TAO Fang-yu, XIE Dan, DIAO Wan-yue, LIU Chang, SUN Hai-zhu,LI Wen-liang,, ZHANG Jing-ping,, WU Xing-long,2,

（1. Faculty of Chemistry, National & Local United Engineering Lab for Power Battery, Northeast Normal University, Changchun, Jilin 130024, China;2. MOE Key Laboratory for UV Light-Emitting Materials and Technology, Northeast Normal University, Changchun, Jilin 130024, China）

Abstract： We report the fabrication of a lithiophilic Ti3C2Tx MXene-modified carbon foam (Ti3C2Tx-MX@CF) for the production of highly-stable LMBs that regulates Li nucleation behavior and reduces the volume change of a lithium metal anode (LMA).The 3D CF skeleton with a high specific surface area not only reduces the local current density to avoiding concentrated polarization,but also provides enough space to absorb the volume expansion during cycling. The excellent lithiophilicity of Ti3C2Tx-MX produced by its abundant functional groups reduces the Li nucleation overpotential, guides uniform Li deposition without the formation of Li dendrites, and maintains a stable SEI on the anode surface. Consequently, a Li infiltrated Ti3C2Tx-MX@CF symmetrical cell has an excellent cycling stability for more than 2 400 h with a low overpotential of 9 mV at a current density of 4 mA cm-2 and has a capacity of 1 mA h cm-2. Furthermore, a Li- Ti3C2Tx-MX@CF||NCM111 full cell has a capacity of 129.6 mA h g-1 even after 330 cycles at 1 C, demonstrating the advantage of this method in constructing stable LMAs.

Key words: MXene；Three-dimensional structure；Dendrite suppression；Li metal anode

1 Introduction

Lithium metal batteries (LMBs) have been the subject of great research efforts due to the fascinating advantages of metallic Li, which include the lowest redox potential (-3.04 Vvs. the standard hydrogen electrode) and the highest gravimetric theoretical capacity (3 860 mAh g-1)[1-4]. Unfortunately, the uncontrollable Li dendrite growth and infinite volume change during Li deposition/stripping processes have greatly hindered the practical application of LMBs[5-7].The dendritic growths caused by the uneven Li+flux,easily pierce the separator and cause battery failure,often posing serious threats to safety[8-9]. The huge volume fluctuation is closely related to the inherent“hostless” property of metallic Li, which causes the solid electrolyte interface (SEI) film of lithium metal anodes (LMAs) to fracture, thereby inducing the reconstruction of SEI film, lowering the reversibility and limiting cycle lifespan of LMAs[10].

Recently, various strategies have been proposed to solve the above-mentioned problems, mainly including the optimization of electrolyte component[11-13], construction of a stable artificial interfacial layer on the surface of LMAs[14-15], and rational design of solid-state electrolyte[16-18]. Although modification of SEI film is able to enhance the stability of LMAs and improve electrochemical performance of LMBs to some extent, infinite volume expansion of LMAs arising from th1 “hostless” nature remains unresolved[19-20]. To this end, constructing a three-dimensional (3D) host has been viewed as the one of the most promising methods to resolve this issue[21]. This is attributed to the elaborately designed 3D skeleton(such as Cu, Ni and carbon foams) with high specific surface area and sufficient space, which not only lowers the local current density for homogenizing the distribution of Li+flux, but also accommodates the deposited Li metal for alleviating volume variations during repeated cycling[22-25]. Among them, accommodating deposited metallic Li into the 3D carbon skeleton with high electronic conductivity, flexibility and lightweight, is an attractive option[26]. The 3D carbon skeleton promotes fast Li deposition kinetics and guarantees high energy density of battery[27]. However, ordinary 3D carbon skeleton have a poor affinity for Li metal, which can result in high nucleation overpotential, triggering uneven Li deposition and Li dendrites growth[6]. Therefore, it is necessary to introduce lithiophilic materials into the 3D carbon skeleton to improve the overall lithiophilicity and facilitate uniform Li nucleation and growth[28].

