarged results from corresponding selected areas of 1T/2H-MCS

To explore the crystalline structure and compositions of different samples, XRD, Raman, XPS, and EPR measurements were conducted. XRD analysis in Fig. 2b reveals that all characteristic peaks correspond to those of cubic Co₃S₄ (JCPDS 47-1738) and 2H-MoS₂ (JCPDS 75-1539), proving that 1T/2H-MCS were successfully synthesized. Figures S6a, S7a, and S8a exhibit XRD patterns of the precursor, Co₃S₄, and 2H-MoS₂, respectively. Raman data in Fig. S9 reveal that the peaks at 185 (A_g), 298 (F_2g), and 348 cm⁻¹ (F_2g) are related to Co₃S₄, while the peaks at 375 (E1 2g) and 399 cm⁻¹ (A_g) are associated with 2H-MoS₂ [39–41]. Four distinctive peaks at 140 (J₁), 193 (J₂), 278 (E_1g), and 333 cm⁻¹ (J₃) can be also observed, originating from the phonon modes of 1T-MoS₂, and this demonstrates successful MoS₂ phase conversion from 2H to 1T after combination [42].

XPS survey spectra in Fig. S10 overview the elements existing on different samples. To provide a direct comparison, the Co 2p XPS spectra of Co₃S₄ and 1T/2H-MCS are presented in Fig. 2e. Apparently, Co 2p XPS peak consists of two spin–orbit doublets, including Co 2p_3/2 and Co 2p_1/2 along with two satellites. Co³⁺ and Co²⁺ XPS peaks of Co₃S₄ position at 778.2/793.1 and 781.1/796.8 eV, whereas those of 1T/2H-MCS locate at 778.8/793.9 and 781.7/797.4 eV, respectively, indicating a binding energy red shift of 0.6 eV in 1T/2H-MCS relative to Co₃S₄. This XPS peak shift was attributed to changes in the oxidation state of the Co₃S₄ owing to MoS₂ presence. Detailed comparisons of Co²⁺/Co³⁺ ratios in 1T/2H-MCS and Co₃S₄ are displayed in Fig. S11, showing the increased content of Co³⁺ from 48% to 65%. Co 2p XPS spectrum of 1T/2H-MCS displays higher oxidation states than that of Co₃S₄ owing to electron transfer at the interfaces of Co₃S₄ and 2H-MoS₂, which is also confirmed by Bader charge results [43]. Co 2p XPS spectrum of 1T/2H-MCS displays higher oxidation states than that of Co₃S₄ owing to electron transfer at the interfaces of Co₃S₄ and 2H-MoS₂ [43]. It is believed that high Mo electronegativity altered the Co electronic structure (σ*-orbital occupancy) by decreasing the Co electron density. Co σ*-orbital (e_g) occupancy is close to unity, which can boost the intrinsic ORR and OER activities toward oxygen species by promoted covalence of Co–O bonds [44, 45]. Figure S12 shows the corresponding high-resolution Mo 3d XPS spectra of 2H-MoS₂ and 1T/2H-MCS. Compared to Mo 3d_3/2 (232.3 eV) and Mo 3d_5/2 (229.2 eV) signals of pristine 2H phase, all the corresponding XPS peaks of 1T/2H-MCS blue-shift to lower binding energy at 231.9 and 228.8 eV. After fitting, the asymmetric signals can be divided into two peaks, and higher binding energy (labeled with green) and lower binding energy (labeled with pink), respectively, correspond to 2H and 1T MoS₂ phases, which is well consistent with Raman results [46, 47]. In terms of the deconvolution result of Mo 3d XPS spectra, the ratio of 1T phase is over 65% in the MoS₂ (Fig. 2d). It can be concluded the increased content of Co³⁺ could induce extra electron, occupying the e′ orbitals of 2H-MoS₂ to realize Mo 4d orbital reorganization and MoS₂ phase transformation from 2H to 1T. Accordingly, the high-resolution S 2p XPS spectra of 1T/2H-MCS in Fig. S13 are fitted into four peaks, which locate at 161.6/162.7 eV for 1T-MoS₂, 162.1/163.2 eV for 2H-MoS₂, and 163.7/164.9 eV for Co₃S₄. It is obvious that different shifts occurred for Co₃S₄ and 2H-MoS₂, demonstrating that the construction of the heterostructures enabled electron transfer from Co₃S₄ to 2H-MoS₂ through S-bridges. As shown in Fig. S14, EPR signal of 1T/2H-MCS (g = 2.003) corresponds to the unpaired Mo-S electrons on coordinatively unsaturated sulfur vacancy sites, which is stronger than those of Co₃S₄ and 2H-MoS₂ [48, 49]. Compared with 2H-MoS₂, the restriction growth of MoS₂ by Co₃S₄ triggered more unpaired electrons on (002) lattice planes, exposing more unsaturated sulfur edges on the surfaces of 1T/2H-MCS, and this is in accordance with TEM results.

