Metal nitride heterostructures capsulated in carbon nanospheres to accommodate lithium metal for constructing a stable composite anode

INTRODUCTION

The energy density of conventional lithium (Li)-ion batteries (LIBs) based on the graphite anode is hard to meet the requirements of the rapid development of automobiles and portable electronic devices^[1-4]. Li metal is the most promising anode for the next generation of high-energy batteries due to its ultrahigh capacity (3860 mA h g^-1) and low reduction potential (-3.04 V vs. the standard hydrogen electrode)^[5-8]. However, the critical problems, including uncontrolled Li dendrite formation^[9-12], infinite volume expansion^[13,14] and severe interfacial side reactions^[15-20], cause safety issues and poor stability of lithium metal batteries (LMBs). Thus far, numerous strategies have been proposed to solve the above problems, such as artificial interface construction^[21-23], host structure design^[24-26] and electrolyte modification engineering^[27-29]. Unfortunately, most of the above works are conducted under low-capacity loadings or with flooded electrolytes, far from the harsh practical conditions^[9,30].

Compared with the pure Li metal anode, the Li metal-based composite anode is more promising for applications because of its better structural stability and capability for mass production^[31-33]. Various materials have been used as hosts to accommodate the Li metal, including carbon-based hosts and three-dimensional metal foams^[24,33]. However, these hosts have poor interfacial compatibility with lithium^[34], and thus, many lithiophilic alloys such as metal nitrides and lithium compounds are used to modify the host surface to guide the uniform plating or infusion in the hosts. Nevertheless, these lithiophilic materials usually suffer the inevitable pulverization and agglomeration during long cycling because of the reaction with Li^[35,36]. Therefore, the structural integrity of lithiophilic components and host materials is important to prepare a stable Li-metal-based composite anode and realize the uniform Li plating and stripping.

Herein, we design and prepare a host of metal nitrides (Mo₂N and WN heterostructures) capsulated in the hollow carbon nanospheres (denoted WN/Mo₂N@HCN). A composite Li metal anode was prepared by die pressing and heating the Li metal powder and WN/Mo₂N@HCN host together [Supplementary Figure 1]. The lithiophilic Mo₂N regulates the preferential metallic Li deposition inner the hollow carbon nanospheres. In addition, the rigid WN, which is well composited with Mo₂N, helps maintain the integrity of Mo₂N in cycling, ensuring continuous and stable Li deposition. The composite Li metal anode showed stable plating/stripping over 1400 h with minimal voltage divergence (16 mV at 1 mA cm^-2 and 1 mA h cm^-2). The full battery assembled with NCM811 cathode also showed excellent long cycling stability even at high mass loading (8.6 mg cm^-2) and lean electrolyte, demonstrating the practicability of such composite anode for high-performance LMBs.

RESULTS AND DISCUSSION

The WN/Mo₂N@HCN host was obtained through micelle-interfacial copolymerization in the presence of soluble metal salts, followed by carbonization and nitridation^[37-39]. The copolymerization of organic monomers triggers the condensation of metal salts at micelle interfaces, which achieves one-pot integration of nanoparticle preparation, encapsulation and shell formation. The WN/Mo₂N@HCN host shows a well-defined spherical morphology and core-shell structure [Figure 1A and B]. In addition, this process also ensures the uniform growth of WN and Mo₂N together to form a heterostructure-like structure due to the similar properties of the used salt precursors and simultaneous nitriding. The cavity thickness in the HCN is about ~24 nm suggested by the high-resolution TEM image [Figure 1C]. The specific surface area calculated by the Brunauer-Emmett-Teller (BET) method and the pore volume of WN/Mo₂N@HCN are 206 m² g^-1 and 0.35 cm³ g^-1, respectively, which are much lower than those of HCN (569 m² g^-1 and 0.54 cm³ g^-1). The decrease in BET specific surface area and the disappearance of hysteresis loop suggest that the interior void of carbon is filled with the alloy nanoparticles after the incorporation of WN/Mo₂N heterostructures [Supplementary Figure 2 and Supplementary Table 1]. Furthermore, HAADF-scanning transmission electron microscopy (STEM) images and corresponding elemental mappings and diffraction peaks, as well as the X-ray diffraction pattern (XRD), suggest the formation of WN and Mo₂N and their uniform integration in the HCNs [Figure 1D and E], proving the formation of heterostructure-like structure between WN and Mo₂N. The contents of Mo and W are 12.13 wt% and 13.03 wt%, respectively, characterized by inductively coupled plasma-mass spectrometry (ICP-MS) [Supplementary Table 2]. The void space and metal nitrides guide the homogeneous Li deposition inner the HCN^[40,41]. Note that the Mo₂N is well composited with WN, and thus, the rigid WN can help avoid pulverization and agglomeration of Mo₂N, ensuring stable Li deposition in HCN.