It is well-accepted that introducing inorganic compounds with abundant functional groups can significantly improve the lithiophilicity of 3D carbon skeleton because of their strong interaction with metallic Li. For instance, two-dimensional (2D) transition metal carbides, nitrides and carbonitrides, also known as MXenes, inherently possess abundant functional groups, high electrical conductivity and low ion diffusion barrier, and have been confirmed to convert the 3D carbon skeleton from the lithiophobic to lithiophilic, thereby inhibiting Li dendrites growth[29-32].More interestingly, MXenes, as a typical 2D layered material, can form lightweight freestanding film through self-assembly method[33-34]. Unfortunately, the formed freestanding MXene film is composed of densely compacted nanosheets with limited space, that is fragile and can hardly withstand the volume change during Li deposition/stripping processes, inevitably leading to the battery failures. Consequently, it is of great significance to integrate the advantages of 3D carbon skeleton and MXenes to construct a 3D lithiophilic host for mitigating volume change and simultaneously guiding uniform Li deposition for realizing highly-stable and dendrite-free LMAs.

Herein, a lithiophilic Ti3C2Tx-MXene decorated carbon foam (Ti3C2Tx-MX@CF) is prepared via a simple wetting process for constructing stable and dendrite-free LMAs. In this structure, the obtained Ti3C2Tx-MX@CF anode presents several advantages:(1) Thanks to their good affinity with metallic Li, —O and —F groups of Ti3C2Tx-MX can serve as nucleation sites to guide the homogeneous Li nucleation and growth, thus suppressing the formation of Li dendrites. (2) The 3D Ti3C2Tx-MX@CF skeleton possesses a large specific surface area, which decreases the local current density and ensures uniform Li+flux.(3) The 3D Ti3C2Tx-MX@CF framework consisting of interconnected networks provides enough space for accommodating deposited Li and mitigating volume fluctuation of LMAs during the repeated cycling. As a result, the Ti3C2Tx-MX@CF-based symmetrical cell achieves a long cycling life up to 2 400 h at a high current density of 4 mA cm-2with an areal capacity of 1 mAh cm-2. Meanwhile, under the higher areal deposition capacity of 10 mAh cm-2, the Ti3C2Tx-MX@CF symmetrical cell still realizes stable cycling for more than 2 000 h. Upon pairing Li-Ti3C2Tx-MX@CF with LiNiCoMnO2, the assembled full batteries show improved cycling and rate performance,indicating the potential of Ti3C2Tx-MX@CF in constructing high-performance LMBs.

2 Experimental

2.1 Fabrication of Ti3C2Tx-MX

Ti3C2Tx-MX was synthesized based on previous reports[30]. First, 2 g of LiF was added to 40 mL of 9 mol L-1HCl solution, followed by stirring for 30 min. Then, 2 g Ti3AlC2powder was slowly added into the above solution and stirred for 24 h at room temperature. Subsequently, the acidic mixture was repeatedly washed with deionized water and centrifuged at 8 000 r min-1until the pH was ≈ 7. Then, the resulting solution was sonicated for 3 h. Finally, the suspension was poured out and centrifuged at 3 500 r min-1for 5 min to obtain a dark green supernatant(7 mg mL-1).

2.2 Fabrication of CF

The CF was obtained through one-step carbonization of the discarded facial mask tissue at 500 °C for 4 h under the Ar atmosphere.

2.3 Fabrication of Ti3C2Tx-MX@CF

The discarded facial mask tissue was first cut into desirable shapes and sizes, and then the Ti3C2Tx-MX was added to it dropwise. The obtained composite material was then dried overnight in an oven, and the Ti3C2Tx-MX@CF mixture was obtained by carbonization at 500 °C for 4 h in an Ar atmosphere.

2.4 Materials characterization

The morphology and crystal structures of materials were characterized by scanning electron microscopy (SEM, XL 30 ESEM-FEG, FEI Co.) and X-ray diffraction (XRD) patterns using a Rigaku P/max 2200VPC diffractometer with Cu K α radiation. The element mapping of samples was performed by using the EDX module equipped with the SEM.

To observe the surface morphology evolution of the electrodes during the Li plating/stripping process,the coin CR2032 cells were disassembled and washed using 1,2-dimethoxyethane (DME, Sigma-Aldrich,99.8%) several times to remove the electrolyte in an argon-filled glove box with concentrations of H2O and O2below 0.1×10-7. Subsequently, the samples were quickly transferred to the vacuum chamber for SEM testing.

