Bismuthene Arrays Harvesting Reversible Plating‐Alloying Electrochemistry Toward Robust Lithium Metal Batteries

3D lithiophilic skeletons have attracted enormous attention in homogenizing local current distribution and optimizing metal deposition in the pursuit of robust Li metal anodes. Nonetheless, their practicability is markedly plagued by the cumbersome production routes and mediocre Coulombic efficiency (CE) of Li plating/stripping. Herein, scalable in situ growth of uniform bismuthene arrays over commercial Cu foam via spontaneous galvanic replacement reaction is demonstrated. Exhaustive structural/electrochemical measurements in combination with theoretical calculations collectively disclose the reversible plating‐alloying mechanism, wherein the formed Li3Bi alloy interphase aids to lower the Li nucleation overpotential and elevate the CE performance. The thus‐designed Li metal electrode sustains a stable cyclic operation at 1 mA cm−2/1 mAh cm−2 for 1600 h. When paired with LiFePO4 and sulfur cathodes, the Li metal batteries enable gratifying rate capability and cycling durability. This straightforward maneuver opens a new frontier in the scalable manufacturing of pragmatic current collectors in an economic fashion.


Introduction
Li metal is an appealing anode candidate on the ground of its high theoretical capacity of 3860 mAh g À1 and low electrochemical potential (À3.04 V vs standard hydrogen electrode). [1][2][3][4][5][6] Nevertheless, the potential practicability of Li metal anode is plagued by uncontrollable dendrite growth and inferior Coulombic efficiency (CE). [7][8][9][10][11][12][13] To circumvent these hurdles, 3D current collector has stimulated a burgeoning interest to enhance the reversibility and durability of Li metal anodes. [14][15][16] Although this promising strategy is conducive to reducing local current density and alleviating dendrite growth during the repeated Li plating/stripping process, such critical effects are prone to degrade continuously due to the intrinsically lithiophobic feature and tremendous nucleation overpotential of common carbonaceous/metallic current collectors. To this end, the conformal modification of lithiophilic moieties is expected to homogenize current distribution and facilitate nucleation behavior, accordingly mitigating the Li dendrite formation. [17][18][19][20] In this regard, the major concerns compromising the large-scale application of 3D lithiophilic current collectors lie in their complicated preparation strategies and mediocre CE performance. [21][22][23] Admittedly, there is a significant lack of straightforward toolboxes to fabricate a 3D lithiophilic current collector, which can rarely bypass the tedious synthetic process and high temperature/pressure reaction environment. [24,25] Encouragingly, the galvanic replacement reaction (GRR) represents a spontaneous redox process without the involvement of heat, pressure, and inert atmosphere, thereby readily lowering the potential production cost. [26][27][28][29] In terms of the reaction process, a metal substrate harnessing the relatively low reduction potential is exposed to soluble metal ions that exhibit a higher reduction potential. As a result, the difference in their redox potentials will give rise to a gradual oxidation of the metal, accompanied by the reduction of soluble metal ions. [30] Even though the hierarchical array architecture can be in target achieved, few related efforts have been devoted to fabricating 3D lithiophilic current collector in the pursuit of advanced Li metal anodes.
Another bottleneck for common 3D current collector pertains to the dramatic increase of solid electrolyte interface (SEI) arising from their abundant specific surface area upon metal deposition, DOI: 10.1002/sstr.202200313 3D lithiophilic skeletons have attracted enormous attention in homogenizing local current distribution and optimizing metal deposition in the pursuit of robust Li metal anodes. Nonetheless, their practicability is markedly plagued by the cumbersome production routes and mediocre Coulombic efficiency (CE) of Li plating/stripping. Herein, scalable in situ growth of uniform bismuthene arrays over commercial Cu foam via spontaneous galvanic replacement reaction is demonstrated. Exhaustive structural/electrochemical measurements in combination with theoretical calculations collectively disclose the reversible platingalloying mechanism, wherein the formed Li 3 Bi alloy interphase aids to lower the Li nucleation overpotential and elevate the CE performance. The thus-designed Li metal electrode sustains a stable cyclic operation at 1 mA cm À2 /1 mAh cm À2 for 1600 h. When paired with LiFePO 4 and sulfur cathodes, the Li metal batteries enable gratifying rate capability and cycling durability. This straightforward maneuver opens a new frontier in the scalable manufacturing of pragmatic current collectors in an economic fashion.
