In Situ Reaction Fabrication of a Mixed‐Ion/Electron‐Conducting Skeleton Toward Stable Lithium Metal Anodes

Lithium metal batteries are emerging as a strong candidate in the future energy storage market due to its extremely high energy density. However, the uncontrollable lithium dendrites and volume change of lithium metal anodes severely hinder its application. In this work, the porous Cu skeleton modified with Cu6Sn5 layer is prepared via dealloying brass foil following a facile electroless process. The porous Cu skeleton with large specific surface area and high electronic conductivity effectively reduces the local current density. The Cu6Sn5 can react with lithium during the discharge process to form lithiophilic Li7Sn2 in situ to promote Li‐ions transport and reduce the nucleation energy barrier of lithium to guide the uniform lithium deposition. Therefore, more than 300 cycles at 1 mA cm−2 are achieved in the half‐cell with an average Coulombic efficiency of 97.5%. The symmetric cell shows a superior cycle life of more than 1000 h at 1 mA cm−2 with a small average hysteresis voltage of 16 mV. When coupled with LiFePO4 cathode, the full cell also maintains excellent cycling and rate performance.


Introduction
With the continuous development of secondary batteries, higher requirements have been raised on the energy density of batteries. Due to the low theoretical specific capacity (372 mAh g −1 ) of the graphite, the energy density of traditional Li-ion batteries is not enough, and a higher-energy-density anode is urgently needed. Lithium metal serves as an anode possessing an ultra-high theoretical specific capacity (3860 mAh g −1 ) with low density (0.53 g cm −3 ) and the most negative electrochemical potential (−3.04 V vs. standard hydrogen electrode), making it an outstanding anode material. [1][2][3][4] However, there are still some problems to be settled to realize the practical application of lithium metal anodes: 1) generation of lithium dendrites. Heterogeneous distribution of the electric field at the interface between electrolyte and lithium metal anode caused by the sluggish Li-ions transport dynamic, which is related to high diffusion energy barrier of lithium due to weak Li-Li bond strength, and uneven electrode surface together induce uneven lithium deposition to form dendrites. The growing lithium dendrites may eventually pierce the separator and cause a short circuit. [5,6] 2) On account of the high electrochemical reactivity and severe volume change caused by the transformation reaction of hostless lithium, the interface between electrolyte and lithium is difficult to form the stable solid electrolyte interface (SEI), especially under the high current density, which results in the continuous consumption of metal lithium and electrolyte. In addition, the unstable interface induces the detachment of lithium dendrites from the lithium metal surface to form 'dead Li', leading to the low Coulombic efficiency (CE) and short cycle life of lithium metal anodes. [7,8] In order to optimize the lithium metal anodes, a series of research methods have been presented, mainly including electrolyte additives, [9][10][11] three-dimensional current collectors, [12][13][14][15][16][17] construction of artificial SEI, [18][19][20] solid-state electrolyte, [21,22] and separator modification. [23] Among them, three-dimensional current collectors have been widely applied in lithium metal anodes, incorporating metal and porous carbon frameworks. Particularly, metal frameworks such as Ni foam, Cu mesh, and porous Cu with large specific surface area, [24][25][26] porous structure, and high electronic conductivity can effectively reduce the local current density and alleviate volume change during repeatedly lithium stripping/plating, which are regarded to be practically potential in lithium metal anodes. However, the poor lithiophilicity of most metal frameworks limits its application in lithium metal anodes. Some modified methods have been identified to be effective to improve the lithiophilicity of substrates, such as introducing inorganic N, S active sites, [27,28] lithiophilic metal Au, Ag, [29,30] and metal oxide ZnO. [31] These substances can reduce the nucleation energy barrier and uniform charge distribution to guide the uniform lithium plating. In the meantime, several Li-ions conductive layers such as lithium halides, [32] metal oxides, [33] and sulfides [34] can support more homogeneous Li-ions flux owing to their low diffusion activation energy of Li ions, which is also crucial for smooth lithium deposition. Hence, it is desirable to construct a three-dimensional skeleton with both fast electronic/ion transport and lithiophilicity.
