Interface engineering of 0D–2D CoSe₂/ZnSe@MXene heterostructured electrodes for high-performance lithium-ion batteries

Abstract

High-capacity conversion-type anodes with high volume expansion and low conductivity face limitations in meeting the high energy density demands of lithium-ion batteries. Herein, MOF-derived CoSe₂/ZnSe bimetallic selenide nanoparticles are confined in layered Ti₃C₂T_x MXene (CoSe₂/ZnSe@MX) as electrodes for high-performance lithium-ion batteries by an in situ self-assembly and selenization strategy. The interconnected conductive MXene networks can not only provide highways for charge transfer but can also effectively accommodate large volume expansion, improving structural stability. Meanwhile, the bimetallic CoSe₂/ZnSe nanoparticles with heterostructures and Se vacancies offer abundant redox reaction sites, promote Li-ion diffusion, and enhance Li-ion adsorption. Thus, the CoSe₂/ZnSe@MX electrodes exhibit a remarkable capacity of 830.8 mA h g⁻¹ at 0.1 A g⁻¹, high-rate capability of 290.8 mA h g⁻¹ at 5 A g⁻¹, and superior cycling stability with 63.1% capacity retention after 2000 cycles. Furthermore, the full cell demonstrated practical applicability with a high capacity of 156 mA h g⁻¹ at 0.1C. This facile technique is promising for constructing high-performance energy storage devices.

---

Mingzheng Ge

Mingzheng Ge is a professor at the School of Textile and Clothing, Nantong University. He received his PhD degree from the College of Textile and Clothing Engineering at Soochow University in 2018. During 2016–2017, he was a visiting scholar at Nanyang Technological University. He was a postdoctoral researcher at the Institute of Applied Physics and Materials Engineering at the University of Macau from 2020 to 2022. He has been listed among World's Top 2% Scientists for 2023 & 2024. His research interest focuses on advanced materials for intelligent textiles, wearable electronics and high energy density/safety energy storage devices, such as lithium-ion batteries and aqueous Zn-ion batteries.

---

1. Introduction

Lithium-ion batteries (LIBs) have dominated the market from portable electronics to electric vehicles (EVs) since their commercialization in the 1990s due to their high energy density.^1–3 However, conventional LIBs using graphite anodes cannot meet the increasing energy demands for the continuous expansion of the EV market.^4–6 A wide range of conversion (i.e., MoS₂ and CoSe₂) and alloy (i.e., Si and Ge)-type anode materials have been reported to replace commercial graphite to satisfy the increasing demands of energy/power density.^7–11 Among them, transition metal selenides have been considered some of the most promising candidates for LIBs due to their high theoretical capacity.^12–14 However, the huge volume expansion during Li-ion insertion/extraction usually causes pulverization of active materials, structural collapse, and continuous growth of the solid electrolyte interface (SEI), resulting in rapid capacity fading.^15–18 Meanwhile, the low intrinsic ionic and electronic conductivity results in sluggish reaction kinetics and inferior rate performance.^19,20 Therefore, rational structure and interface design of high-capacity transition metal selenides is of importance for constructing high-energy LIBs.^21–23

From a methodology perspective, nanostructured strategies, atom substitution on transition metals, and compositing transition metal selenides with other materials (e.g., carbon, graphene, MXenes, etc.) have been widely adopted to overcome these obstacles.^24–27 Although nanostructured transition metal selenide materials have exhibited strain relaxation to volume changes, they still suffer from the inherent fragile feature and the continuous SEI growth due to the large specific surface area.^28,29 Recently, metal–organic framework (MOF) derived carbon-coated transition metal selenides have exhibited high capacity due to the well-defined porous architectures, more active reaction sites and the confinement effect.^30–33 However, MOF-derived discontinuous particles are prone to aggregate to form bigger particles and finally collapse during the charge/discharge process, resulting in an unsatisfactory performance.^34–36 Meanwhile, compared to monometallic selenides, bimetallic selenides have exhibited preferable electrical conductivity and Li-ion adsorption capability due to the heterointerface with several defects and lattice distortions.^37–39 Thus, assembling bimetallic selenide nanoparticles on conductive two-dimensional supports can make them uniformly distributed with improved electronic conductivity.^40–43 For instance, SnSe₂/NiSe₂ heterointerfaces with rich Se vacancies were embedded into N-doped carbon via a facile hydrothermal process and selenization strategy.⁴⁴ The introduction of heterojunction engineering and Se vacancies can accelerate charge transfer efficiency and improve Na-ion adsorption, while N-doped carbon accommodated the volume expansion. Thus, the electrode delivered long term cycling stability with 322.7 mA h g⁻¹ after 7500 cycles and excellent rate capability with 314.6 mA h g⁻¹ at 10 A g⁻¹. From a materials perspective, MXenes (MXs), typical representatives of two-dimensional layered transition metal carbides and nitrides, have been widely used in electrochemical energy storage devices because of their superior conductivity, abundant polar functional groups, and favorable pseudocapacitive performance.^45–50 Inspired by this, we hypothesize that confining bimetallic selenides in layered MXs with 0D–2D heterointerfaces with rich vacancies can alleviate the huge volume expansion, improve charge carrier transfer, and enhance Li-ion adsorption capability, thus enabling hybrid electrodes with large energy storage performance.

Herein, bimetallic CoSe₂/ZnSe nanoparticles were nanoconfined in layered Ti₃C₂T_x MX (CoSe₂/ZnSe@MX) via self-assembly of bimetallic MOFs on MX, followed by an in situ selenization strategy. Ti₃C₂T_x MX nanosheets, working as an elastic skeleton to encapsulate CoSe₂/ZnSe nanoparticles, not only facilitated rapid electron and Li-ion transfer, but also effectively accommodated the large volume expansion and prevented aggregation of bimetallic CoSe₂/ZnSe nanoparticles during Li-ion insertion/extraction processes, thus improving the structure stability. Meanwhile, the CoSe₂/ZnSe nanoparticles with heterojunctions and abundant Se vacancies boosted charge transfer, provided more active sites, and enhanced Li-ion adsorption. Therefore, the CoSe₂/ZnSe@MX electrodes exhibited a remarkable capacity of 830.8 mA h g⁻¹ at 0.1 A g⁻¹ and high-rate capability of 290.8 mA h g⁻¹ at 5 A g⁻¹ owing to the synergetic effects of nanoconfinement, heterostructures and Se vacancies. Meanwhile, it delivered a high capacity of 536.7 mA h g⁻¹ after 1000 cycles and even exhibits superior cycling stability with 63.1% capacity retention after 2000 cycles. Furthermore, the CoSe₂/ZnSe@MX//LFP full cell displayed a high initial capacity of 156 mA h g⁻¹ at 0.1C and exhibited superior cycling stability with only 0.4% capacity decay in each cycle. This facile technique provides new opportunities for constructing high-capacity electrodes for other energy storage devices, such as Na/K-ion batteries, supercapacitors, and solar cells.