Brunauer–Emmett–Teller (BET) measurements based on N₂ adsorption–desorption method were performed to investigate the specific surface areas and pore volumes. As shown in Fig. S15, the isotherms of 2H-MoS₂, Co₃S₄, and 1T/2H-MCS conform to the H3-type hysteresis loop, implying the presence of mesoporous structures in different samples. Typically, the isotherms of 1T/2H-MCS with inconspicuous adsorption platform demonstrate that this mesoporous structure is irregular. Specific surface area of 1T/2H-MCS (47.54 m² g⁻¹) is larger than those of Co₃S₄ (22.74 m² g⁻¹) and 2H-MoS₂ (11.34 m² g⁻¹). Compared with 2H-MoS₂, it is supposed that S²⁻ can quickly react with Co²⁺ to form hollow Co₃S₄ with high surface area at low temperature. The growth of small MoS₂ nanoflakes on the surfaces of Co₃S₄ at higher temperature could further enlarge the specific surface area. Moreover, the hollow Co₃S₄ structure endowed 1T/2H-MCS with a larger pore volume (0.21 cm³ g⁻¹) than those of 2H-MoS₂ (0.071 cm³ g⁻¹) and Co₃S₄ (0.153 cm³ g⁻¹). It is believed that large specific surface area is beneficial to expose more active sites, promoting the formation and decomposition of discharge product. Moreover, more space could be enabled for fast mass transfer and enough discharge product storage. More importantly, the abundant mesopores in 1T/2H-MCS could contribute to effective electrolyte penetration, enlarging the electrolyte–catalyst interface area to accelerate oxygen catalysis reactions [50].

To further detect electronic structure and coordination information of Co species in 1T/2H-MCS, X-ray absorption spectroscopy (XAS) analysis was conducted. X-ray absorption near-edge spectra (XANES) of Co K-edge in Fig. 2e show that the Co K-edge of 1T/2H-MCS shifts to a higher energy position than those of Co foil and Co₃S₄, standing for an increasing Co chemical state in 1T/2H-MCS. The higher Co valence state in 1T/2H-MCS is consistent with the theoretical prediction and XPS results, suggesting that electron transferred from Co to Mo sites when combined with 2H-MoS₂ due to their electronegativity differences. Fourier transform (FT) k³-weighted EXAFS spectra of 1T/2H-MCS in Fig. 2f exhibit two apparent peaks located at 1.63 and 2.52 Å, which were ascribed to the Co–S and Co–Co contribution [51]. EXAFS spectra of Co K-edge were best fitted by including Co–S and Co–Co scattering path in fit model, as depicted in Table S1. Comparing 1T/2H-MCS with Co₃S₄, the first coordinative path of 1T/2H-MCS was expanded, which can be attributed to the asymmetric construction of Co–S–Mo after combining with 2H-MoS₂. This coordination path can be better visualized in Fig. 2g by the maxima in the wavelet transform (WT) of the EXAFS, which is quite different compared to those of Co foil and Co₃S₄.

Morphology and structure information of these samples were captured by FESEM. As depicted in Fig. S6b, ZIF-67 exhibits a rhombic dodecahedral structure, and its surfaces are smooth without apparent wrinkles. Figure 2h presents the FESEM image of 1T/2H-MCS, which retained the typical dodecahedral rhombic shape of ZIF-67 and is decorated with random nanosheets with an average particle size of about 500 nm. For comparison, Figs. S7b and S8b show the FESEM images of Co₃S₄ and 2H-MoS₂ samples, respectively, which display polyhedrons and large nanosheets with severe agglomeration. It is thus demonstrated that the ZIF-67 template can obviously restrict the layer-oriented growth of MoS₂. In addition, AFM was used to detect the flake thickness of 1T/2H-MCS. The result in Fig. S16a reveals that the MoS₂ nanoflakes formed by 6–7 monolayers, which is consisted with the following results of HRTEM images, and its corresponding 3D topography is offered in Fig. S16b. As shown in Figs. 2i and S17a, the high-angle annular dark-field (HAADF) and HRTEM images confirm that 1T/2H-MCS displays hollow structure containing Co₃S₄ nano-polyhedrons and outer MoS₂ nanosheets. Selected area electron diffraction (SAED) pattern agrees well with the crystal plane reflection of Co₃S₄ and 2H-MoS₂, verifying the successful synthesis of MoS₂@Co₃S₄ heterostructures, as pictured in Fig. S18. From higher-magnification observation, the lattice fringe spacing of 0.284 nm corresponds to Co₃S₄ (311) planes, and there are some disordered regions generated by the lattice mismatch between Co₃S₄ and MoS₂ (Fig. S17b). Interestingly, after formation of MoS₂ and Co₃S₄ heterostructures, lattice structures of 1T-MoS₂ in blue area and 2H-MoS₂ in green area are observed simultaneously in Fig. 2j, further confirming the successful synthesis of 1T/2H-MCS. In Fig. S19, energy-dispersive X-ray (EDX) elemental mappings of 1T/2H-MCS manifest that Mo, Co, and S elements are evenly distributed throughout the entire architecture. Similarly, Figs. S20 and S21 display uniformly element distribution on Co₃S₄ and 2H-MoS₂, respectively. Moreover, the element content results from EDX in Fig. S22 show that the molar ratio of Co₃S₄/MoS₂ in 1T/2H-MCS is 53.8/46.2, which is close to the inductively coupled plasma optical emission spectroscopy (ICP-OES) result (52.1/47.9), as demonstrated in Table S2.