Figure 1. Structural characterization of WN/Mo₂N@HCN and the obtained Li metal composite anode. (A) SEM image, (B and C) TEM images and (D) stacked HAADF and element mappings of WN/Mo₂N@HCN; (E) XRD patterns, (F and G) surface and cross-sectional SEM images and (H) AFM image with Young’s modulus mapping of the composite anode.

The Li metal composite anode (WN/Mo₂N@HCN/Li) was prepared by die pressing and heating the Li metal powder and WN/Mo₂N@HCN host together with different ratios (Unless specified, the mass ratio between the WN/Mo₂N@HCN and Li is 1:3), in which the molten Li uniformly infused into the host and formed compact stack with high bulk density [Supplementary Table 3]. After this process, the diffraction peaks of WN/Mo₂N@HCN were weakened. The morphologies of the composite anode are shown in Figure 1F and G and Supplementary Figure 3A and B, which show a similar thickness (around 360 μm) to that of the prepared Li metal foil anode by Li metal powder. The surface modulus of the composite anode and lithium metal foil was characterized by an atomic force microscope (AFM). As shown in Figure 1H, the average Young’s modulus of the composite anode reaches 86.0 GPa, whereas that of Li metal foil is only 10.2 GPa[Supplementary Figure 3C]. The increased mechanical strength should be ascribed to the ultrahigh hardness of metal nitrides and the carbon shell^[42], which can tolerate the stress generated by the Li plating and stripping and guarantee the integrity of the composite anode^[43]. The composite anode also shows a dense and smooth surface with the uniformly distributed WN/Mo₂N@HCN host in it, which can be confirmed by EDS elemental mapping images [Supplementary Figure 4].

The reaction between Li metal powder and WN/Mo₂N@HCN was investigated by X-ray photoelectron spectroscopy (XPS). The N 1s XPS peaks at 398.4 and 395.4 eV in the spectra of the composite anode [Supplementary Figure 5A] are assigned to pyridinic-N and lithium nitride formed by the partial lithiation reaction, which are not observed in that of Li metal foil [Supplementary Figure 5B]^[44,45]. In addition, the Li 1s XPS peak at 57.1 eV corresponds to Li-C species due to the reaction of Li and carbon [Supplementary Figure 5C and D]. The lithiophilic lithium nitride and Li-C can reduce the nucleation overpotential and induce uniform lithium deposition in cycling^[46,47]. A contact angle test was conducted to explore the interfacial wettability between the liquid electrolyte and composite anode. As shown in Supplementary Figure 6, the initial contact angle was 27.7° and the droplet spread out in 10 s, but for the lithium metal foil, the contact angle was 58.6°, suggesting the improved interfacial wettability of the composite anode with the above components^[48].

In order to investigate the electrochemical stability of the composite anode, the galvanostatic stripping/plating voltage profiles of symmetric cells with ether-based electrolyte were tested, as shown in Figure 2A. The composite anode exhibits stable cycling performance for more than 1400 h with an ultralow voltage hysteresis (~16 mV) without distinct voltage fluctuation at 1 mA cm^-2 and 1 mA h cm^-2. In contrast, the composite anodes of WN@HCN/Li and Mo₂N@HCN/Li show enlarged overpotential after 600 h and 700 h, which should be ascribed to the inferior lithiophility of WN and the pulverization of Mo₂N in long cycling. Note that the HCN/Li and Li metal foil anode show distinct voltage fluctuation in short cycling due to the nonuniform Li deposition, which induces uncontrollable dendrite growth. The comparison of semicircles at the high-frequency region of Nyquist plots [Figure 2B] also shows that the composite anode has the smallest electrochemical resistance due to improved ion diffusion and interfacial reaction kinetics, which is benefited from the introduction of WN/Mo₂N@HCN. It is shown that Mo₂N@HCN/Li also has a similar initial overpotential with the composite anode during the Li plating [Figure 2C], which is much smaller than those of other anodes, indicating the function of Mo₂N in lowering the nucleation barrier^[49]. Mo₂N can serve as the preferred nucleation site for Li deposition, which can effectively reduce nucleation barriers, where Li can be deposited on Mo in an epitaxial manner with the lattice strain accommodated at the interface, leading to stronger Li-Mo bond than the Li-Li bond^[41]. When the mass ratio of WN/Mo₂N@HCN and lithium powder was adjusted to 1:2 and 1:4, the cycling performance became worse than that with the ratio of 1:3 [Supplementary Figure 7]. No cracks and dead Li can be observed on the composite anode surface after 200 h stripping/plating [Figure 2D and E]. Even at a high current density of 2 mA cm^-2 and a high capacity of 2 mA h cm^-2, the cells with the composite anode still maintain excellent cycling stability for 900 cycles [Figure 2F and G].