2. Experimental section

2.1. Materials

The MAX powder (Ti₃AlC₂, 400 mesh) was purchased from Jilin 11 Technology Co., Ltd. Hydrofluoric acid (HF, 40%), alcohol (C₂H₅OH, AR) and methanol (CH₃OH, AR) were purchased from Sinopharm Co., Ltd. Cobalt nitrate hexahydrate (Co(NO₃)₂·6H₂O, AR), zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O, AR), 2-methylimidazole (mIM), N-methyl pyrrolidone (NMP), polyvinylidene fluoride (PVDF) and Se powder were purchased from Aladdin Co., Ltd.

2.2. Synthesis of MX (Ti₃C₂T_x)

1 g of Ti₃AlC₂ MAX powder was slowly added to 20 ml of HF acid solution (40%). The mixture was stirred for 24 hours at 45 °C. The sediment was collected and washed with deionized water until the supernatant pH = 7. The sediment was re-dispersed in deionized water and ultrasonically treated with argon for 30 min. Finally, the sample was freeze-dried overnight to get MX nanosheets.

2.3. Synthesis of CoZn-MOFs@MX

30 mg MX nanosheets were dispersed in 20 ml methanol solution and sonicated for 20 min. 200 mg PVP was added to the previous solution and stirred for 15 min. Then, 25 mg Zn(NO₃)₂·6H₂O and 200 mg Co(NO₃)₂·6H₂O were dispersed into the solution and stirred for 30 min, which was named solution A. Meanwhile, 500 mg mIM was added into 20 ml methanol solution and sonicated for 10 min, which was named solution B. Subsequently, solution B was poured into solution A with continuous stirring. After 4 h of aging, CoZn-MOFs@MX powder was obtained by washing with water and ethanol three times and freeze-dried overnight. For the control group, CoZn-MOFs were constructed by a similar method without adding MX.

2.4. Synthesis of 0D–2D CoSe₂/ZnSe@MX

The CoZn-MOFs@MX and Se powders with a mass ratio of 1 : 1 were mixed homogeneously and heated at 600 °C for 3 h with a heating rate of 2 °C min⁻¹ under a nitrogen atmosphere, thus achieving 0D–2D CoSe₂/ZnSe@MX. For the control group, CoSe₂/ZnSe was synthesized using a similar method, maintaining a mass ratio of 1 : 1 between CoZn-MOFs and Se powder.

2.5. Characterization

Field emission scanning electron microscopy (SEM, ZEISS, Gemini 300) and transmission electron microscopy (TEM, FEI, Tecnai G20) were used to observe the morphology and nanostructures of the prepared samples. An energy dispersive X-ray (EDX) spectrometer fitted to TEM was used for the elemental analysis. The crystal structures were observed by X-ray Diffraction (XRD, Philips, X'pert-Pro MRD) with Cu-Kα radiation between 5° and 90°. A Raman spectrometer (Horiba Scientific, LabRam HR Evolution) was used to evaluate the structure and defects of the samples. X-ray photoelectron spectroscopy (XPS, Thermo Scientific, NEXAS) was used to determine the chemical compositions of the samples. Electron paramagnetic response spectroscopy (EPR, Bruker, EMX PLUS) was used to detect Se vacancies. The specific surface area and pore size distribution were measured via nitrogen adsorption isotherms at 77 K (Quantachrome, Autosorb-iQ).

2.6. Electrochemical measurements

The performance of half cells was studied using CR2032 coin cells. Active materials (80 wt%), PVDF (10 wt%), and carbon black (10 wt%) were mixed in NMP and ground for 40 min to form a homogeneous slurry. The mixed slurry was then coated on the Cu foil using a blade and dried at 100 °C for 12 h in a vacuum oven. Then, the Cu foil was cut into discs with a diameter of 10 mm. The mass loading of active materials was ∼1.1 mg cm⁻². Half cells were assembled in an argon-filled glove box by using Li metal as the counter electrode, Celegard 2400 as the separator, and 1.0 M LiPF₆ in EC : DMC : EMC (1 : 1 : 1 vol%) with 1.5% vinylene carbonate as the electrolyte. The galvanostatic charge/discharge tests and galvanostatic intermittent titration technique (GITT) were performed at 0.01–3 V on a NEWARE battery detection system. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) test experiments were conducted on an electrochemical workstation (CHI 660E). The CV was tested with the potential window of 0.01–3 V at the scan rate varying from 0.1 to 1 mV s⁻¹, while EIS was conducted with a frequency ranging from 0.01 to 10⁵ Hz with a voltage amplitude of 5 mV. Regarding the full cells, the CoSe₂/ZnSe@MX electrode was activated by charging and discharging in a half cell after 10 cycles and then obtained as the anode, paired with a commercial LiFePO₄ (LFP) cathode. The weight ratio of LFP, carbon black, and PVDF was 8 : 1 : 1. The mass loading of active materials was ∼2.2 mg cm⁻² for the LFP cathode. The capacity ratio of negative to positive electrodes was controlled at ∼1.1. The electrolyte and separator were the same as those used in half cells. The full cells were galvanostatically charged/discharged from 1 to 3.8 V at different current densities.