In this work, the porous Cu skeleton modified with Cu 6 Sn 5 layer (P-Cu@Cu 6 Sn 5 ) has been fabricated via chemical dealloying brass foil, following a simple electroless plating method. Porous Cu skeleton (P-Cu) with high electronic conductivity effectively alleviates the volume change during lithium plating/stripping and reduces the local current Lithium metal batteries are emerging as a strong candidate in the future energy storage market due to its extremely high energy density. However, the uncontrollable lithium dendrites and volume change of lithium metal anodes severely hinder its application. In this work, the porous Cu skeleton modified with Cu 6 Sn 5 layer is prepared via dealloying brass foil following a facile electroless process. The porous Cu skeleton with large specific surface area and high electronic conductivity effectively reduces the local current density. The Cu 6 Sn 5 can react with lithium during the discharge process to form lithiophilic Li 7 Sn 2 in situ to promote Li-ions transport and reduce the nucleation energy barrier of lithium to guide the uniform lithium deposition. Therefore, more than 300 cycles at 1 mA cm −2 are achieved in the half-cell with an average Coulombic efficiency of 97.5%. The symmetric cell shows a superior cycle life of more than 1000 h at 1 mA cm −2 with a small average hysteresis voltage of 16 mV. When coupled with LiFePO 4 cathode, the full cell also maintains excellent cycling and rate performance. density. Li 7 Sn 2 obtained by Cu 6 Sn 5 is not only beneficial to the transport of Li ions to homogenize electric field distribution but also reduces the nucleation energy barrier to guide the even plating of lithium. Owing to the advantages above, the P-Cu@Cu 6 Sn 5 current collector exhibits low nucleation overpotential and stable CE for over 300 cycles at 1 mA cm −2 . The symmetric cell exhibits remarkable performance for working more than 1000 h with small increase in hysteresis voltage. Moreover, P-Cu@Cu 6 Sn 5 /Li‖LiFePO 4 full cell also displays superior cycle life and excellent rate performance.

Results and Discussion
The P-Cu@Cu 6 Sn 5 was fabricated by a simple method. First, the bright yellow brass foil was etched with the color turned to orange-red, and the orange-red P-Cu was processed by electroless strategy to obtain the target sample as the color changed to silver-gray ( Figure 1a). From the scanning electron microscope (SEM) characterization, there is a smooth surface with only a few scratches left by rolling for the mere cleaned raw brass foil ( Figure S1a, Supporting Information). After chemical etching, the obtained P-Cu is distributed with many sub-micron-scale pores in the shape of ant nests with a porosity of~38%, which greatly increases the specific surface area and the surface of the P-Cu skeleton after etching is smooth ( Figure S1b, Supporting Information). The thickness of the framework is about 50 μm with channel structure running through the whole bulk phase ( Figure S2, Supporting Information). Compared with P-Cu, P-Cu@Cu 6 Sn 5 obtained by electroless treatment displays rougher surface for uniformly distributed Cu 6 Sn 5 particles connected to each other with the size of 50-100 nm (Figure 1b), and the thickness of the Cu 6 Sn 5 layer is about 411.5 nm ( Figure S3, Supporting Information). In addition, the energy-dispersive spectroscopic analysis (EDS) has been applied to analyze the distribution of the elements. Compared with the original brass foil, the zinc content of P-Cu after chemical etching is extremely low, indicating the radical dealloying of the brass foil ( Figure S4, Supporting Information). The uniformly distributed Sn element demonstrates the successful modification of Cu 6 Sn 5 layer onto the P-Cu surface (Figure 1c). The highresolution transmission electron microscope (HRTEM) image reveals that the crystal lattice spacing is estimated to be 0.304 nm, [35] which is in line with the (22-1) crystal face of Cu 6 Sn 5 (Figure 1d). From the Xray diffraction (XRD) (Figure 1e), the characteristic peaks of (111) and (200) lattice planes attributed to α-CuZn disappear (PDF#50-1333), and are replaced by characteristic peaks of pure Cu after chemical etching (PDF#04-0836), certifying the successful dealloying. After electroless treatment, additional characteristic peaks regarded as the lattice planes of Cu 6 Sn 5 appear (PDF#45-1488), corroborating with the TEM result. The electronic states of different elements on the surface were derived from X-ray photoelectron spectroscopy (XPS). In the highresolution spectrum of Cu 2p (Figure 1f), two distinct peaks located at 952.56 and 932.80 eV belong to 2p 1/2 and 2p 3/2 of Cu 0 in Cu 6 Sn 5 and other pair of peaks situated in 953.70 and 934.4 eV correspond to 2p 1/2 and 2p 3/2 of Cu 2+ in CuO on account of the sensitivity of fresh Cu to air. In Figure 1g, Sn 3d are divided into four peaks, and two of them derived from Sn 0 in Cu 6 Sn 5, [35] while the peaks at 495 and 486.6 eV belong to 3d 3/2 and 3d 5/2 of Sn 4+ in SnO 2 owing to the oxidation of the Cu 6 Sn 5 surface. [36] Furthermore, characteristic peaks evolution of Cu 6 Sn 5 can be recognized in the XRD patterns of the samples at different electroless plating time ( Figure S5a, Supporting Information). With reaction time increased, the signals of Cu 6 Sn 5 gradually increase and achieve the strongest intensity after 120 s. Meanwhile, the corresponding digital photograph also indicates that the Cu 6 Sn 5 layer turns to be uniform and dense as the processing time prolongs, and the loading mass of Cu 6 Sn 5 achieves the upper limit after 120 s ( Figure S5b, Supporting Information). Moreover, the sample with reaction time of 120 s exhibits the best electrochemical performance, which will be discussed below. Hence, it is concluded that the time of electroless plating process mainly affects the loading mass, uniformity, and density of Cu 6 Sn 5 layer to ultimately influence the electrochemical performance of P-Cu@Cu 6 Sn 5 . And it should be noted that P-Cu@Cu 6 Sn 5 with reacting time of 120 s is taken as the most typical sample for study.
In order to explore the internal mechanism of P-Cu@Cu 6 Sn 5 in lithium metal anodes, the cyclic voltammetry (CV) test and density functional theory (DFT) calculations were carried out. During the first discharge process of P-Cu@Cu 6 Sn 5 in the half-cell, a distinct voltage plateau appears (Figure 2a), corresponding to the displacement reaction of lithium with Li 2 CuSn according to the previous research, [36][37][38] while the Li 2 CuSn is derived from the solution reaction between Cu 6 Sn 5 and lithium. The reaction processes are as follows: xLi þ Cu 6 Sn 5 ! Li x Cu 6 Sn 5 , 3Li þ 2Li 2 CuSn ! Li 7 Sn 2 þ 2Cu: The CV test was applied to explore the specific reaction process (Figure 2b and Figure S6, Supporting Information). During the initial cathodic scan for Li‖P-Cu@Cu 6 Sn 5 , two different peaks appear compared with Li‖P-Cu and Li‖Cu foil. One broad peak located at 0.38 V (vs. Li + /Li) attributes to the phase transition from Cu 6 Sn 5 to Li 2 CuSn. The other peak at 0.08 V (vs. Li + /Li) is related to reacting process of Li 2 CuSn to Li 7 Sn 2, [37] which can be detected by the XRD spectrum of P-Cu@Cu 6 Sn 5 discharged to 0.001 V ( Figure S7, Supporting Information). For the first anodic scan, the peaks around 0.5-0.8 V belong to the reaction from Li 7 Sn 2 to Cu 6 Sn 5 . This can also be certified by the XRD spectrum of P-Cu@Cu 6 Sn 5 after being recharged to 1 V ( Figure S7, Supporting Information). In the two cycles later, the cathodic peaks at 0.08 and 0.38 V move to 0.25 and 0.40 V, respectively, and the two CV curves overlapped well with each other, indicating that the reversible Li 7 Sn 2 layer can be generated. DFT calculations were employed to explore the interaction of lithium and the current collectors. Li 7 Sn 2 derived from Cu 6 Sn 5 has a higher binding energy (−3.96 eV) than Cu substrate (−3.04 eV), indicating the better lithiophilicity of Li 7 Sn 2 ( Figure 2c). Meanwhile, the Li-ions diffusion energy barriers on the Li 7 Sn 2 and Cu substrate were calculated as well, which presents the ability of Li-ions diffusion on the different components. The calculation results indicate that