which inevitably triggers the irreversible loss of active Li meal and deteriorating the initial CE. [31,32] To cope with this challenge, Li alloying has emerged as an ingenious strategy to promote the lithiophilicity of Li metal anodes and enable homogeneous Li nucleation. In this sense, multifarious Li-based alloys (e.g., Li-Ag, Li-Au, Li-Zn, Li-Sn, Li-Hg) can not only function as excellent ion conductor for facilitating Li þ adsorption and transport, but also play a pivotal role in enhancing the interfacial stability. [33][34][35][36][37] Such an alloying strategy harnessing the highly reversible alloy interphase evolution can accordingly avoid detrimental surface deposition, which further alleviates the Li dendrite formation and imparts excellent cyclic durability. Nonetheless, it remains elusive to elucidate the underlying Li plating-alloying mechanism. In further contexts, a multitude of novel Li alloys still lack in-depth investigations with respect to their design philosophy and fundamental property. Notably, barely no attention has been paid to probing Li-Bi alloy owing to the low melting point and high oxophilicity of metallic Bi. [38,39] In this contribution, we demonstrate the feasible proof-ofconcept of GRR between metallic Cu and Bi 3þ by using commercial Cu foam (denoted as CF) as the growth substrate, which can readily contribute to the formation of bismuthene nanosheet arrays on Cu foam (denoted as Bi-ene@CF). The growth principle of bismuthene array is probed, which is attributed to the considerable concentration difference between Cu 2þ and Bi 3þ in the reaction solution. Equally importantly, the Li alloying nucleation mechanism is disclosed with the aid of elaborate structural characterizations and electrochemical measurements. It is revealed that the Bi-ene@CF is in favor of simultaneously lowering the nucleation overpotential and elevating the CE, thereby sustaining a stable stripping/plating operation at 1 mA cm À2 /1 mAh cm À2 for 1600 h in the symmetric cell test. More encouragingly, the Bi-ene@CF-Li can also deliver satisfying electrochemical performance in Li metal full batteries. This work offers a straightforward maneuver to fabricate a 3D current collector harnessing the reversible Li alloying nucleation electrochemistry, which opens an avenue to achieve pragmatic Li metal anodes in an economic fashion.

Results and Discussion
Commercial CF was employed as the substrate for bismuthene growth and the host for Li metal deposition, which is expected to accommodate volume change and homogenize current distribution because of the 3D porous skeleton architecture. Upon soaking CF pieces into the N,N-dimethylformamide (DMF) solution containing 0.02 M BiCl 3 , GRR would spontaneously take place and accordingly lead to the controllable formation of aligned bismuthene nanosheet arrays (Figure 1a). Although the standard reduction potential of Bi 3þ /Bi (0.31 V vs standard hydrogen electrode) is slightly lower than that of Cu 2þ /Cu (0.34 V vs standard hydrogen electrode), the Nernstian contribution arising from the huge ion concentration difference in solution could easily drive the occurrence of redox reaction shown as follows It is well established that the direction of redox reaction is predicted by the Nernst equation where E represents the actual cell potential, E 0 is the standard cell potential, R is the ideal gas constant, T is the temperature, n is the number of electrons transferred, F is the Faraday constant, a ox stands for the activity of the oxidized species and a red refers to the activity of the reduced species. In terms of the aforementioned redox reaction, E 0 ¼ À0.03 V, hence it is not spontaneous under the standard working condition. Note that the RT nF ln a ox a red reflects the Nernstian contribution to a realistic reaction system. In this case, the considerable difference in ion concentration (the concentration of Bi 3þ is significantly higher than that of Cu 2þ ) will yield a sufficient Nernstian contribution to drive the gradual oxidation of Cu and the concurrent reduction of Bi 3þ to Bi. It is noted that the employment of Cu foam enables the controlled growth of hierarchical bismuthene arrays without any violent reactions. Figure 1b demonstrates that a piece of Bi-ene@CF with a size of 10 cm Â 10 cm can be readily achieved by this straightforward GRR, which is envisaged to implement the scalable application on protecting Li metal anode.