2.5 Electrochemical measurement

All the coin cells were assembled in an argonfilled glovebox. The galvanostatic charge-discharge experiments were carried out using the Neware battery testing system at room temperature. For half cells,Li||Ti3C2Tx-MX@CF and Li||CF cells were assembled with Li foil as the counter electrode, and Ti3C2Tx-MX@CF or CF electrode as the working electrode to investigate the reversibility of Li plating/stripping.Half cells were tested at 0.5 mA cm-2with the plating capacity of 0.5 mAh cm-2and 1 mAh cm-2at 1 mA cm-2. For symmetrical cells, Li foil and Ti3C2Tx-MX@CF or CF electrode as counter electrode and working electrode to assemble a symmetrical cell, and then 10 mAh cm-2of Li was electrodeposited onto the Ti3C2Tx-MX@CF or CF electrode at a current density of 0.5 mA cm-2. The cycling stability of symmetrical cells was studied at a fixed current density of 4 mA cm-2and a capacity of 1 or 10 mA h cm-2. Glass microfiber filter film was used as the separator and the 100 μL ether mixture consisting of 1 mol L-1LiTFSI(Malklin, 99%) solution in 1∶1 (v/v) DOL (SigmaAldrich, 99%) and DME (Sigma-Aldrich, 99.8%)with 0.5 mol L-1LiNO3(Aladdin, 99%) was used as the electrolyte. For full cells, the cathode electrode was fabricated by coating the cathode slurry (composed of LiNiCoMnO2(NCM111) powder∶PVDF∶super P = 7∶2∶1 using N-methyl-2-pyrrolidone(NMP) as the solvent) on Al foil and then drying in vacuum overnight at 100 °C. The loading amount of the NCM111 electrode was about 1.0-1.2 mg cm-2.The 1 mol L-1LiPF6in EC/DEC (v/v = 1/1) was used as the electrolyte. The pre-cycled anode was obtained by disassembling cells to get working electrodes that had been deposited 5 mA h cm-2Li on Ti3C2Tx-MX@CF or CF. The cycle life and rate performance of full cells were tested by galvanostatic charge/discharge at 1 C and 5 C (1 C = 150 mAh g-1) and the voltage window was set to 2.0-4.3 V.

3 Results and discussion

3.1 Characterization of Ti3C2Tx-MX@CF

The Ti3C2Tx-MX was successfully fabricated through a convenient acid etching method. As displayed in Fig. S1a and 1b, Ti3AlC2precursor presents a typical layered structure with a large number of small irregular granules, while Ti3C2Tx-MX shows an obvious multilayer structure with expanded interlayer spacing after selectively etching th1 Al layers(Fig. S1c and 1d). The corresponding X-ray diffraction (XRD) patterns were employed to reflect the chemical composition changes before and after the etching process of the Ti3C2Tx-MX (Fig. 1d). As shown, the diffraction peak of (104) plane in Ti3C2Tx-MX disappeared, representing the successful etching of Al layers. Moreover, the characteristic peak of(002) plane shifts from 9.44° (Ti3AlC2) to 8.73°(Ti3C2Tx-MX) after etching, indicating the expansion of interlayer spacing and demonstrating the successful preparation o1 Ti3C2Tx-MX. Subsequently,Ti3C2Tx-MX suspension was added dropwise on the surface of non-woven fabric composed of parallelly aligned fibers (Fig. 1a and Fig. S2a) and then subjected to annealing in a furnace to prepare the Ti3C2Tx-MX@CF. As demonstrated, MXene was firmly and uniformly coated on the CF after the full infiltration and carbonization without any aggregation (Fig. 1b and Fig. S2b). The corresponding energy dispersive X-ray spectrometer (EDS) elemental mapping characterization indicates the uniform distribution of C, Ti, O and F elements on the CF surface and the content of elements is 8.87%, 35.92%, 45.97% and 10.08%, respectively, which can provide enough nucleation sites and effectively reduce the nucleation overpotential of metallic Li through strong interaction between metallic Li and heteroatoms (Fig. 1c and e).