the diffusion energy barriers among four pathways of Li 7 Sn 2 are all distinctly lower than that of lithium diffusion on Cu, especially for pathway 1 with diffusion energy barriers of mere 0.0674 eV (Figure 2d-g). The ultra-low activation energy of Li 7 Sn 2 guarantees quick transport of Li ions, which prevents the gathering of Li ions to consequently suppress the generation of lithium dendrites. Furthermore, ab initio molecular dynamics simulations at 300 k were performed to calculate the coefficient of Li-ions diffusion in Li 7 Sn 2 , and the Li-ions diffusion coefficient of 1.1 × 10 −5 cm 2 s −1 also indicates its fast Li-ions transport kinetics ( Figure S8, Supporting Information). Electrochemical impedance spectroscopy (EIS) of the half-cell with various current collectors was applied to further identify the fast Li-ions transport in Li 7 Sn 2 . Before cycling, the charge-transfer impedance (R ct ) of P-Cu@Cu 6 Sn 5 and P-Cu, which refers to the incomplete semicircle corresponding to the high-frequency region in the impedance spectrum, are lower than that of Cu foil owing to the good wettability of the porous structure ( Figure S9a, Supporting Information). After discharge to 0.001 V, the in-situ-formed Li 7 Sn 2 reduces the R ct of P-Cu@Cu 6 Sn 5 . However, the increase of R ct of P-Cu and Cu foil is probably related to the formation of heterogeneous SEI ( Figure S9b, Supporting Information). For the advantages discussed above, it is concluded that the P-Cu@Cu 6 Sn 5 greatly promotes the charge transfer and reduces the nucleation barrier of lithium to eventually achieve prominent electrochemical performance during the following tests.
In order to gain a deeper understanding of the plating behavior of lithium on P-Cu@Cu 6 Sn 5 , SEM was applied to investigate the microscopic morphology of plated lithium at different cycles under the current density of 2 mA cm −2 . After the first discharge of the Li-Cu halfcells, the plated lithium on P-Cu@Cu 6 Sn 5 presents no dendrites and voids. Even after 50 cycles, the plated lithium on P-Cu@Cu 6 Sn 5 is compact and even without obvious cracks (Figure 3a Figure S10c, Supporting Information). Simultaneously, the crosssection images of three current collectors after 50 cycles indicate that the thickness of plated lithium of P-Cu@Cu 6 Sn 5 is only 6.8 μm, which is less than that of P-Cu (21.2 μm) and Cu foil (82.6 μm) (Figure 3g-i). This is owing to the porous structure effectively reducing the current density and inducing lithium to plate inside the porous channel, which can be identified from the morphology changes in the pore structure of P-Cu@Cu 6 Sn 5 and P-Cu before and after cycling (the insets of Figure 3g,h and Figure S11, Supporting Information). Moreover, compared with P-Cu, the Li 7 Sn 2 generated in situ further promotes the even and dense lithium plating on P-Cu@Cu 6 Sn 5 . For Cu foil, on account of its dense two-dimensional planar structure ( Figure S12, Supporting Information), the lithium gradually becomes looser and looser during repeated plating/stripping, leading the thickness of plated lithium to dramatically increase. From all the evidence above in SEM, it is convincing that P-Cu@Cu 6 Sn 5 is helpful to guide flat lithium plating. On the one hand, due to the fast electron/ion transport dynamics promoted by porous Cu skeleton and Li 7 Sn 2 layer, the charge aggregation on the electrode surface is greatly reduced. On the other hand, ultra-low nuclear barrier of Li ions on account of the Li 7 Sn 2 can further induce the uniform nucleation of Li ions and the formation of stable SEI to eventually form the caky-plated lithium in the initial deposition and maintain the smooth surface after several cycles (Figure 3j). On the contrary, uneven nucleation and dendrite growth  will form during lithium plating/stripping on Cu foil owing to the uneven charge distribution, which generates tiny fluffy lithium and cracks as the cycle time prolongs. These will greatly escalate the contact area between lithium and the electrolyte as well as rapidly consume them, resulting in a sharp decrease in cycle life (Figure 3k).