As illustrated in the scanning electron microscopy (SEM) images, vertically erected bismuthene nanosheets are uniformly grown on the CF substrate ( Figure 1c; S1, Supporting Information), which are anticipated to accommodate volume change and homogenize current distribution during repeated Li plating/stripping. With the aim of investigating the timedependent morphology evolution of bismuthene and further probing the morphology-dependent protective effect on Li metal anode, the reaction time is accordingly adjusted. It is revealed that the bismuthene continuously grows with the increase of reaction time, accordingly giving rise to the size enlargement along with gap extension between nanosheets ( Figure S2, Supporting Information). Moreover, the transmission electron microscopy (TEM) image was acquired to verify the morphologic feature of bismuthene nanosheets ( Figure 1d). Note that the high-resolution TEM (HRTEM) image in Figure 1e showcases the interlayer spacing of 0.23 nm, which is in agreement with the (110) lattice fringes of metallic Bi. The thickness of bismuthene nanosheets was measured to be %2.0 nm by atomic force microscopy (AFM) inspection (Figure 1f ), which approximately corresponds to two atomic layer thickness.
The Bi-ene@CF was further subject to the X-Ray diffraction (XRD) measurement (Figure 1g). The presented peaks at 27.2°, 37.9°, and 39.6°are assigned to the (012), (104), and (110) facets of metallic Bi, manifesting the high crystallinity of bismuthene nanosheet arrays. [40] Meanwhile, the peaks corresponding to the Cu substrate can also be observed for Bi-ene@CF and CF ( Figure S3, Supporting Information). Raman spectrum was collected to further identify the structural fingerprints of Bi ( Figure S4, Supporting Information). Note that the signals arising from in-plane E g (60.2 cm À1 ) and out-of-plane A 1g (89.0 cm À1 ) mode could be indexed as two-dimensional metallic bismuthene. [41] Furthermore, an X-Ray photoelectron spectroscopy (XPS) spectrum of Bi-ene@CF was obtained to disclose the surface states and chemical environments (Figure 1h).
Remarkably, the Bi 4f spectrum can be divided into two peaks locating at %159.1 and 164.3 eV, which are assigned to 4f 7/2 and 4f 5/2 signals of Bi 3þ , respectively. [42] These spectroscopic characterizations collectively validate the successful synthesis of high-crystallinity bismuthene nanosheet arrays.
To uncover the enhancement of electrochemical performance under the regulation of bismuthene arrays, CE measurement was carried out to evaluate the cyclic durability of Li plating/ stripping in a half-cell configuration. Note that such cells were assembled by pairing routine Li metal with Bi-ene@CF (or CF). [43,44] To begin with, Bi-ene@CF electrodes with different reaction times of n (n ¼ 1, 3, 6, 12, and 24) h (denoted as Bi-ene@CF-n) were subject to the CE test to confirm the optimal reaction condition ( Figure 2a). It is revealed that either Bi-ene@CF-1 or Bi-ene@CF-24 will trigger the rapid attenuation of CE, whereas the Bi-ene@CF-3 (or 6, 12) can effectively promote the CE performance. It is, therefore, deduced that the appropriate reaction time of GRR is key to obtaining the favorable morphological features of bismuthene nanosheets, thus further facilitating the Li plating/stripping process. In consequence, Bi-ene@CF-3 was employed to showcase the protective effect of bismuthene nanosheet arrays in the following electrochemical  measurements. It is known that the high specific area of 3D CF structure will cause the augmentation of SEI at the initial electrochemical cycling process, thereby giving rise to the inferior cycling