Fig. 1 SEM images of (a) bare CF and (b) Ti3C2Tx-MX@CF. (c) Elemental mapping of Ti3C2Tx-MX@CF. (d) XRD of Ti3AlC2 precursor and Ti3C2Tx-MX.(e) Element contents obtained by XRD measurement in Ti3C2Tx-MX@CF

3.2 Electrochemical behaviors o1 Ti3C2Tx-MX@CF

To investigate the effect of the Ti3C2Tx-MX@CF composites on the Li deposition/stripping behavior and constructing dendrite-free LMAs, the surface morphology evolution and thickness change of the CF and Ti3C2Tx-MX@CF electrode was observed via exsitu SEM and cross-section images at 0.5 mA cm-2with various Li plating/stripping capacities. Fig. 2a shows the plating and stripping voltage profile of the Ti3C2Tx-MX@CF electrode. As shown in Fig. 2b and 2c, when deposition capacity is 2 mAh cm-2, Li metal is uniformly nucleated and grown on the wrinkled Ti3C2Tx-MX of CF surface benefited from the strong Li affinity of Ti3C2Tx-MX. Meanwhile, the thickness of Ti3C2Tx-MX@CF electrode exhibits a negligible change in comparison to that of the original state(255.65vs. 255.25 μm, Fig. S4a and Fig. 5a), suggesting that the Li deposition process tends to start from the lithiophilic Ti3C2Tx-MX@CF surface. As the plating capacity increases to 4 and 6 mAh cm-2, the Ti3C2Tx-MX is fully wrapped with deposited Li, and the edges o1 Ti3C2Tx-MX@CF become plump(Fig. 2d-g). However, the Ti3C2Tx-MX@CF surface always remains flat without the formation of Li dendrites during the entire Li deposition processes, demonstrating that the introduction of Ti3C2Tx-MX can offer numerous lithiophilic sites to guide homogeneous Li deposition. Moreover, the thickness of the Ti3C2Tx-MX@CF electrode varies from 256.70 to 257.49 μm,meaning that the Ti3C2Tx-MX@CF with interconnected fiber structure provides enough space for accommodating the deposited Li metal. During the subsequent charging process, the parallel-aligned Ti3C2Tx-MX@CF architecture gradually becomes clear and the wrinkle1 Ti3C2Tx-MX reemerges(Fig. 2h-m). It is worth noting that the surface of the Ti3C2Tx-MX@CF skeleton is smooth without residual Li metal, and the thickness of the electrode is close to the initial state after charging to 0.5 V (Fig. S4f), further demonstrating the effectiveness of the Ti3C2Tx-MX@CF in suppressing Li dendrites and alleviating the volume change during the Li plating/stripping processes. As for the CF, owing to the interconnected network of CF accommodating the deposited metal Li and buffering the volume change to some extent, the thickness of the CF electrode was slightly increased from 254.64 to 258.10 μm (Fig. S4g-i). Nevertheless,the deposited Li metal on the CF surface tends to aggregate into large chunks and then grows into filamentary Li dendrites due to the lack of nucleation sites on the CF surface (Fig. S3a-f). After the Li was stripped, a large number of the “dead Li” and dendritic Li is left behind the surface of the bare CF skeleton(Fig. S3g-l and Fig. S4l). These results indicate that Ti3C2Tx-MX@CF can be expected to improve the reversibility of LMAs and realize highly-stable LMBs.

Fig. 2 (a) Voltage profile of the Ti3C2Tx-MX@CF electrode at a current density of 0.5 mA cm-2 and capacity of 6 mAh cm-2. Morphology evolution of the Ti3C2Tx-MX@CF electrodes during the initial cycle of the Li plating/stripping process: After the anode was plated with (b, c) 1 mAh cm-2, (d, e) 2 mAh cm-2,and (f, g) 5 mAh cm-2 of Li metal; after stripping Li of (h, i) 2 mAh cm-2, (j, k) 4 mAh cm-2 and (l, m) charged to 0.5 V from the Ti3C2Tx-MX@CF anode