The half-cell measurements of P-Cu@Cu 6 Sn 5 were carried out via lithium repeated plating/stripping at different current densities. By the control experiments in CE combined with the conclusion above, electroless plating with 120 s is regarded as the optimum experiment parameter, which is used for the subsequent experiments ( Figure S13, Supporting Information). The lowest nucleation overpotentials are obtained at different current densities (0.01, 0.5, and 1 mA cm −2 ) for P-Cu@Cu 6 Sn 5 , which is beneficial for smooth lithium plating ( Figure  4a and Figure S14, Supporting Information). The CE of commercial Cu foil drops sharply after 50 cycles at 1 mA cm −2 , meanwhile, the CE of P-Cu oscillates up and down near 94.7% until 170 cycles and decreases below 85% later. In contrast, P-Cu@Cu 6 Sn 5 still retains stability after 300 cycles with the average CE achieved at 97.5% and lowest hysteresis voltage (Figure 4b and Figure S15, Supporting Information). Even increased to 2 mA cm −2 and 3 mA cm −2 , the P-Cu@Cu 6 Sn 5 presents better cycling performance than P-Cu and commercial Cu foil (Figure 4c,d), and P-Cu@Cu 6 Sn 5 exhibits the smallest hysteresis voltage of only 44 mV in the 10 th cycle at 3 mA cm −2 , which identifies its excellent performance at high current density (Figure 4e,f). Furthermore, P-Cu@Cu 6 Sn 5 can maintain stability for more than 50 cycles with a high stripping/plating capacity of 4 mAh cm −2 ( Figure S16, Supporting Information).
Symmetric cells were assembled to investigate the effect of P-Cu@Cu 6 Sn 5 on the cycling stability of lithium metal anodes, and symmetrical cells assembled with bare Li foils were control samples. The P-Cu@Cu 6 Sn 5 /Li composite anodes are formed by pre-plating lithium on the P-Cu@Cu 6 Sn 5 , and the electrolyte is the same as the halfcells used. At the current density of 1 mA cm −2 , P-Cu@Cu 6 Sn 5 /Li remains stable for more than 1000 h with a small average hysteresis voltage of 16 mV, while the bare Li foil shows a continuous increase in hysteresis voltage and becomes short-circuited after 430 h (Figure 5a). The corresponding highly overlapped charge/discharge curves and selected hysteresis voltage profiles also indicate the superior cycling stability and fast kinetic process of P-Cu@Cu 6 Sn 5 /Li (Figure 5b,c and Figure S17, Supporting Information). When increased to 2 and 3 mA cm −2 , the P-Cu@Cu 6 Sn 5 /Li still maintains stability for over 250 and 120 h, respectively, with small hysteresis voltages, which are much better than that of bare Li foil ( Figure S18a,b, Supporting Information). Notably, at the current density of 3 mA cm −2 with the Li plating/stripping capacity of 3 mAh cm −2 , more than 210 h is achieved during Li plating/stripping for the P-Cu@Cu 6 Sn 5 /Li with a small hysteresis voltage of 50 mV ( Figure S18c, Supporting Information). Even more, onlỹ 65 mV of hysteresis voltage is obtained at large current density of 5 mA cm −2 for P-Cu@Cu 6 Sn 5 /Li. In contrast, the hysteresis voltage of the control sample becomes disordered probably due to the bare Li foil that cannot withstand sudden current change (Figure 5d). The notable performance of the P-Cu@Cu 6 Sn 5 /Li in symmetrical cells is compared with related literatures on lithium metal composite anodes in Figure 5e and Table S1. EIS was introduced to further explore the P-Cu@Cu 6 Sn 5 mechanism in lithium metal anodes. [39] Before cycling, the R ct of the interface between the bare Li foil and the electrolyte is 126 Ω, while that of the P-Cu@Cu 6 Sn 5 /Li is only 44 Ω (Figure 5f). After 50 cycles, the R ct of P-Cu@Cu 6 Sn 5 /Li decreases to the minimum value of 9 Ω. This is