Electrocatalytic Performance

Electrochemical tests were carried out to obtain insights into the ORR/OER catalytic activities of the prepared catalysts in LOBs. Figure 3a shows CV profiles of different cathodes within the voltage range of 2.35–4.50 V. It is found that 1T/2H-MCS cathode exhibits a higher onset potential for ORR (2.88 V) and a lower onset potential for OER (3.18 and 3.99 V), and its integral area is larger than those of other cathodes, implying better reaction kinetics and enhanced specific capacities [52]. During the cathodic scan, two small redox peaks during the ORR/OER processes demonstrate the stepwise formation/decomposition of intermediate LiO₂, indicating the fast ORR/OER kinetics of 1T/2H-MCS cathode. Figure 3b displays the initial galvanostatic discharge/charge performance for different cathodes at 100 mA g⁻¹. It is traced that 1T/2H-MCS cathode shows the highest discharge/charge specific capacities of 18,721/18,313 mAh g⁻¹, whereas those of the Co₃S₄, 2H-MoS₂, and KB cathodes are 12,823/10,775, 8346/7744, and 5,195/3,394 mAh g⁻¹, respectively. Figure S23 shows the first discharge/recharge differential plots of 1T/2H-MCS cathode. The obvious peak corresponds to ORR voltage plateau at first discharging, while the two main peaks during OER process mean that charging reactions went through with two plateaus. It is apparent that 1T/2H-MCS cathode presents superior discharge/charge capacities with highest Coulombic efficiency and lowest overpotentials, agreeing well with CV results.

Fig. 3.

a CV curves, b initial discharge/charge curves at 100 mA g⁻¹, c rate performance, as well as cycling duration e at 1000 mA g⁻¹ under a specific capacity limit of 600 mAh g⁻¹ with d selected charge–discharge profiles and f at 200 mA g⁻¹ under a specific capacity limit of 1000 mAh g⁻¹ of different cathodes. g Cycling stability comparison of 1T/2H-MCS cathode with representative and state-of-the-art cathodes containing Mo-, Co-, and noble metal-based compounds

In Fig. 3c, the discharge/charge polarization of 1T/2H-MCS cathode gradually increased, and it still delivers impressive discharge specific capacities of 17,452, 12,804, 9,852, and 5,203 mAh g⁻¹ at current densities of 200, 500, 800, and 1,000 mA g⁻¹, respectively, with corresponding Coulombic efficiencies of 97.2%, 91.3%, 86.7%, and 68.9%. In addition, the rate performance of different cathodes was also evaluated by terminal discharge/charge voltage fluctuations at various current densities in Fig. S24. As current density increased (100, 200, 400, 800, and 1000 mA g⁻¹), the discharge and charge terminal voltages of the 1T/2H-MCS cathode only show slight variations of 0.20 and 0.24 V, respectively. After the current density was set to a primitive value, the discharge/charge terminal voltages exhibit highly reversible recovery, further manifesting the excellent rate reliability of 1T/2H-MCS cathode. The resistivity value of 1T/2H-MCS was measured as 82 Ω cm by using four-point probe techniques, which is much lower than that of 2H-MoS₂ (7.68 × 10⁴ Ω cm) and Co₃S₄ (1.86 × 10³ Ω cm) [53, 54]. It is believed that by the formation of 1T-MoS₂ with the fast charge transfer channels of Co–S–Mo at the interfaces, the electronic conductivity of the 1T/2H-MCS cathode could dramatically increase, which enabled fast charge transfer and electrochemical performance improvement at high current densities.

Long-term cycling performance of different cathodes in LOBs was investigated. Figure S25 illustrates that 1T/2H-MCS cathode can stably work for 344 cycles, whereas the 2H-MoS₂, Co₃S₄, and KB cathodes gradually decayed after 176, 149, and 60 cycles with a cutoff specific capacity of 600 mAh g⁻¹ at 500 mA g⁻¹. It can also be seen in Fig. 3d–f that 1T/2H-MCS cathodes display an impressive cycle duration of 535 cycles at a high current density of 1000 mA g⁻¹ and an ultralong cycle life of 2420 h with a fixed specific capacity of 1,000 mAh g⁻¹. Figure 3g and Table S3 show battery performance comparison of the representative and state-of-the-art cathodes containing Mo-, Co-, and noble metal-based compounds with the result in this work. It is worth noting that the cycling stability of 1T/2H-MCS cathode exhibits comparable performance and even outperforms some of noble metal-based counterparts. These improved electrocatalytic activities are attributed to the phase transformation of MoS₂ from 2H to 1T by electron transfer, delivering a higher electrical conductivity, while the 2H phase can act as a stabilizer to keep the 1T phase structure stable during cycling. Meanwhile, the Co ions in Co₃S₄ with modulated e_g orbital occupancy after the electron transfer could promote the ORR/OER kinetics. In addition, after combining with 2H-MoS₂, the E_d of Co slightly moved away from the Fermi level, which optimized the adsorption ability for the discharge product and further improved the battery performance.