stability of Li plating/stripping. Encouragingly, the CE value of Bi-ene@CF exhibits an obvious advantage over that of commercial CF under a current density of 1 mA cm À2 and a plating capacity of 1 mAh cm À2 (denoted as 1 mA cm À2 / 1 mAh cm À2 ) during the initial 50 cycles (Figure 2b). In addition, the Bi-ene@CF also displays a more stable cycling operation under a more stringent condition of 5 mA cm À2 /1 mAh cm À2 . It is worth noting that the Li nucleation overpotential plays a pivotal role during the repeated Li plating/stripping process. As illustrated in Figure 2c, the overpotential of Bi-ene@CF (32 mV) is considerably lower than that of CF (132 mV), indicative of the elevated reaction kinetics of Li deposition. Even at the higher current densities of 3 mA cm À2 /1 mAh cm À2 and 5 mA cm À2 /1 mAh cm À2 , the Bi-ene@CF can also possess a relatively lower overpotential value (Figure 2d; S5, Supporting Information). In terms of the voltage-capacity curves after different cycles at 1 mA cm À2 /1 mAh cm À2 , the Bi-ene@CF exhibits less capacity loss and lower voltage hysteresis (Figure 2e; S6, Supporting Information), manifesting the stable Li plating/stripping process under the regulation of bismuthene arrays. Moreover, the half cells were subject to the electrochemical impedance spectroscopy (EIS) test after 25 cycles, wherein the Bi-ene@CF displays a lower charge transfer resistance (Figure 2f ), corroborating that Bi-ene@CF can ensure faster interfacial reaction kinetics as compared to the bare CF. In further contexts, the Bi-ene@CF-Li (or CF-Li) electrodes with a prestored Li capacity of 10 mAh cm À2 were acquired via a typical electrodeposition process, which were further used as the identical electrode to fabricate symmetric cells. Figure 2. Electrochemical performance of Bi-ene@CF. a) Coulombic efficiency (CE) curves of Bi-ene@CF with different reaction times at 1 mA cm À2 / 1 mAh cm À2 . b) CE curves of Bi-ene@CF and CF at 5 mA cm À2 /1 mAh cm À2 and 1 mA cm À2 /5 mAh cm À2 . c) Voltage-capacity curves of Bi-ene@CF and CF at 1 mA cm À2 /1 mAh cm À2 . d) Comparison of the Li nucleation overpotential on Bi-ene@CF and CF at different current densities. e) Voltage-capacity curves of Bi-ene@CF after different cycles at 1 mA cm À2 /1 mAh cm À2 . f )Electrochemical impedance spectroscopy (EIS) curves of Bi-ene@CF and CF after 25 cycles. g) Cycling performance of symmetric cells at 1 mA cm À2 /1 mAh cm À2 . Inset: Enlarged voltage-time profiles at the initial cycling stage. h) Cycling performance of Bi-ene@CF-Li-based symmetric cells with a depth of discharge (DOD) of 80% at 8 mA cm À2 /8 mAh cm À2 . As demonstrated in Figure 2g, Bi-ene@CF-Li can sustain a stable cycling operation with a low voltage hysteresis of %14 mV over 1600 h at 1 mA cm À2 /1 mAh cm À2 , whereas the CF-Li displays a higher voltage hysteresis of %25 mV. Even worse, an apparent short circuit would occur for the latter after only 370 h because of the severe dendrite growth. The Bi-ene@CF-Li also exhibits an overwhelming advantage over CF-Li with respect to the cyclic stability at 5 mA cm À2 /5 mAh cm À2 ( Figure S7, Supporting Information), which can steadily operate over 500 h. It is noted that the high Li utilization is of critical significance to the pragmatic application of Li metal anodes. When the depth of discharge (DOD) is elevated to 80%, the Bi-ene@CF-Li can also cycle over 60 h without obvious battery failure (Figure 2h). Such electrochemical results strongly bear out that the as-synthesized bismuthene is beneficial to mitigating the dendrite formation and prolonging the cyclic lifespan of lithium metal batteries.