because the porous Cu skeleton and Li 7 Sn 2 generated in situ promote charge diffusion and then smoothen the surface of plated lithium (Figure 5g). To further identify this conclusion, the Tafel curve test of symmetric cells was conducted. [40][41][42] The exchange current density of P-Cu@Cu 6 Sn 5 /Li is j 0 = 0.109 mA cm −2 , which is larger than that of bare Li foil. This again demonstrates the fast Li-ions/electron transport of P-Cu@Cu 6 Sn 5 /Li (Figure 5h). In order to intuitively certify the positive effect of P-Cu@Cu 6 Sn 5 /Li on stabilizing lithium metal anodes, in situ electrochemical optical microscope test was conducted to observe the dendrite growth process during lithium plating. [43] For bare Li foil, no dendritic clusters appear at the interface before lithium plating. With the extension of time to 30 min, obvious dense linear branch dendrites can be observed and begin to entangle into clusters, and the dendrites cover the entire observation window at the end (Figure 5i). In contrast, dense and uniform-plated lithium is formed on P-Cu@Cu 6 Sn 5 /Li during the whole process, efficaciously guaranteeing the stability of the deposition surface (Figure 5j). Full cells were assembled to investigate the application of P-Cu@Cu 6 Sn 5 in practice. 6 mAh cm −2 lithium was first plated onto P-Cu@Cu 6 Sn 5 to match the LiFePO 4 (loading mass: ≈9 mg) to assemble the full cell, and bare Li foil anode (thickness: ≈300 μm, diameter: 12 mm) was used as the control sample. The P-Cu@Cu 6 Sn 5 / Li‖LiFePO 4 full cell exhibits stable cycle performance after 600 cycles with the specific capacity of 120.3 mAh g −1 retained under a low N/ P (defined as the ratio of negative-to-positive capacity) of~4.4, corresponding to the capacity retention of 81%. In reverse, the specific capacity of bare Li‖LiFePO 4 rapidly decreases to 78.6 mAh g −1 after the same cycles under a high N/P of~45.0, and the capacity retention is only 55.4% (Figure 6a). Even more, the 1st, 10th, 50th, 100th, 200th, 400th, and 600th charge/discharge curves also show that the P-Cu@Cu 6 Sn 5 /Li‖LiFePO 4 presents more stable and smaller hysteresis voltages than that of bare Li‖LiFePO 4 (Figure 6b and Figure S19a, Supporting Information), which indicate better diffusion kinetics of P-Cu@Cu 6 Sn 5 /Li. For the rate test, the full cells were first activated at the low current density of 0.2C for five cycles. When the current density is increased from 0.5C to 5C, the specific capacities of P-Cu@Cu 6 Sn 5 /Li‖LiFePO 4 are 150.1, 138.7, 126.5, 117.1, and 109.6 mAh g −1 . After returning to 1C, the specific capacity of 136.5 mAh g −1 is maintained while the specific capacity of the bare Li‖ LiFePO 4 full cell drops rapidly at the same process (Figure 6c). The charge/discharge curves of P-Cu@Cu 6 Sn 5 /Li‖LiFePO 4 also reveal obviously smaller hysteresis voltages with increasing rates than that of bare Li‖ LiFePO 4 (Figure 6d and Figure S19b, Supporting Information), indicating the smaller polarization of P-Cu@Cu 6 Sn 5 /Li than that of bare Li. The excellent cycling performance of P-Cu@Cu 6 Sn 5 /Li can be attributed to the lithophilic Li 7 Sn 2 derived from Cu 6 Sn 5 , which guarantees uniform lithium nucleation/ growth and reversible lithium plating/stripping. Meanwhile, the P-Cu@Cu 6 Sn 5 featuring high ionic/electron conductivity supplies rapid Li-ions and electrons diffusion and reduces the interface impedance in high current density, thereby leading to better rate performance of the full cells. All these results manifest that P-Cu@Cu 6 Sn 5 can serve as a stable skeleton for lithium metal anodes with promising practical application potential.