Symmetric cells were assembled to evaluate the stability of Li-Ti3C2Tx-MX@CF during long-term cycling. Initially, 5 mAh cm-2of Li is deposited onto the Ti3C2Tx-MX@CF to construct the Li-Ti3C2Tx-MX@CF electrode. Meanwhile, 5 mAh cm-2of Li is pre-deposited onto the CF for comparison. As shown in Fig. 3a, the symmetric Li-Ti3C2Tx-MX@CF cell presents a steady voltage hysteresis with a low overpotential of 9 mV after 2 400 h at 4 mA cm-2with an areal capacity of 1 mAh cm-2. On the contrary, the voltage hysteresis of the symmetric Li-CF cell show obvious fluctuations and sudden increases in voltages caused by the dendrite growth and accumulation of“dead Li”, indicating the failure of the cells. This result is further verified in the voltage hysteresis of Li charge/discharge with Li-Ti3C2Tx-MX@CF and Li-CF electrodes (Fig. 3b). When the Li deposition capacity increased to 10 mAh cm-2(Fig. 3c), the Li-Ti3C2Tx-MX@CF electrode continued to show stable cycle for 2 200 h with low overpotential of 22 mV. In sharp contrast, the symmetric battery with Li-CF presents a drastic voltage fluctuation and the overpotential increases to 2 V after 88 h. Besides, Table S1 compares the cycling performance of Ti3C2Tx-MX@CF with that of the previously reported LMAs based on 3D carbon frameworks. It is clearly seen that the Ti3C2Tx-MX@CF electrode exhibits superior cycling stability with the lowest voltage hysteresis, demonstrating its application potential for advanced dendrite-free LMAs. The significantly improved cycling performance of Li-Ti3C2Tx-MX@CF is mainly attributed to the Ti3C2Tx-MX@CF with enriched lithiophilic sites and high electronic conductivity lowering the nucleation overpotential and promoting Li deposition kinetics, thus directing uniform Li deposition behaviors and inhibiting Li dendrites formation. Subsequently,the morphology variation of the electrodes after cycling was observed by theex-situSEM images. As shown in Fig. S6a and S6b, bulk and loose “dead Li”with pores and cracks is accumulated on the surface of the CF. By contrast, a smooth and dendrite-free surface is observed in the Ti3C2Tx-MX@CF skeleton(Fig. S6c and d), implying that the Ti3C2Tx-MX@CF electrode induces uniform Li metal depositions and suppresses dendrites growth. Fig. 3d shows the rate performance of symmetric cells at various current densities ranging from 0.5 to 4 mA cm-2at a constant capacity of 1 mAh cm-2. With the current density increasing from 0.5 to 4 mA cm-2, the Li-Ti3C2Tx-MX@CF symmetric cells deliver gradually elevated overpotential of 10, 14, 19, 24 and 28 mV respectively. However, the Li-CF symmetric cell shows much larger voltage hysteresis compared with Li-Ti3C2Tx-MX@CF, which is attributed to the lack of lithiophilic sites of CF. This causes the uneven nucleation of Li metal, thereby resulting in the generation of Li dendrites and “dead Li” during the long cycles.Impressively, when the current density changes back from 4 to 0.5 mA cm-2, the overpotential of the Li-Ti3C2Tx-MX@CF still returns to 7 mV, indicating fast ion/electron transport kinetics and excellent structural stability of Li-Ti3C2Tx-MX@CF composite anode.The advantages of the Li-Ti3C2Tx-MX@CF anode in terms of interfacial stability were proven by electrochemical impedance spectroscopy (EIS) analysis.Fig. 3e and 3f display the Nyquist plots of the impedance spectra and the corresponding equivalent circuit diagram, and the simulated impedance parameters are listed in Table S2 (Supporting Information). It is observed that theRSEIandRctof the Li-Ti3C2Tx-MX@CF cell gradually reduce within 100 cycles, implying a gradual stabilization process. Besides, the Li-Ti3C2Tx-MX@CF electrode shows lower SEI resistance (RSEI)and charge transfer resistance (Rct) compared with that of the Li-CF electrode at a different number of cycles,suggesting the formation of a stable SEI layer and improved charge-transfer kinetics in th1 Ti3C2Tx-MX@CF host[35].