Practical Application in Pouch-Type Batteries

It is demonstrated in Fig. 4a that the flexible pouch-type LOB was constructed. As shown in Fig. 4b, it can obtain discharge/charge specific capacities of 16,875/16252 mAh g⁻¹ and a high capacity retention of 96%, indicating excellent ORR/OER kinetics of 1T/2H-MCS cathode. Impressively, it could be continuously operated for 375 cycles with a limited specific capacity of 1 mAh cm⁻² at 0.5 mA cm⁻², revealing a considerable cycling stability, as offered in Fig. 4c. To clearly demonstrate battery electrochemical stability, the open-circuit voltages of the LOBs with 1T/2H-MCS cathodes were tested under different bending angles (Fig. 4d), displaying almost no change even battery corner was cut and was under 180° bent at the same time. It is also demonstrated that two pouch-type LOBs could be connected in series with the open-circuit voltage of 5.84 V (Fig. S26). As illustrated in Fig. 4e, the smartphone can be successfully charged using the above batteries, which can also charge a smartwatch with its power increasing from 73% to 84% in 21 min (Fig. 4f). It is suggested that high energy density and durability of the pouch-type LOBs with 1T/2H-MCS cathodes are highly beneficial for supporting next-generation electronic devices.

Fig. 4.

a Schematic illustration of the configuration, b initial discharge/charge performance, c cycling duration at 600 mA g⁻¹ under a specific capacity limit of 1800 mAh g⁻¹, d open-circuit voltages of different deformation states, and e, f practical applications of pouch-type LOBs

Electrochemical Analysis

To investigate the morphology and structural evolution of discharge product on the different cathodes, ex situ XRD, ex situ FESEM, ex situ XPS, and in situ differential electrochemical mass spectrometry (DEMS) analysis were applied to characterize the reaction processes at different states during discharge/charge processes. Ex situ XRD results show that the main discharge product on 1T/2H-MCS cathode is Li₂O₂ with poor crystallinity (Fig. S27), which disappeared after recharging. Even after 100 cycles, Li₂O₂ was nearly decomposed to expose catalyst surfaces again, demonstrating negligible accumulation of the discharge product and efficient catalysis activities of 1T/2H-MCS. Similar conclusion can also be drawn by Raman spectra in Fig. S28, and it is demonstrated that the side reactions were significantly inhibited, and the Li₂O₂ was fully decomposed [55]. Interestingly, the peaks of 1T-MoS₂ were detected at different stages during cycling, indicating improved stability. According to previous studied, the existence of 2H-MoS₂ can contribute to stabilizing 1T-MoS₂, avoiding restacking as well as transformation into the 2H phase [35]. Furthermore, the morphology evolution of the 1T/2H-MCS, Co₃S₄, and 2H-MoS₂ cathodes at different stages during 1st discharging/charging was observed from ex situ FESEM images. It is apparent that the dense film-like discharge product was adhered to the surfaces of Co₃S₄ cathodes, as shown in Fig. 5a, b, leading to cathode passivation with limited specific capacities [56]. However, since the E_ads of 2H-MoS₂ for reaction intermediates is weak, the discharge product tends to form the rod-like Li₂O₂ through a solution-mediated pathway. It is noted that the poor catalytic activities of 2H-MoS₂ as well as limited contact between it and the rod-like Li₂O₂ result in higher overpotentials and unsatisfactory cycle performance due to incomplete decomposition of discharge product, as can be seen in Fig. S29a, b. Interestingly, the film- and rod-like Li₂O₂ generated simultaneously on the surfaces of 1T/2H-MCS cathode, which were caused by moderate adsorption energy of active sites for reaction intermediates. Actually, the discharge product on 1T/2H-MCS cathode coexisted in two forms, allowing for more accommodation of discharge product and a significant increase in discharge specific capacity. Figure S30 displays the morphology of the discharge product for 1T/2H-MCS cathode after discharging to 1,000 mAh g⁻¹. It is clear that the loose film-like and fine rod-like discharge product first formed, and it then grew into a uniform film and thick rod morphology with discharge capacity increase, which revealed the co-growth mechanism of the film- and rod-like Li₂O₂ on 1T/2H-MCS cathode. As shown in Fig. S31a, after 100th discharging, the film- and rod-like Li₂O₂ still could be observed on the surfaces of 1T/2H-MCS cathode, which is similar with that at 1st discharging, indicating the stability of the solution and surface dual reaction pathways on 1T/2H-MCS cathodes. Furthermore, the 1T/2H-MCS cathode could still maintain the original morphology after 100th recharging (Fig. S31b). In contrast, as shown in Figs. 5c, f and S29c, a large number of residues still remained on the surfaces of Co₃S₄ and 2H-MoS₂ cathodes. They not only covered the active sites on the cathode surfaces, but also blocked the Li⁺/O₂ transfer channels, thus ultimately leading to battery performance fails [57].