2.7. Theoretical analysis

The detailed procedure for density functional theory calculations can be found in the ESI.†

3. Results and discussion

The 0D–2D CoSe₂/ZnSe@MX heterostructured composites were successfully synthesized via a combination of self-assembly and selenization strategies, the fabrication process of which was illustrated in Fig. 1. First, Ti₃C₂T_x MX was synthesized by selectively etching out the metallic Al element in the precursor Ti₃AlC₂ MAX phase with HF solutions and then sonicated to form accordion-like delaminated 2D Ti₃C₂T_x MX nanosheets. Second, electropositive Co and Zn ions were adsorbed on the surface of electronegative MX with functional groups (i.e., –OH and –F) via electrostatic interactions. Third, with the addition of mIM, CoZn-MOF nanoparticles were uniformly anchored on both sides of MX nanosheets through self-assembly (CoZn-MOFs@MX). Finally, CoSe₂/ZnSe nanoparticles were in situ formed as derivatives of CoZn-MOFs after selenization (CoSe₂/ZnSe@MX), inducing CoSe₂/ZnSe heterojunctions and rich Se vacancies in the composites. This unique structure design provided CoSe₂/ZnSe@MX with multiple advantages: (i) the 2D MX nanosheets provided fast pathways for electrons and ion transportation, accommodated the huge volume changes and prevented aggregation of CoSe₂/ZnSe nanoparticles, guaranteeing the structure stability; (ii) the formed CoSe₂/ZnSe heterointerface could enhance the Li-ion adsorption and boost charge transfer to accelerate the reaction kinetics, achieving outstanding fast-charging performance; (iii) introduction of Se vacancies induced more reaction active sites and promoted ion diffusion, enabling high-performance LIBs.

Fig. 1 Schematic illustration of the fabrication process of CoSe₂/ZnSe@MX via a combination of self-assembly and selenization strategies. The interconnected conductive MX networks can not only provide highways for charge transfer, but can also effectively accommodate the large volume expansion, improving the structural stability of the electrode. Meanwhile, the bimetallic CoSe₂/ZnSe nanoparticles with heterostructures and Se vacancies offer abundant redox reaction sites, boost charge transfer, and enhance Li-ion adsorption.

The morphology and microstructure of the as-prepared CoSe₂/ZnSe@MX were characterized by SEM and TEM. The chemically exfoliated Ti₃C₂T_x MX nanosheets exhibited uniform size with a large lateral size of ∼10 μm and layer spacing of 100–200 nm (Fig. 2a). Meanwhile, it can be seen that the CoZn-MOFs in each layer of the MX nanosheets had a regular polyhedral morphology and a smooth surface with a size of ∼150 nm (Fig. S1†) due to the confinement of the MX layer, promoting Li-ion diffusion and enabling fast charging capability. After selenization, ultrafine CoSe₂/ZnSe nanoparticles with similar morphology were uniformly distributed on both sides of each MX nanosheet (Fig. 2b). This observation is also evident in the TEM image (Fig. 2c). The hybrid materials maintained their original morphology with a 0D–2D heterointerface after selenization. As shown in high-resolution transmission electron microscope (HRTEM) images, the lattice spacings of 0.26, 0.17 and 0.17 nm corresponded to the (210), (112) and (108) crystal planes of CoSe₂, ZnSe and MX, respectively. The phase boundaries between CoSe₂ and ZnSe with heterojunctions enabled more active sites and defects for Li-ion storage (Fig. 2d). Furthermore, the selected area electron diffraction (SAED) image in Fig. 2e was explored to confirm the crystal structure of CoSe₂/ZnSe@MX, where the yellow rings represented the (200) planes of ZnSe, the blue rings represented the (210) planes of CoSe₂, and the green rings represented the (104) planes of MX. The corresponding EDX analysis of CoSe₂/ZnSe@MX indicated that Ti, C, Co, Zn and Se elements were distributed homogeneously in the whole nanosheet (Fig. 2f & S2†), further confirming that ultrafine CoSe₂/ZnSe heterostructured nanoparticles were uniformly distributed in the MX matrix.

Fig. 2 Morphology of pristine MX and CoSe₂/ZnSe@MX composites. SEM images of (a) MX and (b) CoSe₂/ZnSe@MX composites; (c) TEM image, (d) HRTEM image, (e) SAED diffraction patterns and (f) the corresponding EDX mapping of CoSe₂/ZnSe@MX composites.

The crystalline phases of pure MX, CoSe₂/ZnSe and CoSe₂/ZnSe@MX composites were analyzed with XRD as shown in Fig. 3a. After selenization, the main diffraction peaks of CoSe₂/ZnSe and CoSe₂/ZnSe@MX were correlated well with the hexagonal ZnSe (JCPDS no. 89-2940) and trogtalite CoSe₂ (JCPDS no. 89-2002). The representative peaks at 30.5, 34.2, 37.6, 43.7, 51.7, 58.9, 74.1 and 86.2° matched well with the (200), (210), (211), (220), (311), (321), (421), and (511) planes of phase CoSe₂, while the representative peaks at 25.7, 45.4, 48.9 and 55.4° match well with the (100), (110), (103), and (004) crystalline planes of phase ZnSe. Furthermore, the disappearance of the peak at 39.04°, corresponding to the (104) plane of MX (JCPDS no. 52-0875), confirms that the “Al” component has been successfully etched (Fig. S3†). Meanwhile, the diffraction peak of the (002) crystal plane at 8.8° was related to Ti₃C₂T_x MX in CoSe₂/ZnSe@MX composites.⁴⁵ These results all demonstrated the successful synthesis of CoSe₂/ZnSe@MX heterostructures. The Raman spectrum of the MX exhibited peaks at around 216, 395, and 597 cm⁻¹, which are attributed to its E_g and A_1g vibration modes, respectively (Fig. 3b).⁵¹ Besides, the peaks at 1340 and 1590 cm⁻¹ of CoSe₂/ZnSe@MX composites belonged to amorphous and graphitized carbon derived from the selenization of CoZn-MOFs. The ratio of I_D/I_G is 1.37, indicating that the interfacial interaction between CoSe₂/ZnSe and MX induced a large number of defects, increasing the electronic conductivity and promoting ion adsorption.

Fig. 3 Structure characterization of pristine MX, CoSe₂/ZnSe, and CoSe₂/ZnSe@MX composites. (a) XRD of MX, CoSe₂/ZnSe and CoSe₂/ZnSe@MX; (b) Raman spectra of MX and CoSe₂/ZnSe@MX; (c) XPS of MX, CoSe₂/ZnSe and CoSe₂/ZnSe@MX composites; high-resolution (d) Co 2p, (e) Zn 2p, and (f) Se 3d XPS spectra of CoSe₂/ZnSe@MX composites; (g) EPR spectrum of CoSe₂/ZnSe@MX composites; (h) N₂ adsorption–desorption isotherms and (i) pore size distribution of MX and CoSe₂/ZnSe@MX composites.