Fig. 3 Voltage profiles of Li-CF and Li-Ti3C2Tx-MX@CF electrodes in symmetric cells at (a) 4 mA cm-2 with a plating/stripping capacity of 1 mAh cm-2 and(b) the corresponding voltage hysteresis. (c) Voltage profile of Li-CF and Li-Ti3C2Tx-MX@CF symmetric cells under a large plating/stripping capacity of 10 mAh cm-2. (d) Rate performances of Li-CF and Li-Ti3C2Tx-MX@CF symmetric cells. (e, f) Electrochemical impedance spectroscopy of Li-CF and Li-Ti3C2Tx-MX@CF symmetric cells at different cycles

The coulombic efficiency (CE) is a crucial parameter for evaluating the sustainability of composite Li anode during the cycling processes, which is defined as the ratio of stripped capacity to plated capacity[36].The CE was studied by assembling asymmetric cells with Ti3C2Tx-MX@CF and bare CF as the working electrodes and Li foil as the counter electrode in both cells. The CE of CF anode shows an obvious overcharge phenomenon after 110 cycles under a plating capacity of 0.5 mAh cm-2at the current density of 0.5 mA cm-2, which is the result of the reestablishment of electrical connection of “dead Li” during the followin1 Li deposition processes[37]. By contrast,Li||Ti3C2Tx-MX@CF cell exhibits stable cycling over 120 cycles with a high average CE of 99.5% (Fig. 4a),benefiting from the Ti3C2Tx-MX@CF with abundant lithiophilic sites inducing a homogeneous Li deposition, maintaining a stable SEI layer, mitigating the volume change of LMAs during repeated cycling. As the deposition capacity increased to 1.0 mAh cm-2with the same current density (Fig. 4b), the Ti3C2Tx-MX@CF continues to display a stable CE for more than 120 cycles with an average CE as high as 99.8%.In the case of CF, an obvious decline of CE is observed after merely 30 cycles resulting from uneven Li deposition and dendritic Li growth caused by the lithiophobic nature of CF. These results indicate that the Ti3C2Tx-MX decorated CF host can significantly enhance the reversibility during the Li plating/stripping process due to the excellent lithiophilicity of Ti3C2Tx-MX. Besides, Li||Ti3C2Tx-MX@CF cell exhibits a smaller nucleation overpotential (20 mV) at 0.5 mA cm-2in comparison with Li||CF cells (24.2 mV), attributed to a large number of nucleation sites provided by Ti3C2Tx-MX which increase the affinity with Li and reduce the nucleation energy barrier(Fig. 4c). The surface morphology of Li plating/stripping for 50 cycles is displayed in Fig. 4d-g. Massive Li dendrites and “dead Li” are accumulated on the bare CF surface, while the Ti3C2Tx-MX@CF surface remains flat and smooth, further confirming the positive effect of designed Ti3C2Tx-MX@CF on inhibiting Li dendrites formation and enhancing the interfacial stability between electrode and electrolyte.

Fig. 4 Comparative coulombic efficiencies of Li plating/stripping on Ti3C2Tx-MX@CF and CF at (a) 0.5 mA cm-2, 0.5 mAh cm-2 and (b) 1 mA cm-2,1 mAh cm-2. (c) Nucleation overpotential of Li on the Ti3C2Tx-MX@CF and CF electrode. SEM images of (d, e) Ti3C2Tx-MX@CF and(g, e) CF electrode after the 50th cycle in the charge state