To further shed light on the underlying Li nucleation mechanism, a variety of electrochemical measurements and structural characterizations were performed. As depicted in Figure 3a, cyclic voltammetry (CV) profiles were collected within a voltage range of 0-1 V. As for the Bi-ene@CF electrode, two cathodic peaks (peak i and ii) are assigned to the stepwise alloying reactions upon Li plating process. In the meantime, an anodic peak (peak iii) stemming from the reversible dealloying reaction can be observed. In contrast, the CF electrode can scarcely manifest an apparent current response due to the routine Li plating/stripping behavior without the alloying formation. When the half cell based on the Bi-ene@CF electrode was discharged to 0 V, the corresponding XRD pattern was collected to decipher the Li-Bi alloy interphase formed during the Li plating process (Figure 3b). Note that two peaks at %22.9°and 26.5°are indexed as the (111) and (200)    assigned to Li (PDF#15-0401) can be detected except for the diffraction peaks of the Cu substrate ( Figure S8, Supporting Information). To analyze the surface component and bonding information of Li metal anodes, the thus-derived electrodes were subject to the XPS test (Figure 3c). With regard to the CF-Li electrode, the characteristic XPS peaks are assigned to several Libased compounds (i.e., LiF, LiOH, and Li 2 CO 3 ). [45,46] It is noted that these peaks of the Bi-ene@CF-Li electrode shift toward lower binding energy, which is accompanied by the appearance of a new peak that is assigned to Li 3 Bi alloy. Taken together, the above characterization results verify the Li plating-alloying mechanism and decode the generation of Li 3 Bi alloy interphase.
To gain in-depth insight into the impact of Li 3 Bi alloy interphase on effectively modulating Li metal deposition, density functional theory (DFT) calculations pertaining to the thermodynamics and kinetics of the Li nucleation process were carried out. [47] First, the density of states (DOS) near the Fermi level were examined for Li (110) and Li 3 Bi (111) facets (Figure 3d) with optimized crystal configurations ( Figure S9, Supporting Information). Note that the DOS characteristics of Li 3 Bi are quite weak around the Fermi level as compared to the pure Li scenario. Therefore, it is reasonable to deduce that the Li 3 Bi interphase can restrict the electron conduction throughout the SEI layer and hence facilitate the Li deposition at the interface, thereby alleviating the side reactions with extra electrolyte consumption and improving the cyclic stability of Li plating/stripping. Furthermore, the adsorption and diffusion behaviors of Li atoms on Li 3 Bi and Li were probed. It is well established that the enhanced adsorption energy for Li adatoms can lead to the effective suppression of dendrite growth. In this case, the possible adsorption sites between a Li adatom and Li (110) or Li 3 Bi (111) facet were taken into account ( Figure S10 and S11, Supporting Information). As such, the adsorption energies of Li 3 Bi (111) are always higher than those of Li (110) (Figure 3e), which can aid in enhancing the lithiophilicity and mitigating the dendrite formation. Given the fact that the Li atom diffusion is also of importance for the ultimate Li deposition, the surface diffusion behavior of Li adatoms was monitored, with the diffusion paths on Li (110) or Li 3 Bi (111) presented in Figure 3f,g, respectively. It is intriguing that the diffusion energy barrier on the Li 3 Bi (111) is significantly higher than that on the Li (110) (Figure 3h). This suggests that the migration of Li adatoms across the surface of Li 3 Bi alloy can be effectively restrained at the initial Li nucleation stage, accordingly suppressing the early formation of Li dendrites and optimizing the subsequent Li plating/stripping behavior. Meanwhile, the energy barrier also remains low with regard to another diffusion path on the surface of Li (110) facet ( Figure S12 and S13, Supporting Information), Figure 4. Deposition morphology of Bi-ene@CF. SEM images of a-d) Bi-ene@CF-Li and e-h) CF-Li with a plating current density/capacity of a,e) 1 mA cm À2 /1 mAh cm À2 , b,f ) 3 mA cm À2 /3 mAh cm À2 , c,g) 5 mA cm À2 /5 mAh cm À2 , and d,h) 1 mA cm À2 /1 mAh cm À2 after 100 cycles. Scale bars: a-h) 20 μm. i,j) Schematic diagrams of the Li plating behavior and cycling performance on i) Bi-ene@CF and j) CF.  Figure 4a-c that metallic Li could smoothly deposit on Bi-ene@CF, which is free of apparent agglomerated lumps with a plating capacity varying from 1 through 3 to 5 mAh cm À2 . More impressively, the precipitated Li metal also exhibits favorable spherical morphology instead of mossy dendritic architecture after 100 cycles at 1 mA cm À2 /1 mAh cm À2 (Figure 4d), indicative of the ameliorated Li nucleation behavior regulated by the Li 3 Bi interphase. In contrast, a multitude of uneven pits and dendrites can be observed on the surface of the CF electrode under similar test conditions (Figure 4e-h). This uneven and loose Li precipitation would not only consume abundant electrolytes to generate new SEI but also trigger the formation of "dead Li" to increase interfacial resistance. Moreover, the favorable Li deposition morphology of Bi-ene@CF is also evidenced by the low-magnification SEM images ( Figure S14, Supporting Information). Inspired by the Li plating-alloying mechanism and dendrite-free deposition morphology, schematic diagrams showing the Li deposition behavior on Bi-ene@CF and CF are presented in Figure 4g,i, respectively. Furthermore, EIS measurements after the activation of three cycles were supplemented to further confirm the favorable ion diffusion in Li 3 Bi alloy ( Figure S15, Supporting Information). The deposited Li metal would react with bismuthene film to derive Li 3 Bi alloy interphase upon the Li plating, which is in favor of further guiding the homogeneous deposition of Li metal and ultimately leading to a dendrite-free morphology after the repeated Li plating/stripping process.