To further investigate compositions and chemical interactions of MX and CoSe₂/ZnSe@MX composites, an XPS test was conducted (Fig. 3c), which revealed the existence of C, O, N, Co, Zn, Se, and Ti elements in CoSe₂/ZnSe@MX. In the high-resolution Ti 2p spectrum of the CoSe₂/ZnSe@MX electrode (Fig. S4a†), the peaks at around 464.0, 463.3, 460.1, 458.1, 457.6, and 452.7 eV corresponded to the Ti–O 2p_1/2, Ti²⁺/Ti³⁺ 2p_1/2, Ti–C 2p_1/2, Ti–O 2p_3/2, Ti²⁺/Ti³⁺ 2p_3/2, and Ti–C 2p_3/2 bonds, demonstrating successful selective etching of the “Al” element in MAX,⁵² whereas in the C 1s high-resolution XPS spectrum, the four main peaks at 283.5, 283.9, 285.9, and 294.6 eV corresponded to Ti–C, C–C, C–O and O–C=O bonds, respectively (Fig. S4b†), indicating the carbon coating around CoSe₂/ZnSe nanoparticles after selenization, which was consistent with the Raman results. The peaks located at 777.2 and 779.5 eV in the Co 2p high-resolution XPS spectrum corresponded to Se–Co(II) and Co–O(III) of Co 2p_3/2, respectively. Likely, the peaks appearing at 792.3 and 795.6 eV corresponded to Co–Se(II) and Co–O(III) of Co 2p_1/2, confirming the existence of Co²⁺ in CoSe₂ (Fig. 3d).⁵³ The high-resolution spectra of Zn 2p showed two peaks at 1043.7 and 1020.6 eV, corresponding to Zn 2p_3/2 and Zn 2p_1/2, respectively (Fig. 3e). Furthermore, the Se–O–Se bond that appeared at 58.3 eV can be attributed to the chemical reaction between oxygen and the slightly residuary metalloid Se, indicating the +2 valence state of Zn in ZnSe (Fig. 3f).⁵⁴ For the Se 3d high-resolution XPS spectrum, the peaks at 53.5 and 54.5 eV could be assigned to the core level bands of Se 3d_5/2 and Se 3d_3/2, respectively (Fig. 3g). The N 1s spectrum exhibited deconvoluted peaks at 397.8, 399.6 and 400.3 eV that can be attributed to pyridinic N, pyrrolic N, and graphitic N, respectively (Fig. S4c†). The N-doped carbon coating from selenization of CoZn-MOFs contributed to enhanced electronic conductivity. What's more, the O 1s spectrum indicated that the oxygen-related functional groups existed in CoSe₂/ZnSe@MX (Fig. S4d†). The presence of Se vacancies in CoSe₂/ZnSe@MX was further substantiated through EPR spectroscopy (Fig. 3g). The EPR signal observed at g = 2.000 in the CoSe₂/ZnSe@MX composites indicated the successful introduction of Se vacancies and heterojunction after selenization.⁵⁵ The presence of Se vacancies can offer abundant defect sites, thereby providing more sites for Li⁺ adsorption and improving ionic diffusion kinetics. The N₂ adsorption/desorption isotherm of CoSe₂/ZnSe@MX displayed typical type-IV isotherms, suggesting its mesoporous structure. The specific surface area of CoSe₂/ZnSe@MX (39.7 m² g⁻¹) was larger than that of MX (8.5 m² g⁻¹), indicating that CoSe₂/ZnSe@MX could provide large surface area electrolyte absorption and more active sites for Li-ion adsorption to improve the electrochemical performance (Fig. 3h and i). Meanwhile, mesopores (∼11.20 nm) in MOF-derived CoSe₂/ZnSe heterostructured nanoparticles could effectively accelerate the Li-ion diffusion kinetics.

The initial five cycles of CV curves of CoSe₂/ZnSe@MX in the voltage range of 0.1–3.0 V at a scan rate of 0.1 mV s⁻¹ are displayed in Fig. 4a. The peak at 0.31 V can be attributed to the formation of the SEI layer and the decomposition of electrolytes in the initial cathodic reduction process. However, the reduction peaks disappeared in the subsequent cycles, indicating that a stable SEI layer had been formed after the first cycle. The other peak at 0.72 V could be attributed to the alloying reaction of Ti₃C₂T_x (eqn (1)). During the cathodic scan, the peak at 1.44 V represented an alloying process of CoSe₂. With the intercalation of Li⁺ into CoSe₂, it transformed into CoSe and then into Co (eqn (2) & (3)).⁵⁶ Similarly, another peak at 1.75 V belonged to the transformation of ZnSe into Zn (eqn (4)). The appearance of the oxidation peaks located at 2.09 and 2.27 V could be ascribed to the oxidation of metallic Co and Zn to CoSe₂ and ZnSe, respectively, which represented the lithium ion extraction process. The highly overlapped CV curves from the 2nd to the 5th cycle indicated that CoSe₂/ZnSe@MX had good electrochemical stability and reversibility.

Ti₃C₂T_x + yLi⁺ + ye⁻ → Ti₃C₂T_xLi_y (1)

2Li⁺ + 2e⁻ + CoSe₂ ↔ Li₂Se + CoSe (2)

CoSe + 2Li⁺ + 2e⁻ ↔ Li₂Se + Co (3)

ZnSe + 2Li⁺ + 2e⁻ ↔ Li₂Se + Zn (4)

Fig. 4 Electrochemical characterization of pristine MX, CoSe₂/ZnSe, and CoSe₂/ZnSe@MX composites. (a) Initial five CV curves of the CoSe/ZnSe@MX electrode at scan rates of 0.1 mV s⁻¹; (b) the b value at the oxidation and reduction peaks of the CoSe₂/ZnSe@MX electrode; (c) capacitive contribution of the CoSe₂/ZnSe@MX electrode at 5 mV s⁻¹ and (d) at different scan rates ranging from 0.1 to 5 mV s⁻¹; (e) EIS curve of MX, CoSe₂/ZnSe and CoSe₂/ZnSe@MX; (f) GITT profiles of CoSe₂/ZnSe and CoSe₂/ZnSe@MX electrodes at 0.1 A g⁻¹.