Figure 2. Electrochemical performance of the different anodes in symmetrical cells. (A) long-term plating/stripping profiles of composite anodes with the different hosts in symmetrical cells at 1 mA cm^-2 and 1 mA h cm^-2 and (B) the corresponding Nyquist plots; (C) overpotential profiles during the first Li plating cycle; SEM images of (D) WN/Mo₂N@HCN/Li and (E) pure Li metal anode after 200 h cycles; (F) long-term plating/stripping profiles of WN/Mo₂N@HCN/Li and pure Li metal anode in symmetrical cells at 2 mA cm^-2 and 2 mA h cm^-2 and (G) the corresponding partial enlargement at 500-510 h.

The morphology of the composite and pure Li metal anodes was characterized to reveal the different Li stripping/plating behavior at the current density of 1 mA cm^-2. With the deposition capacity of 1 mA h cm^-2, numerous disk-shaped Li pieces were evenly distributed on the composite anode surface, which finally grew into a dense and compact layer without any Li dendrite after the capacity increased to 3 mA h cm^-2[Figure 3A and B]. In sharp contrast, unevenly distributed needle-like Li dendrites were formed on the Li metal foil anode surface under 1 mA h cm^-2 to 3 mA h cm^-2 [Figure 3C and D]. Furthermore, the AFM image in Supplementary Figure 8 shows a smooth deposition morphology with a minor surface roughness on the composite anode surface compared with that on Li metal foil anode, in agreement with the SEM images. In order to reveal the mechanism of uniform lithium deposition on different surfaces, the surface potential distributions of anodes were measured by the Kelvin probe force microscopy (KPFM), as shown in Figure 3E and F. It is seen that WN/Mo₂N@HCN/Li has a well-distributed surface potential with the roughness of 35.7 mV, much lower than that of the Li metal foil (70.7 mV) [Figure 3G and H], which is a benefit to uniform Li deposition, as illustrated in Figure 3I.

Figure 3. Lithium deposition behavior of different anodes. Lithium deposition morphology at 1 mA cm^-2 for (A) 1 h and (B) 3 h for the composite anode, (C) 1 h and (D) 3 h for pure Li metal anode; surface potential mappings of (E) the composite anode and (F) pure Li metal anode through Kelvin probe force microscopy (KPFM) and corresponding (G and H) statistical distribution; (I) schematic illustration of Li deposition process on these two anodes.

The smooth and flat surface of the composite anode was shown after 100 cycles in the full cell, but the surface of Li metal foil anode showed many cracks [Figure 4A and B]. The above results clearly prove that the lithiophilic metal nitride nanoparticles effectively guide homogeneous Li stripping/plating and the carbon shells well buffer volume changes to release the stress^[22], which strengthens the structural stability of the composite anode during cycling. The compositions of the SEI formed on the composite and Li metal foil anodes in the carbonate electrolyte were analyzed by XPS [Figure 4C-F]. In the C 1s spectrum, the peaks at 284.4, 286.4, 288.1 and 290.2 eV correspond to the C-C/C=C, COR, CO₃^2- and C-F species, respectively^[50]. The disappearance of Li₂CO₃ peak in the composite anode indicates that the side reaction was suppressed significantly benefited from improved interfacial kinetics. The peaks at 687.6 and 684.7 eV in the F 1s spectrum are assigned to C-F and LiF^[51]. The results show more LiF contained in the SEI formed on the composite anode surface due to the existence of lithium nitride in the composite anode, which can regulate lithium deposition behavior by boosting rapid ion transport. Such differences are also suggested by the electrochemical impedance spectroscopy [Figure 4G and H]. The interfacial impedances of Li metal foil anode not only fluctuated wildly at different cycles in the Nyquist plots, but also were larger than that of the composite anode. The apparent activation energy of Li-ions migration across SEI film on the interface was also calculated by the electrochemical impedance spectroscopy measurements under different temperatures^[52]. As shown in Figure 4I, the E_SEI of the composite anode (52.95 kJ mol^-1) is obviously lower than that of Li metal foil anode (60.38 kJ mol^-1), suggesting the improved dynamics of Li-ion migration through SEI for the uniform Li deposition^[53].