Conclusion
In summary, an excellent three-dimensional current collector was obtained by introducing Cu 6 Sn 5 layer onto chemically dealloyed micron-sized pores Cu surface via electroless plating means. The porous Cu can supply high electron transport efficiency and reduce local current density during repeated lithium plating/stripping processes. Li 7 Sn 2 generated in the discharge process promotes Li-ions transport and dramatically reduces the nucleation overpotential. With these advantages, more than 300 cycles at 1 mA cm −2 are achieved in the half-cell of P-Cu@Cu 6 Sn 5 with an average CE of 97.5%. The symmetric cells of P-Cu@Cu 6 Sn 5 /Li maintained excellent performance for cycling more than 1000 h with a small average hysteresis voltage of 16 mV. The P-Cu@Cu 6 Sn 5 /Li‖LiFePO 4 also exhibits excellent cycling stability for more than 600 cycles and rate performance at 5C. This tactic may offer a facile and wholesale route to manufacture advanced lithium metal skeletons.

Experimental Section
Preparation of P-Cu@Cu 6 Sn 5 skeleton: First, 50-μm-thick brass foil (α-CuZn, Cu 0.64 Zn 0.36 ) was punched into 12-mm-diameter discs with the ultrasonic clean in deionized water for 5 min to remove surface impurities. The pre-cleaned brass foils were immerged into 3 mol/L HCl + 2 mol/L NH 4 Cl mixed solution, then heated at 80°C for 8.5 h until no obvious bubbles appeared. The color of the brass foils changed from yellow to orange-red. The treated samples were immersed in the electroless plating solution for a period of time with water bath heating under 50°C, and the surface of P-Cu changed from orange-red to silvergray. The electroless plating solution was composed of weight of 3% tin tetrafluoroborate, 2% sodium hypophosphite, 6% fluoroboric acid, 8% thiourea, and Materials characterizations: The phase structure of the material was characterized by powder X-ray diffraction (X'Pert3 Powder, PANalytical, Netherlands) with Cu Kα radiation (λ = 0.154178 nm) at a scanning speed of 5°min −1 between 10°and 80°. SEM (LYRA3, TESCAN, Czech) equipped with energydispersive X-ray analyzer (EDX) and TEM (FEI Talos F200X G2, Thermo Fisher Scientific, USA) were applied for the morphology characterization. For the preparation of lithium metal deposition sample, CR2016 half-cell was assembled followed by 2 mAh cm −2 amount of lithium plated on the current collector at 2 mA cm −2 . XPS measurements were achieved by PHI-5000versaprobe (Thermo Fisher Scientific, USA) with Al Kα (1486.6 eV) as the X-ray source. In situ optical microscope experiments were observed by galvanostatic discharge of lithium metal symmetric cells in a transparent mold under an optical microscope (WST-H3800, WST Electronic Technology, Shenzhen). The porosity of P-Cu was calculated by the following formula: where M is the mass of the sample, and the unit is g; V is the volume of the sample, and the unit is cm 3 ; ρ s is the theoretical density of the substance constituting the porous metal, and the unit is g cm −3 . In our research, M corresponds to the mass of P-Cu obtained by the electronic balance (31.25 mg). V corresponds to the volume of P-Cu (diameter: 12 mm, thickness: 50 μm). ρ s corresponds to the theoretical density of pure copper. Electrochemical measurements: All electrochemical tests were performed by