To further demonstrate the practical potential of this 3D lithiophilic current collector harnessing reversible plating-alloying mechanism, Bi-ene@CF-Li electrodes with a prestored capacity of 10 mAh cm À2 Li were employed as the anodes and paired with common cathodes to assemble Li metal full cells. [48][49][50][51] To begin with, when coupled with the commercial LiFePO 4 cathode, the as-fabricated Bi-ene@CF-Li||LiFePO 4 cell showcases a better rate capability than CF-Li||LiFePO 4 cell (Figure 5a). More encouragingly, when the current density is switched back to 0.1 C (1 C ¼ 170 mA g À1 ), the Bi-ene@CF-Li||LiFePO 4 cell can still maintain a stable operation, whereas the CF-Li||LiFePO 4 cell exhibits a rapid capacity decay. Galvanostatic charge/discharge (GCD) curves of both cells under varied current densities are displayed (Figure 5b; S16, Supporting Information). Note that the Li metal cell under the regulation of bismuthene arrays possesses a higher CE value and lower voltage polarization (Figure 5c), implying that the plating-alloying mechanism can effectively facilitate Li nucleation and enhance reaction kinetics. When assembled into a Li-S battery, reduced graphene oxide (rGO) was employed as the sulfur host to fabricate the S/rGO cathode. [52,53] Apparently, Bi-ene@CF-Li||S/rGO cell delivers the capacities of 1100, 990, 875, and 810 mAh g À1 at 0.1, 0.2, 0.5, and 1.0C (1.0C ¼ 1672 mA g À1 ), outperforming that of the CF-Li||S/rGO counterpart under various current densities (Figure 5d). Their corresponding GCD curves indicate that Bi-ene@CF-Li||S/ rGO cell showcases a smaller voltage polarization and faster www.advancedsciencenews.com www.small-structures.com redox kinetics (Figure 5e; S17, Supporting Information), which could be attributed to the highly reversible Li alloying nucleation mechanism. As-assembled cells were further subject to cyclic performance tests at 0.5C, in which the Bi-ene@CF-Li||S/rGO cell affords superior performances with regard to either specific capacity or cyclic stability (with an initial capacity of 875 mAh g À1 and a capacity retention of 83%) (Figure 5f ). These results collectively corroborate that the Bi-ene@CF current collector can significantly enhance the electrochemical performance of the Li metal batteries, holding great potential in the pursuit of pragmatic applications.

Conclusions
In summary, we have developed a straightforward maneuver to realize the controllable growth of hierarchical bismuthene nanosheet arrays on commercial CF via spontaneous GRR, which readily bypasses tedious synthetic procedures to enable the in situ decoration of lithiophilic moiety. The thus-derived Bi-ene@CF current collector can not only lower the Li nucleation overpotential but also boost the CE performance owing to the generated Li 3 Bi alloy interphase. Furthermore, the underlying plating-alloying electrochemistry has been probed via exhaustive electrochemical tests, structural characterizations, and theoretical calculations. Consequently, the symmetric cell based on Bi-ene@CF-Li can maintain a stable cyclic operation at 1 mA cm À2 /1 mAh cm À2 over 1600 h. The assembled Li metal batteries also harvest satisfying rate capability and cycling stability. It is anticipated that our work would promote the pragmatic application of 3D lithiophilic current collector in the realm of metal anode protection.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.