Fig. 5.

FESEM images of a–c Co₃S₄ and d–f 1T/2H-MCS cathodes at different states. g Initial discharge/charge curves at 200 mA g⁻¹ under a cutoff specific capacity of 1000 mAh g⁻¹, h high-resolution Li 1s XPS spectra at different states, and i in situ DEMS profiles with the charge curve at 200 mA g⁻¹ of 1T/2H-MCS cathodes

EIS plots of different cathodes were used to identify the intrinsic kinetics characteristic at different stages. As shown in Fig. S32, 1T/2H-MCS cathode presents smaller semicircles, and this implies a superior electrical conductivity for the phase transformation of MoS₂ from 2H to 1T, which is well consistent with TDOS results in Fig. 1d. After discharging, it is clear that charge transfer resistance (R_ct) increased from 21.9 to 85 Ω, which is still lower than those of Co₃S₄ and 2H-MoS₂ cathodes due to the unique morphology evolution of the discharge product on cathode surfaces, promoting the charge transfer on 1T/2H-MCS cathodes. In addition, after 1st and 100th recharging, their R_ct shows insignificant changes compared to that of the fresh cathode. In contrast, a remarkable R_ct increment of Co₃S₄ and 2H-MoS₂ cathodes can be detected after 1st discharging and charging, indicating that the dense film- and rod-like Li₂O₂/catalyst interfaces increased the charge transfer resistance and lead to obvious reaction polarization [58]. It is thus concluded that the 1T/2H-MCS cathodes show excellent electrical conductivity and moderate adsorption ability toward discharge product during cycling.

To further explore the chemical conditions of discharge product and reaction mechanisms in discharge/charge processes, ex situ high-resolution Li 1s XPS spectra of 1T/2H-MCS cathodes with selected pivotal states (Fig. 5g, h) at 200 mA g⁻¹ were studied. It is worth noting that the primarily discharge product on 1T/2H-MCS cathode is Li_2-xO₂ (a low crystalline mixture of Li₂O₂ and LiO₂) with the peak at 56.1 eV, and it transformed to Li₂O₂ during subsequent discharging, which is similar on Co₃S₄ and 2H-MoS₂ cathodes [59]. However, signals of Li₂CO₃ appeared at 55.6 eV for Co₃S₄ and 2H-MoS₂ cathodes. This indicates that side reactions occurred during the discharge process, and Li₂CO₃ remained when recharged to a limited specific capacity of 500 mAh g⁻¹, which were accumulated during cycling and degraded battery performance, as evidenced in Figs. S33 and S34. Subsequently, in situ DEMS was conducted to determine the evolution of gas species in discharge/charge reactions. To calculate the electron-to-O₂ (e⁻/O₂) ratio, the intrinsic micro-electrochemical processes were evaluated using the following equations, as determined by the amount of e⁻ passed during charging and the integrating curves of the O₂ consumption rate:

$$\text{Q}=\text{I}\times \text{t}$$ 1

$$\nu \left({\text{e}}^{-}\right)=\text{Q}\times {\text{e}}^{-}$$ 2

$$\nu \left({\text{O}}_2\right)=\text{GasEvol}.\text{Rate}/60\times {\text{N}}_{\text{A}}$$ 3

In situ DEMS results of different cathodes were obtained at 500 mA g⁻¹ during charging, as provided in Figs. 5i, S35b, and S36b. It is clear that the e⁻/O₂ ratio for 1T/2H-MCS was found to be close to 1.92, indicating the simultaneous decomposition of Li₂O₂ (2e⁻/O₂) and LiO₂ (e⁻/O₂). In contrast, the values of Co₃S₄ and 2H-MoS₂ are, respectively, 2.88 and 2.58, much higher than that for 1T/2H-MCS. This indicates incomplete decomposition of the discharge product and severe parasitic reactions, which is consistent with ex situ FESEM and ex situ XPS results [60]. Typically, O₂ evolution plots for each cathode present typical “M” shape during charging, demonstrating stepwise decomposition of Li₂O₂, and an increased maximum O₂ gas precipitation was presented at lower potential for 1T/2H-MCS cathode, which could be attributed to enhanced OER catalytic abilities [61]. Furthermore, it is apparent that an amount of CO₂ gas evolved at the end of charging state for both Co₃S₄ and 2H-MoS₂ cathodes, providing strong evidence that severe parasitic reactions occurred in recharge process to pose poor reversible specific capacity. During discharging, all curves are similar, and e⁻/O₂ ratios are all close to the 2 for desired O₂ reduction to Li₂O₂, as listed in Figs. S35a, S36a, and S37. These results demonstrate that the 1T/2H-MCS cathode could deliver moderate adsorption ability and significantly inhibit side reactions, thus exhibiting superior electrocatalytic activities for ORR/OER processes in LOBs.