The energy storage performance of CoSe₂/ZnSe@MX was also evaluated using CV curves. Fig. S5† presents the CV curves of the CoSe₂/ZnSe@MX electrode at scan rates ranging from 0.1 to 5 mV s⁻¹. The relationship between current (i) and sweep rate (v) can be explained as i = av^b, where a and b are constants.⁵⁷ The b value is calculated in reduction and oxidation processes of CoSe₂/ZnSe@MX (Fig. 4b). The b value can be determined using the slope of log(v) − log(i). When the b value is 0.5, the diffusion behavior dominates the charge/discharge process, while when the b value is 1.0, the pseudocapacitive behavior is dominant. The b values of the CoSe₂/ZnSe@MX electrode in the charging and discharging processes were 0.79 and 0.91, respectively, indicating the surface-controlled pseudocapacitive behavior. Furthermore, the pseudocapacitive contribution can be calculated based on i = k₁v + k₂v^1/2 in the overall reaction, where k₁ represents capacitive contributions and k₂ represents diffusion contributions. At a scanning rate of 5 mV s⁻¹, the capacitance contribution of the CoSe₂/ZnSe@MX electrode was approaching 89.9%, indicating that it had a high-rate charging capability (Fig. 4c). The high capacitive contribution was due to the accelerated charge transfer at the heterojunction interfaces, enhanced Li-ion adsorption, and improved electronic conductivity. The capacitive contribution of CoSe₂/ZnSe@MX electrodes at different scanning rates is summarized in Fig. 4d, indicating that the CoSe₂/ZnSe@MX electrode exhibited a high-capacity storage and ultrafast charging/discharging capability. EIS and the GITT were implemented to explore the reaction kinetics of CoSe₂/ZnSe@MX electrodes. Fig. 4e shows the EIS diagram of CoSe₂/ZnSe@MX, which was obtained after 50 charge/discharge cycles at 1 A g⁻¹. The EIS curve typically includes a straight line and a semicircular line. The straight line is related to the lithium-ion diffusion in the electrode, which represents the low-frequency region of interface charge transfer impedance, which is called Warburg impedance. The semicircle is related to the electron transfer process, indicating the high lithium diffusion impedance frequency region, representing the charge transfer resistance (R_ct). The charge transfer resistance of the CoSe₂/ZnSe@MX electrode (77.6 Ω) was smaller than that of the CoSe₂/ZnSe (88.8 Ω) and MX (132.3 Ω) electrodes. Meanwhile, the slope of CoSe₂/ZnSe@MX in the low frequency region was larger than that of the others, indicating the low solid-state Li-ion diffusion resistance. In addition, the Li⁺ ion diffusion coefficient of the CoSe₂/ZnSe@MX electrode (2.15 × 10⁻⁷ cm² s⁻¹) was nearly 4 times higher than that of CoSe₂/ZnSe (5.17 × 10⁻⁸ cm² s⁻¹), indicating that the synergistic effects of 2D MX layers and heterostructures with rich interfacial phase boundaries, along with the presence of Se vacancies boosted charge transfer and accelerated the Li-ion diffusion (Fig. 4f & S6†).

The rate capabilities of MX, CoSe₂/ZnSe, and CoSe₂/ZnSe@MX electrodes at different current densities are shown in Fig. 5a. The CoSe₂/ZnSe@MX electrode delivered reversible specific capacities of 830.7, 524.5, 498.6, 456.5, 397.2, and 290.8 mA h g⁻¹ at current densities of 0.1, 0.2, 0.5, 1, 2, and 5 A g⁻¹, respectively. The initial charge/discharge specific capacity of CoSe₂/ZnSe@MX is 830.7 and 528.5 mA h g⁻¹ at 0.1 A g⁻¹, with an initial coulombic efficiency of 69.1% (Fig. S7†). Furthermore, the capacity could recover to 687.8 mA h g⁻¹ when the current density was increased to 0.1 A g⁻¹, indicating excellent reversibility. Although the CoSe₂/ZnSe electrode displayed a higher capacity than the CoSe₂/ZnSe@MX electrode at a low current density of 0.1 and 0.3 A g⁻¹, the capacity decreased quickly when increasing current density with an inferior rate capability. However, pure MX electrodes exhibited a relatively low specific capacity when compared to CoSe₂/ZnSe and CoSe₂/ZnSe@MX electrodes (Fig. S8†), which was selected as the matrix for CoSe₂/ZnSe to suppress the volume expansion and enhance charge transfer efficiency. After 300 charge/discharge cycles at 1 A g⁻¹, the CoSe₂/ZnSe@MX electrode delivered a high capacity of 728.2 mA h g⁻¹, which was higher than that of MX (218.5 mA h g⁻¹) and CoSe₂/ZnSe (180.6 mA h g⁻¹). In particular, CoSe₂/ZnSe@MX still exhibited a specific capacity of 536.7 mA h g⁻¹ after 1000 cycles at 1 A g⁻¹. Even up to 2000 cycles, the CoSe₂/ZnSe@MX electrode maintained a specific capacity of 293.9 mA h g⁻¹ with 63.1% capacity retention (Fig. 5b). The increase in the initial specific capacity of the CoSe₂/ZnSe@MX electrode can be attributed to the activation of the electrode material. Furthermore, it can be noticed that the long cycling performance of the CoSe₂/ZnSe@MX electrode was superior to that of most of the previously published literature on transition metal selenide-based anodes (Fig. 5c & Table S2†).^56–70 Besides, the CoSe₂/ZnSe@MX electrode still maintained an intact structure similar to that of the original state, and the CoSe₂/ZnSe nanoparticles were uniformly anchored on the surface of the MX layers after 500 charge/discharge cycles (Fig. 5d & S9†). In contrast, CoSe₂/ZnSe electrodes without MX confinement exhibited an inferior rate of long-term cycling stability. CoSe₂/ZnSe electrodes displayed a low capacity of 171.7 mA h g⁻¹ after 2000 charging/discharging cycles. To further demonstrate the practicability of CoSe₂/ZnSe@MX, a full cell was assembled using the CoSe₂/ZnSe@MX anode coupled with a commercial LFP cathode, and the electrochemical performance is displayed in Fig. 5e. The CoSe₂/ZnSe@MX//LFP full cell displayed a high initial capacity of 156 mA h g⁻¹ at 0.1C (1C = 140 mA h g⁻¹) and exhibited superior cycling stability with only 0.4% capacity decay in each cycle (Fig. S10 & S11†).