assembling CR2016 coin cells in an argon-filled glove box. To test the CE and nucleation overpotential (CT3002A, LAND, Wuhan), bare Li foil was used as the counter electrode together with Cu foil, P-Cu, or P-Cu@Cu 6 Sn 5 as the working electrode, Celgard 2325 as the separator and 1 M lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) in cosolvent of 1,3-dioxolane (DOL) and 1,2-dimethoxyethane (DME) (volume ratio: 1:1) as electrolyte solution. To keep the experiment consistent, 60 μL electrolyte was dripped into each cell. Before the test, the cell was pre-cycled at 100 μA for five cycles between 0 and 0.5 V (vs. Li + /Li) to remove impurities and stabilize the SEI. A fixed capacity of 1 mAh cm −2 of lithium was deposited onto the current collectors at different current densities (1, 2, and 3 mA cm −2 ), then charged to 0.5 V for stripping at each cycle. For the symmetric cell test, 6 mAh cm −2 of lithium was pre-plated onto P-Cu@Cu 6 Sn 5 at 0.5 mA cm −2 to obtain P-Cu@Cu 6 Sn 5 /Li. After being washed with DME, two pieces of P-Cu@Cu 6 Sn 5 were assembled for testing under the fixed capacity of 1 mAh cm −2 , 3 mAh cm −2 , or 4 mAh cm −2 at different current densities, and bare Li foil was assembled for control sample. EIS and linear sweep voltammetry (LSV) tests of symmetric cells were conducted by electrochemical workstation (ZIVE SP2, WonATech, Korea). Frequency range of EIS was from 100 KHz to 10 mHz at a signal amplitude of 5.0 mV. LSV test was carried out between −0.2 and 0.2 V with the scan speed of 0.2 mV s −1 . CV test was conducted by CS2350H dual potentiostat (CS150H, Corrtest, Wuhan).
The full cells were assembled with LiFePO 4 as cathodes, bare Li foil or P-Cu@Cu 6 Sn 5 /Li (with 6 mAh cm −2 of lithium) as anodes. Electrolyte and separator were consistent with the CE tests possessed. To fabricate LiFePO 4 cathodes, the slurry composed of LiFePO 4 powder, acetylene black, and polyvinylidene fluoride mixed in N-methyl pyrrolidone at a mass ratio of 8:1:1 was coated onto carboncoated aluminum foil through a four-sided preparer. Then, the sample was put into a vacuum drying oven at 110°C for 12 h to obtain the LiFePO 4 electrodes. The full cell measurements were carried out via galvanostatically cycled between 2.5 and 3.8 V.
DFT calculations: The Vienna Abinitio Simulation Package (VASP) [44,45] was used for the theoretical study on account of the density functional theory (DFT). The exchange potential was expressed by generalized gradient approximation (GGA) accompanied by the Perdew-Burke-Emzerhof (PBE) function. [46] The plane-wave energy cutoff was set to 520 eV and the convergence criterion of crystal structure relaxation was 0.05 eVÅ −1 in force and 10 −5 eV per atom in energy. The Brillouin zones (BZ) were sampled by K-point grids of 2 × 1 × 5. The binding energy (E b ) was calculated by the following equation: E total , E substrate , and E Li represent the entire energy of the adsorbed system, the entire energy of the substrate, and the entire energy of the single lithium atom, respectively. A vacuum of 15Å was applied to prevent the interaction among the adjacent atomic slabs in model for investigation of Li-ions diffusion. Transition states of Li-ions diffusion on different pathways were calculated by the nudged elastic band (NEB) method. [47]