To further elucidate the effect of double heterojunction effects in 1T/2H-MCS, samples of MoS₂ mixed Co₃S₄ in the ratio of 1:1 (MoS₂ Mixed Co₃S₄) were prepared. Since the DOS distributions of Co 3d, S 2p, and Mo 4d are mismatched, and it is hard to form chemical bonds between Co₃S₄ and MoS₂ through simple mixing. In Fig. S38a, the ratio of Co²⁺/Co³⁺ in MoS₂ Mixed Co₃S₄ is 46/54, which is similar to that in pure Co₃S₄, proving the limited charge donation from Co to Mo ions, and only 2H-MoS₂ is present in MoS₂ Mixed Co₃S₄, as shown in Fig. S38b. Consequently, MoS₂ Mixed Co₃S₄ cathode shows unsatisfactory initial discharge/charge capacities of 11,037/8,400 mAh g⁻¹ with poor Coulombic efficiency, as depicted in Fig. S39a. As for rate performance, discharge/charge terminal voltages decreased/increased obviously as the variation of current density (Fig. S39b). Furthermore, the EIS plots of MoS₂ Mixed Co₃S₄ cathodes were collected at different stages in Fig. S40. The MoS₂ Mixed Co₃S₄ cathode still displays a large R_ct of 169 Ω after recharging, indicating sluggish reaction kinetics during OER process. The above results confirm that the double heterojunction effects in 1T/2H-MCS play a critical role in enhancing electrochemical performance.

Proposed Mechanisms by DFT Calculations

To further clarify the catalytic mechanisms, DFT calculations were carried out to simulate the formation/decomposition of the discharge product on the cathode catalysts. Figure S41 shows different optimized structures after adsorbing different oxygen species (LiO₂, Li₂O₄, Li₃O₄, and Li₄O₄), and corresponding charge density difference plots and Bader charge transfer of different oxygen species on the surfaces of Co₃S₄, 2H-MoS₂, 2H-MoS₂@Co₃S₄, and 1T-MoS₂@Co₃S₄ are shown in Figs. 6a and S42-S45. In those models, the yellow and blue areas denote the gained electrons and lost electrons, respectively. As for the adsorption models of 2H-MoS₂, 2H-MoS₂@Co₃S₄, and 1T-MoS₂@Co₃S₄, the charge could transfer from oxygen-containing intermediates to catalysts, with more charge transferred to the active sites on 2H-MoS₂@Co₃S₄, efficiently promoting desorption of LiO₂ and thus greatly enhancing catalyst reaction kinetics [62]. While for the Co₃S₄, the charge transferred to oxygen-containing intermediates, which is unfavorable for the catalytic conversion and lead to increased overpotentials [63].

Fig. 6.

a Charge density differences with Bader charge transfer, b d-band and p-band centers of Co 3d and S 2p of different catalysts. c Schematic illustration of d-p orbital hybridization for 1T/2H-MCS. d Adsorption energy and the relationship between adsorption energy and overpotentials, e–g Calculated energy diagrams, and h relationship between adsorption energies and overpotentials of different catalysts

It is noted that covalency of metal–oxygen bonds serves as another essential descriptor for the electrocatalytic activities during the reaction processes, as it is closely linked to the interfacial electron transfer kinetics involved in oxygen redox reactions [64]. Typically, a modest increased covalency of metal–oxygen bonds tends to mitigate the electron transfer energy barrier. It can be found in Fig. 6b that PDOS shows a middle overlap degree between Co 3d- and O 2p-orbitals near the Fermi level in 2H-MoS₂@Co₃S₄, compared to those in Co₃S₄ and 1T-MoS₂@Co₃S₄, which demonstrates the moderate adsorption of LiO₂ on the 2H-MoS₂@Co₃S₄. Simultaneously, an increased energy gap between the Co 3d and O 2p band centers can be observed in Fig. 6b. The expanding of the energy difference between the Co 3d and O 2p band centers also manifests the weak hybridization between the Co 3d and O 2p, undermining the covalency of Co–O bonds. According to Sabatier principle [65], the strong covalency of Co–O bonds could result in the deactivation of active sites to increase the overpotentials. On the contrary, weak covalency of Co–O bonds could lead to unrestricted growth of Li₂O₂ and thus limit the contact between Li₂O₂ and cathode surfaces, deteriorating LOB performance. Besides, the overloaded adsorbed discharge product was prone to fall off the cathode, inducing discharge product loss at deep discharge state. Thus, it is concluded that 2H-MoS₂@Co₃S₄ with moderate d-p orbital hybridization of LiO₂ could exhibit abundant adsorption sites and superior OER/ORR kinetics. Furthermore, the 2H-MoS₂@Co₃S₄ possesses optimized E_d value of Co with moderate Co–O antibonding orbital occupancy, which contributed to the optimal bonding interaction between Co sites and oxygen-containing intermediates for suitable oxygen adsorption and boosted electrocatalysis (Fig. 6c). Distinct differences between four models before and after adsorption of LiO₂ can be traced in Fig. S46, indicating effective adsorption of 2H-MoS₂@Co₃S₄ due to the optimized d-p orbital hybridization with LiO₂. Detailed adsorption energy of the four catalysts for the reaction intermediates is clearly observed in Fig. 6d. It is widely accepted that the intrinsic adsorption energy (ΔE_ads) of LiO₂ intermediate on catalyst surfaces is a crucial factor in regulating the nucleation and growth mechanisms of Li₂O₂ during discharge process [66, 67]. Strong affinity of LiO₂ on Co sites of Co₃S₄ promoted the accumulation of LiO₂ with high local concentration and hence resulted in generation of dense film-like Li₂O₂ through surface growth pathway, which is unfavorable for mass diffusion and thus suppressed the subsequent Li₂O₂ formation. As for 1T-MoS₂@Co₃S₄ and 2H-MoS₂, due to their poor adsorption capability of LiO₂, LiO₂ tended to be dissolved in electrolyte and disproportionate to form Li₂O₂ nanorods. The oxidation of Li₂O₂ nanorods brought about high overpotentials because of poor interfacial contact between nanorods and catalytic active surfaces.