Fig. 5 Electrochemical performance and DFT calculations of pristine MX, CoSe₂/ZnSe, and CoSe₂/ZnSe@MX composites. (a) Rate capabilities of MX, CoSe₂/ZnSe and CoSe₂/ZnSe@MX; (b) long cycling performance at 1 A g⁻¹ after 2000 cycles of CoSe₂/ZnSe and CoSe₂/ZnSe@MX; (c) a comparison of cycling performance with that of previously reported anodes; (d) SEM of CoSe₂/ZnSe@MX at 1 A g⁻¹ after 500 charge/discharge cycles; (e) rate capability of the CoSe₂/ZnSe@MX//LFP full cell; (f) the DOS for CoSe₂, ZnSe and CoSe₂/ZnSe heterostructures; (g) the charge density difference plots for CoSe₂, ZnSe and CoSe₂/ZnSe heterostructures; (h) the adsorption energy of Li-ions on the surfaces of CoSe₂, ZnSe, and CoSe₂/ZnSe heterostructures.

To theoretically validate the distinct electrochemical properties of the CoSe₂/ZnSe heterostructure, we employed first-principles calculations to investigate Li⁺ adsorption energy and electronic attributes at atomic and electronic scales. ZnSe exhibits a relatively low density of states (DOS) near the Fermi level, indicating its semiconducting behavior, indicated by a non-zero bandgap. In contrast, CoSe₂ shows significant states intersecting the Fermi level, suggesting the metallic behavior. Interestingly, the mixed-state DOS of CoSe₂/ZnSe increases significantly, indicating an interaction between CoSe₂ and ZnSe, which enhances the metallic properties of CoSe₂ (Fig. 5f). This improvement in the composited electrode's electronic characteristics is attributed to the formation of the heterostructures. Fig. 5g illustrates the charge density differences for ZnSe, CoSe₂, and CoSe₂/ZnSe, with yellow regions indicating electron accumulation and blue regions indicating electron depletion. Upon forming the heterostructure, CoSe₂/ZnSe exhibits electron redistribution at the interface, suggesting that the heterointerface promotes chemical bonding and implies better Li-ion adsorption, as well as the change of DOS. Moreover, the adsorption energy on the CoSe₂/ZnSe heterostructure surface is the highest (−2.17 eV) when compared to CoSe₂ (−1.61 eV) and ZnSe (−1.76 eV), indicating that the heterostructure enhances the Li-ion adsorption and boosted charge transfer (Fig. 5h). In summary, the improved electrochemical performance of the CoSe₂/ZnSe@MX electrode can be tentatively ascribed to (i) MX nanosheets served as an elastic framework to enhance the ionic/electronic conductivity and accommodated the volume expansion of CoSe_2/ZnSe nanoparticles; (ii) the introduction of Se vacancies played a crucial role in promoting ion diffusion and providing more active sites for ion adsorption; (iii) the heterostructure between CoSe₂ and ZnSe facilitated internal charge transfer and enhanced adsorption capability, resulting in outstanding fast-charging performance of the CoSe₂/ZnSe@MX electrode in the field of batteries.

4. Conclusions

In summary, a CoSe₂/ZnSe@MX electrode is programmatically designed toward high-performance LIBs via a combination of self-assembly and selenization strategies. CoSe₂/ZnSe nanoparticles were uniformly dispersed as derivatives of CoZn-MOFs on the both sides of MX nanosheets. The interconnected conductive MX networks can not only provide highways for electron/ion transportation, but can also effectively accommodate the large volume expansion and prevent aggregation of bimetallic CoSe₂/ZnSe nanoparticles during Li-ion insertion/extraction processes, improving the structural stability of the electrode. Meanwhile, the bimetallic CoSe₂/ZnSe nanoparticles with heterostructures and Se vacancies offer abundant redox reaction sites, enhance Li-ion adsorption, and suppress the restacking of MX nanosheets as a spacer. Under the synergistic effects of 2D MX nanoconfinement, CoSe₂/ZnSe heterojunction and Se vacancies, the CoSe₂/ZnSe@MX electrodes exhibit a remarkable capacity of 830.8 mA h g⁻¹ at 0.1 A g⁻¹ and high-rate capability of 290.8 mA h g⁻¹ at 5 A g⁻¹. Meanwhile, they deliver a high capacity of 536.7 mA h g⁻¹ at 1 A g⁻¹ after 1000 cycles and even exhibit superior cycling stability with 63.1% capacity retention after 2000 cycles. What's more, the CoSe₂/ZnSe@MX//LFP full cell displayed a high initial capacity of 156 mA h g⁻¹ at 0.1C and exhibited superior cycling stability with only 0.4% capacity decay in each cycle. This work provides an effective strategy for the synergistic engineering of vacancies and heterostructures to design advanced negative electrode materials for lithium-ion batteries.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

F. Wei and F. Liang contributed equally to this work. T. Yan, H. Liu, R. Li, C. Cao, and M. Ge conceived the project and designed the experiments. F. Liang, Z. Ji, H. He, and Y. Zhao fabricated the samples, conducted the characterization and performed the battery tests. Y. Kong performed the theoretical analysis. R. Li, B. Fei, W. Zhang, and M. Ge revised the manuscript. All authors analyzed the data and contributed to the discussions.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

This work was supported by the National Key R&D Program (2022YFE0206400), the National Natural Science Foundation of China (52202256 and 52102105), the Natural Science Foundation of Jiangsu Province of China (BK20220612), and the Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX23_3405). The authors also acknowledge the funds from the Young Elite Scientists Sponsorship Program of the Jiangsu Association for Science and Technology (JSTJ-2023-089). The authors thank the Nantong University Analysis and Testing Center for the technical support.