To demonstrate the feasibility o1 Ti3C2Tx-MX@CF in practical application o1 LMBs,5 mAh cm-2of Li metal was electrodeposited on Ti3C2Tx-MX@CF to fabricate th1 Li-Ti3C2Tx-MX@CF anode, and then matched with LiNiCoMnO2(NCM111) cathode to assemble the full cell. Fig. 5a shows the rate performance of Li-Ti3C2Tx-MX@CF||NCM111 and Li-CF||NCM111 full cell. As expected,the Li-Ti3C2Tx-MX@CF||NCM111 full cell exhibits the discharge capacities of 202.4, 186.2, 180.1, 174.8,171.9, 162.2, 149.2, 141.1, 132.7 and 124.7 mAh g-1at the current densities of 0.1, 0.2, 0.3, 0.4, 0.5, 1, 2, 3,4 and 5 C, respectively. More interestingly, when the current density is switched back to 0.1 C, the discharge capacity of full cell remains close to the original value, indicative of the excellent reversibility of Li-Ti3C2Tx-MX@CF aided by Ti3C2Tx-MX. In contrast,the Li-CF||NCM111 full cell exhibits a relatively high discharge capacity before 0.5 C rate, but dropped sharply from 1 to 5 C due to the generation of Li dendrite growth and “dead Li”. The long-term cycling stability of these batteries is tested at the current density of 1 C shown in Fig. 5b. As expected, the full cell assembled with the Li-Ti3C2Tx-MX@CF anode displays an initial specific capacity of 168.8 mAh g-1and retains 129.6 mAh g-1after 330 cycles (capacity retention of 76.8%). For Li-CF anode, an initial specific capacity of 157.5 mAh g-1is obtained and delivers a capacity of 69.7 mAh g-1after 330 cycles. Further elevating the current density to 5 C, the Li-Ti3C2Tx-MX@CF||NCM111 full cell still maintains a high discharge capacity of 101.5 mAh g-1after cycling for 500 cycles, corresponding to an ultralow capacity fade of 0.14% per cycle, which is much better than that of Li-CF||NCM111 (discharge capacity of 73 mA h g-1after 500 cycles, Fig. 5c). Moreover, it is clear from theex-situSEM images after 100 cycles(Fig. 5d and 5e) that the surface of bare CF accumulates Li dendrites and “dead Li”, while the Li-Ti3C2Tx-MX@CF anode presents a smooth and clean surface without the formation of “dead Li” and Li dendrites.These results indicate that the Li anode modified with Ti3C2Tx-MX@CF can effectively suppress the growth of Li dendrites, maintain the integrity of the electrode structure, and alleviate the volume change of LMAs during the repeated cycling and will likely find practical applications in LMBs.

Fig. 5 (a) Rate performance of Li-Ti3C2Tx-MX@CF||NCM111 and Li-CF||NCM111 full cells at different current densities. Cycling performance of Li-Ti3C2Tx-MX@CF||NCM111 and Li-CF||NCM111 full cells at a current density of (b)1 C (1 C = 150 mAh g-1) and (c) 5 C. The top-view SEM image of (d) Li-Ti3C2Tx-MX@CF anode and (e) Li-CF in the NCM111 full cell after 50 cycles at 1 C

4 Conclusion

In summary, we have successfully constructed dendrite-free LMAs by using CF modified with lithiophilic Ti3C2Tx-MX as 3D host. The Ti3C2Tx-MX with abundant lithiophilic functional groups serves as nucleation sites to lower the Li nucleation barrier, and consequently induce uniform Li nucleation and suppress dendrite formation. Meanwhile, the 3D CF with interconnected network provides enough space for buffering the volume variation and lower the local current density through its high specific surface area,further inhibiting Li dendrite growth. Benefitting from the synergistic effect of lithiophilic Ti3C2Tx-MX and CF, a dendrite-free morphology is obtained and the assembled symmetrical cell displays excellent cycling stability up to 2 000 h at current density of 4 mA cm-2with capacity of 10 mAh cm-2. Moreover, the Li-Ti3C2Tx-MX@CF||NCM11full cells exhibit the improved cycling stability and rate performance. This work provides an instructive guidance for protecting LMAs and realizing high-performance LMBs.

Data availability statement

The data that support the findings of this study are openly available in Science Data Bank at https://www.doi.org/10.57760/sciencedb.j00125.00032 or https://resolve.pid21.cn/31253.11.sciencedb.j00125.000 32.

Acknowledgements

This work was supported by the financial support from the Natural Science Foundation of Jilin Provinc1 (20220508141RC), th1 111 Project(B13013), the National Natural Science Foundation of China (21873018), the Education Department of Jilin Province (JJKH20221154KJ), Jilin Provincial Research Center of Advanced Energy Materials (Northeast Normal University).