The TDOS data for Co₃S₄, 1T-MoS₂@Co₃S₄, 2H-MoS₂@Co₃S₄, and 2H-MoS₂ before and after adsorptions of different reaction product are given in Fig. S47. TDOS result of 2H-MoS₂@Co₃S₄ around the Fermi energy level exhibits a significant increase, and this represents enhanced electrical conductivity upon loading of the intermediates, which due to the optimized interaction between 2H-MoS₂@Co₃S₄ and discharge product. However, opposite change for Co₃S₄ appears, suggesting that the insulated discharged product decayed the electrical conductivity of Co₃S₄ due to the strong adsorption ability. Phase diagram of the cathode reactions provides in-depth insights into the process of discharge product evolution, as depicted in Fig. S48. As for 2H-MoS₂@Co₃S₄, LiO₂ with a low ΔG formed spontaneously in the interval of 2.17–2.37 V. When the discharge voltage is below 2.17 V (the cross point between Li₂O₄ and Li₃O₄ line), Li₂O₂ with a thermodynamic advantage began nucleation. As the voltage dropped to 2.05 V, steady Li₄O₄ clusters generated. These results demonstrate that Li_2-xO₂ easily grew on 2H-MoS₂@Co₃S₄ surfaces, which could change the nucleation route of Co₃S₄ after the formation of 1T/2H-MCS. To understand the configuration of oxygen-containing intermediates during the discharging process, O–O bond lengths, O–Li–O bond angles, and adsorption distances of oxygen-containing intermediates of four catalysts are compared in Figs. S49–S52. Due to the strongest/weakest d-p orbital hybridization of Co₃S₄ and MoS₂, different oxygen-containing intermediates are all closest/farthest to/from the catalyst surfaces, which is in accordance with adsorption energy results. Furthermore, the bond lengths of O–O in oxygen-containing intermediates on 2H–MoS₂@Co₃S₄ are in the middle of those on 1T-MoS₂@Co₃S₄ and Co₃S₄, and the O–Li–O bond angles on 2H-MoS₂@Co₃S₄ are twisted to a larger angle (the standard value is 45.6°), leading to a facile cleavage of the O–O bonds. Accordingly, heterojunctions of 2H-MoS₂@Co₃S₄ could play an effective role in optimizing the adsorption of oxygen species and accelerating the O–O bond breakages.

Free energy calculations were used to determine the energy required for each reaction stage. OER/ORR overpotentials are defined as η_OER = U_C-U₀/η_ORR = U₀-U_DC. U_C, U_DC, and U_0, respectively, denote minimum charge, maximum discharge, and equilibrium potentials. 2H-MoS₂@Co₃S₄ cathode exhibits the lowest η_OER/η_ORR (0.16/0.18 V) among those of Co₃S₄ (0.86/2.27 V), 1T-MoS₂@Co₃S₄ (1.17/0.81 V), and MoS₂ (0.70/1.54 V) in Fig. 6f, revealing significantly promoted electrocatalytic activities of the 2H-MoS₂@Co₃S₄ toward oxygen redox reactions in LOBs due to the moderate d-p orbital hybridization with oxygen-containing intermediates (Fig. 6e, g). Based on the above results, volcano-like relationship between binding energy of oxygen-containing intermediates with different catalysts and overpotentials can be established in Fig. 6h. Optimal binding energy of oxygen-containing intermediates on 2H-MoS₂@Co₃S₄ is of great essence to improve its electrocatalytic activities, which is in conformity with the Sabatier principle. Particularly, owing to weak adsorption of oxygen-containing intermediates, the coverages of reactive oxygen species on 1T-MoS₂@Co₃S₄ and 2H-MoS₂ are relatively low and thus contributed less to oxygen electrocatalytic activities and electron transfer. On the contrary, excessive adsorption of reactive oxygen species on Co₃S₄ poisoned active sites and thus impeded reaction kinetics. Therefore, an appropriate binding energy of oxygen species on 2H-MoS₂@Co₃S₄ could not only increase reactive oxygen species coverage, but also prevent poisoning effect caused by blockage of active sites, thereby beneficial for remarkable electrochemical performance of 1T/2H-MCS cathodes.