Notes and references

1. J. Li, Y. Cai, H. Wu, Z. Yu, X. Yan, Q. Zhang, T. Z. Gao, K. Liu, X. Jia and Z. Bao, Adv. Energy Mater., 2021, 11, 2003239.
2. C. D. Quilty, D. Wu, W. Li, D. C. Bock, L. Wang, L. M. Housel, A. Abraham, K. J. Takeuchi, A. C. Marschilok and E. S. Takeuchi, Chem. Rev., 2023, 123, 1327–1363.
3. Y. Jeong-Myeong, K. Deok-Gyu, K. Do-Hyeon, L. Young-Han and P. Cheol-Min, Energy Mater., 2024, 4, 400063.
4. J. Y. Cheong, S. H. Cho, J. Lee, J. W. Jung, C. Kim and I. D. Kim, Adv. Funct. Mater., 2022, 32, 2208374.
5. S. Li, K. Wang, G. Zhang, S. Li, Y. Xu, X. Zhang, X. Zhang, S. Zheng, X. Sun and Y. Ma, Adv. Funct. Mater., 2022, 32, 2200796.
6. N. Mahmood, T. Tang and Y. Hou, Adv. Energy Mater., 2016, 6, 1600374.
7. C. Cao, H. Dong, F. Liang, Y. Zhang, W. Zhang, H. Wang, H. Shao, H. Liu, K. Dong, Y. Tang, Y. Lai and M. Ge, Chem. Eng. J., 2021, 416, 129094.
8. D. Liu, Z. j. Liu, X. Li, W. Xie, Q. Wang, Q. Liu, Y. Fu and D. He, Small, 2017, 13, 1702000.
9. M. Z. Ge, Y. X. Tang, O. I. Malyi, Y. Y. Zhang, Z. Q. Zhu, Z. S. Lv, X. Ge, H. R. Xia, J. Y. Huang, Y. K. Lai and X. D. Chen, Small, 2020, 16, 2002094.
10. S. Fu, X. Wang, F. Yao, Q. He, F. Xie, X. Wu, S. Tong and M. Wu, Chem. Eng. J., 2024, 489, 151151.
11. R. H. Wang, W. Wang, Y. Z. Zhang, W. Hu, L. Yue, J. H. Ni, W. Q. Zhang, G. Pei, S. Yang and L. F. Chen, Angew. Chem., Int. Ed., 2025, 64, e202417605.
12. D. Chen, Z. Zhao, G. Chen, T. Li, J. Chen, Z. Ye and J. Lu, Coord. Chem. Rev., 2023, 479, 214984.
13. I. Hussain, S. Sahoo, C. Lamiel, T. T. Nguyen, M. Ahmed, C. Xi, S. Iqbal, A. Ali, N. Abbas, M. S. Javed and K. Zhang, Energy Storage Mater., 2022, 47, 13–43.
14. Z. Liang, M. Yang, S. Wang, B. Chang, H. Tu, Y. Shao, B. Zhang, H. Zhao, Y. Lei, J. Shen, Y. Wu and X. Hao, Chem. Eng. J., 2021, 405, 126724.
15. F. Liang, H. Dong, Z. Ji, W. Zhang, H. Zhang, C. Cao, H. Li, H. Liu, K. Q. Zhang, Y. Lai, Y. Tang and M. Ge, Sci. China Mater., 2023, 66, 1736–1746.
16. S. Zheng, H. Zhang, J. Fan, Q. Xu and Y. Min, Small, 2021, 17, 2102347.
17. H. Du, Y. Wang, Y. Kang, Y. Zhao, Y. Tian, X. Wang, Y. Tan, Z. Liang, J. Wozny, T. Li, D. Ren, L. Wang, X. He, P. Xiao, E. Mao, N. Tavajohi, F. Kang and B. Li, Adv. Mater., 2024, 36, 2401482.
18. J. H. Wang, L. F. Chen, W. X. Dong, K. Zhang, Y. F. Qu, J. W. Qian and S. H. Yu, ACS Nano, 2023, 17, 19087–19097.
19. F. Liang, H. Dong, J. Dai, H. He, W. Zhang, S. Chen, D. Lv, H. Liu, I. S. Kim, Y. Lai, Y. Tang and M. Ge, Adv. Sci, 2024, 11, 2306711.
20. Y. Tang, Y. Zhang, W. Li, B. Ma and X. Chen, Chem. Soc. Rev., 2015, 44, 5926–5940.
21. H. Wang, N. Deng, S. Wang, X. Wang, Y. Li, Q. Zeng, S. Luo, X. Cui, B. Cheng and W. Kang, J. Mater. Chem. A, 2022, 10, 23433–23466.
22. Y. Cheng, J. Huang, F. Yu, Y. Zhou, G. Li, W. Cheng, P. Duan, H. Qi and H. Xie, Chem. Eng. J., 2024, 481, 148737.
23. T. Lu, S. Dong, C. Zhang, L. Zhang and G. Cui, Coord. Chem. Rev., 2017, 332, 75–99.
24. C. Y. Cao, F. H. Liang, W. Zhang, H. C. Liu, H. Liu, H. F. Zhang, J. J. Mao, Y. Y. Zhang, Y. Feng, X. Yao, M. Z. Ge and Y. X. Tang, Small, 2021, 17, 2102233.
25. X. Zhou, Y. Cui, X. Huang, Q. Zhang, B. Wang and S. Tang, Chem. Eng. J., 2023, 457, 141139.
26. Z. Hou, X. Zhang, J. Chen, Y. Qian, L. F. Chen and P. S. Lee, Adv. Energy Mater., 2022, 12, 2104053.
27. Y. Li, Z. Yuan, D. Li, J. Li, Y. Zhang, M. Wang, G. Li, L. Wang and W. Han, ACS Nano, 2024, 18, 4733–4745.
28. Y. Jiang, M. Xie, F. Wu, Z. Ye, Y. Zhang, Z. Wang, Y. Zhou, L. Li and R. Chen, Small, 2021, 17, 2102893.
29. S. Liang, Z. Yu, T. Ma, H. Shi, Q. Wu, L. Ci, Y. Tong, J. Wang and Z. Xu, ACS Nano, 2021, 15, 14697–14708.
30. W. Bi, C. Li, D. Yang, Y.-Z. Zhang, L. Hu, Q. Gong, J. Zhang, Y. Zhang, M. Li, J. Wei, Y. Zhou, D. Zhou, T. Wu, L. F. Chen and A. Cabot, Energy Environ. Sci., 2025, 18, 1929–1940.
31. V. Shrivastav, S. Sundriyal, P. Goel, H. Kaur, S. K. Tuteja, K. Vikrant, K. H. Kim, U. K. Tiwari and A. Deep, Coord. Chem. Rev., 2019, 393, 48–78.
32. X. Xu, J. Liu, J. Liu, L. Ouyang, R. Hu, H. Wang, L. Yang and M. Zhu, Adv. Funct. Mater., 2018, 28, 1707573.
33. X. Yang, S. Wang, D. Y. W. Yu and A. L. Rogach, Nano Energy, 2019, 58, 392–398.