Figure 4. Structural stability and SEI composition of different anodes. SEM images of surface topography for (A) composite anode and (B) pure Li metal anode after cycling; high-resolution XPS analysis of (C and D) C 1s and (E and F) F 1s spectra for the composite anode and pure Li metal anode after cycling; (G and H) Nyquist plots of NCM811 full cells with the above two anodes after different cycles; (I) the E_a values calculated by the electrochemical impedance spectroscopy measurements under different temperatures.

The full battery was assembled with NCM811 cathode to show the practicability of the composite anode. The full cell using the composite anode delivers a higher initial specific capacity of 182.4 mA h g^-1 and superior cycling performance with a capacity retention of 61.5% after 400 cycles, which are obviously higher than that of other composite anodes due to the synergistic effect of Mo₂N and WN heterostructures [Supplementary Figure 9], where the lithiophilic Mo₂N induces uniform Li deposition and the rigid WN matrix helps to suppress the pulverization and agglomeration of Mo₂N. Rapid decay occurred with the Mo₂N@HCN/Li after 300 cycles, while WN@HCN/Li and HCN/Li also delivered low capacity retention rates (51.2% and 40.9%, respectively). However, for the cell with the Li metal foil anode, the initial specific capacity and capacity retention after 400 cycles are only 162.9 mA h g^-1 and 39.5% [Figure 5A]. The coulombic efficiency of the cell with Li metal foil anode decays to 91.7% after cycling, which is much lower than that with the composite anode (99.1%), suggesting good reversibility and high lithium utilization with the WN/Mo₂N@HCN host.

Figure 5. Electrochemical performance of NCM811 full cells with different anodes. (A) Cycling performance, (B) rate performance, (C and D) galvanostatic discharge-charge curves and (E) Nyquist plots of the NCM811 full cells with composite and pure Li metal anodes; (F) Tafel curves of WN/Mo₂N@HCN/Li and pure Li metal anodes; cycling performance of the composite anode and pure Li metal anode at (G) high cathode loading and (H) lean electrolyte amount.

The rate performances of the above two full cells were also tested from 0.2 C to 10 C, as shown in Figure 5B-D. The cell with the composite anode shows specific capacities of 203.2, 194.7, 186.9, 177.2, 158.5, and 138.5 mA h g^-1 at 0.2, 0.5,1, 2, 5, and 10 C, respectively, which are higher than the cell with Li metal foil anode. The improved rate performance is mainly ascribed to reduced electrochemical resistance indicated by the Nyquist plots [Figure 5E], where fitting results show that the composite anode has a much smaller charge transfer resistance (41.21 Ω) than that of pure Li metal anode (222.70 Ω) [Supplementary Figure 10 and Supplementary Table 4]. Furthermore, the exchange current density derived from the Tafel plots [Figure 5F] also confirms the fast interfacial reaction kinetics^[54,55], where the cell with the composite anode shows about four times higher exchange current density than that with Li metal foil anode (1.181 mA cm^-2vs.0.293 mA cm^-2). Besides, the cell with the composite anode also demonstrates stable cycling performance under high-loading cathode (8.6 mg cm^-2) and lean electrolyte (E/C ratio of 2 μL mg^-1), revealing high capacity retention of 88.2% and 77.1% after 150 cycles, respectively [Figure 5G and H]. In contrast, rapid capacity decay followed by the failure of cells occurred under the same test conditions for the cells with Li metal foil anode, which should be attributed to active lithium loss and electrolyte consumption resulting from uncontrollable lithium dendrites and severe interfacial side reactions.

CONCLUSIONS

In summary, a host of metal nitrides (Mo₂N and WN heterostructure) capsulated in the hollow carbon nanospheres was developed, which can accommodate Li metal to form a composite anode with excellent stability. The lithiophilic Mo₂N guides the uniform infusion and the electrochemical plating of Li in HCN. At the same time, WN, which is well composited with Mo₂N, acts as a rigid matrix to conquer the pulverization of Mo₂N during the Li plating/stripping, enhancing the structural stability of the host and regulated deposition during long cycling. The obtained composite anode thus shows excellent electrochemical performance and stability. The symmetric battery can work stably for more than 1400 h at 1 mA cm^-2 with a capacity of 1 mA h cm^-2. Furthermore, the assembled full battery with NCM811 cathode maintains a capacity of 112.2 mA h g^-1 after 400 cycles with a high coulombic efficiency of 99.1%. Even with a high-loading cathode and lean electrolyte conditions, the full batteries still show good cycling stability. This work provides a simple way to prepare a stable Li metal-based composite anode for lithium metal batteries.