Rational formation and decomposition pathways of the discharge product on different cathodes are elucidated in Scheme 1. It is proposed that intermediate product LiO₂ first generated via a one-electron reduction reaction on the cathode surfaces (Eq. (4)). For the 2H-MoS₂ cathode, due to extremely weak binding energy for LiO₂, the intermediate was then dissolved in the electrolyte, defined as LiO_2(sol), which was further converted to the rod-like Li₂O₂ via solution growth pathway (Fig. S29b and Eq. (5)). During charging process, low catalytic activities, poor electrical conductivity, and limited contact with rod-like Li₂O₂ resulted in extremely low discharge specific capacities and unsatisfactory cycle performance of 2H-MoS₂ cathodes. Inversely, the strong adsorption energy of LiO₂* on the Co₃S₄ surfaces promoted the formation of dense film-like Li₂O₂ through surface growth pathway (Fig. 5b and Eq. (6)). They were closely adhered to the cathode surfaces to slow down the kinetic diffusion of electrons/ions on the cathode catalyst, which lowered the activation and led to high interfacial polarization. Fortunately, owing to moderate adsorption energy of 1T/2H-MCS for LiO₂*, Li₂O₂ were grown on 1T/2H-MCS cathode via solution and surface dual pathways, forming two kinds of different morphologies (Fig. 5e). Enhanced catalytic activity of 1T/2H-MCS via the synergistic effects of double complementary heterojunctions from 1T-MoS₂@Co₃S₄ and 2H-MoS₂@Co₃S₄ changed the affinity between the catalyst surfaces and the reaction intermediates to affect LiO₂ nucleation, successfully achieving the ideal Li₂O₂ morphologies.

$${\text{O}}_2+{\text{e}}^{-}+{\text{Li}}^{+}\to {\text{LiO}}_2$$ 4

$${\text{2LiO}}_{2(\text{sol})}\to {\text{Li}}_2{\text{O}}_2+{\text{O}}_2$$ 5

$${\text{LiO}}_2\ast +{\text{Li}}^{+}+{\text{e}}^{-}\to {\text{Li}}_2{\text{O}}_2\ast$$ 6

Scheme 1.

Proposed electrocatalytic mechanisms of a 1T/2H-MCS, b Co₃S₄, and c 2H-MoS₂ cathodes during ORR/OER processes

Conclusions

In summary, ZIF-67 as a template was used to synthesize 1T/2H-MCS composites via one-pot hydrothermal method with two-step temperature-raising process, which served as efficient bifunctional electrocatalysts for LOBs. It is evident that 1T/2H-MCS cathodes display extraordinary electrochemical properties, including high reversible discharge/charge specific capacities of 18,721/18,500 mAh g⁻¹ at 100 mA g⁻¹ and ultralong cycle life of 2420 h at 100 mA g⁻¹ under a cutoff specific capacity of 1000 mAh g⁻¹. When applied in pouch-type LOBs, 1T/2H-MCS cathodes exhibit an impressive and promising potential for practical applications. Based on DFT calculations and experimental results, it is proposed that significantly improved OER/ORR performance of 1T/2H-MCS cathodes is owing to the complementary effects of double heterojunctions derived from the charge donation form Co to Mo ions. Firstly, 1T-MoS₂@Co₃S₄ provided unblocked electron transport channels of Co–S–Mo bonds, well redistributing electrons within 1T/2H-MCS and hence significantly improving its interfacial electron transfer kinetics and electrocatalytic activities. Then, 2H-MoS₂@Co₃S₄ possesses a moderate antibonding orbital occupancy when absorbed oxygen-containing intermediates, optimizing the morphology of discharge product and accelerating their decomposition by modulating the adsorption energy. Furthermore, 1T/2H-MCS holds a dodecahedral hollow structure inherited from ZIF-67 and is decorated with MoS₂ nanosheets, which could restrict layer-oriented growth of MoS₂ and increase specific surface area to expose abundant stable active sites. This complementary relationship provides guidance for designing efficient electrocatalysts with double heterojunctions for LOBs.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Y. C. Dou and Z. Liu conducted experimental measurements and result analyses, as well as drafted the manuscript. J. Wang and L. L. Zhao conceived the research and polished the manuscript. X. A. Li, Z. Shang and L. Wang contributed to the theoretical model construction and DFT calculations. J. Zhang, F. P. Meng, Y. Liu, and Z. D. Zhang discussed the results and revised the manuscript. All authors commented and agreed on the final version.

Declarations

Conflict of interest

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