34. C. Li, T. Chen, W. Xu, X. Lou, L. Pan, Q. Chen and B. Hu, J. Mater. Chem. A, 2015, 3, 5585–5591.
35. T. Li, Y. Bai, Y. Wang, H. Xu and H. Jin, Coord. Chem. Rev., 2020, 410, 213221.
36. C. Liu, B. Wu, T. Liu, Y. Zhang, J. Cui, L. Huang, G. Tan, L. Zhang, Y. Su and F. Wu, J. Energy Chem., 2024, 89, 449–470.
37. Z. Cao, J. Cui, D. Yu, Y. Wang, J. Liu, J. Zhang, J. Yan, Y. Zhang, S. Sun and Y. Wu, Adv. Funct. Mater., 2023, 33, 2306862.
38. M. Huang, K. Mi, J. H. Zhang, H. L. Liu, T. T. Yu, A. H. Yuan, Q. H. Kong and S. L. Xiong, J. Mater. Chem. A, 2017, 5, 266–274.
39. T. Yang, M. Fang, J. Liu, D. Yang, Y. Liang, J. Zhong, Y. J. Yuan, Y. Zhang, X. Liu, R. Zheng, K. Davey, J. Zhang and Z. Guo, Adv. Funct. Mater., 2022, 32, 2205880.
40. Z. Q. Ye, Y. Jiang, L. Li, F. Wu and R. J. Chen, Adv. Mater., 2021, 33, 2101204.
41. W. Sun, C. Cai, X. Tang, L. P. Lv and Y. Wang, Chem. Eng. J., 2018, 351, 169–176.
42. T. Yang, Y. Liu, D. Yang, B. Deng, Z. Huang, C. D. Ling, H. Liu, G. Wang, Z. Guo and R. Zheng, Energy Storage Mater., 2019, 17, 374–384.
43. Z. Zhang, Y. Huang, X. Liu, X. Wang and P. Liu, ACS Sustain. Chem. Eng., 2020, 8, 8381–8390.
44. H. Li, Y. He, Q. Wang, S. Gu, L. Wang, J. Yu, G. Zhou and L. Xu, Adv. Energy Mater., 2023, 13, 2302901.
45. S. L. Zhang, H. J. Ying, P. F. Huang, J. L. Wang, Z. Zhang, T. T. Yang and W. Q. Han, ACS Nano, 2020, 14, 17665–17674.
46. F. Liu, Y. Liu, X. Zhao, K. Liu, H. Yin and L. Z. Fan, Small, 2020, 16, 1906076.
47. Z. Cao, Y.-B. Zhu, K. Chen, Q. Wang, Y. Li, X. Xing, J. Ru, L. G. Meng, J. Shu, N. Shpigel and L. F. Chen, Adv. Mater., 2024, 36, 2401271.
48. X. Lu, J. Cheng, D. Zhou, Y. Chen, H. Jiang, Y. Lu, D. Zhang, D. Kong, P. K. Chu, H. Y. Yang and Y. Luo, Chem. Eng. J., 2023, 478, 147366.
49. F. Ming, H. Liang, G. Huang, Z. Bayhan and H. N. Alshareef, Adv. Mater., 2021, 33, 2004039.
50. J. Wang, C.-F. Du, Y. Xue, X. Tan, J. Kang, Y. Gao, H. Yu and Q. Yan, Exploration, 2021, 1, 20210024.
51. A. Sarycheva and Y. Gogotsi, Chem. Mater., 2020, 32, 3480–3488.
52. C. Cao, H. Su, L. Ai, D. Lv, J. Gu, R. Li, D. Li, W. Zhang, M. Ge and X. Yao, Adv. Funct. Mater., 2024, 34, 2409586.
53. H. Su, L. Lu, M. Yang, F. Cai, W. Liu, M. Li, X. Hu, M. Ren, X. Zhang and Z. Zhou, Chem. Eng. J., 2022, 429, 132167.
54. W. W. Sun, C. Cai, X. X. Tang, L. P. Lv and Y. Wang, Chem. Eng. J., 2018, 351, 169–176.
55. J. Liu, L. Zhang and H. Wu, Adv. Funct. Mater., 2022, 32, 2200544.
56. K. Wang, Y. P. Wang, Y. F. Zhang, F. Liu, J. R. Shi, S. Y. Liu, X. F. Xie, G. Z. Cao and A. Q. Pan, Nanoscale, 2020, 12, 12623–12631.
57. P. Zhou, M. Y. Zhang, L. P. Wang, Q. Z. Huang, Z. A. Su, L. W. Li, X. D. Wang, Y. H. Li, C. Zeng and Z. H. Guo, Front. Chem., 2019, 7, 569.
58. J. Yang, H. Gao, S. Men, Z. Shi, Z. Lin, X. Kang and S. Chen, Adv. Sci, 2018, 5, 1800763.
59. L. Zhang, W. Liu, Q. Ma, Y. Xu, Z. Liu and G. Wang, ChemElectroChem, 2021, 8, 4047–4053.
60. X. B. Liu, Y. C. Liu, M. Feng and L. Z. Fan, J. Mater. Chem. A, 2018, 6, 23621–23627.
61. N. Yu, L. Zou, C. Li and K. Guo, Appl. Surf. Sci., 2019, 483, 85–90.
62. H. Liu, Y. Lin, M. Li, L. Zhang, Y. Lu and W. Xiao, Int. J. Hydrogen Energy, 2021, 46, 30818–30827.
63. Q. Xiao, Q. Song, K. Zheng, L. Zheng, Y. Zhu and Z. Chen, Nano Energy, 2022, 98, 107326.
64. C. Dong, Y. Zhou, W. Liu, W. Du, X. Zhang, X. Sun, L. Kang, X. Zhang and F. Jiang, J. Alloys Compd., 2021, 875, 159903.
65. J. Zhong, L. Qin, J. Li, Z. Yang, K. Yang and M. Zhang, Int. J. Miner., Metall. Mater., 2022, 29, 1061–1072.
66. D. Chen, G. Chen, J. Pei, Y. Hu, Z. Qin, J. Wang and F. Wu, ChemElectroChem, 2017, 4, 2158–2163.
67. Y. M. Lin, H. S. Fan, C. Z. Zhu and J. Xu, Rare Met., 2022, 41, 4104–4115.
68. R. Liu, W. Cao, D. Han, Y. Mo, H. Zeng, H. Yang and W. Li, J. Alloys Compd., 2019, 793, 505–511.
69. N. Xue, X. Li, M. Zhang, L. Han, Y. Liu and X. Tao, ACS Appl. Energy Mater., 2020, 3, 10234–10241.
70. C. Li, Z. Xue, J. Qin, M. Sawangphruk, P. Yu, X. Zhang and R. Liu, J. Alloys Compd., 2020, 842, 155812.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta01040c

‡ These authors contributed equally.

---