Construction of Dynamic Alloy Interface for Uniform Li Deposition in Li-Metal Batteries

It is well accepted that a lithiophilic interface can eﬀectively regulate Li deposition behaviors, but the inﬂuence of the lithiophilic interface is gradually diminished upon continuous Li deposition that completely isolates Li from the lithiophilic metals. Herein, we perform in-depth studies on the creation of dynamic alloy interface upon Li deposition, arising from the exceptionally high diﬀusion coeﬃcient of Hg in the amalgam solid solution. As a comparison, other metals such as Au, Ag and Zn have typical diﬀusion coeﬃcients of 10-20 orders of magnitude lower than for Hg in the similar solid solution phases. This diﬀerence induced compact Li deposition pattern with an amalgam substrate even with a high areal capacity of 55 mAh cm-2. This ﬁnding provides new insight into the rational design of Li anode substrate for the stable cycling of Li metal batteries


Introduction
10] On the other hand, Pzone metal anodes (Si, Sb, Sn, Ge. ..) possess high capacities with little risks of dendrites due to the inward diffusion of deposited Li atoms driven by the essential alloying reactions. [11- 18]26][27][28] Certain substrate with definite solubility in Li is preferable for reducing Li deposition barriers. [29]For example, the formation of solidsolution alloy at the Li-rich region allows much reduced Li nucleation potential for Au substrate. [30]The doping of small amount of Mg into Li substrate also greatly alters the surface energy and results in distinct deposition morphologies from pure Li substrate. [31,32]Nonetheless, the substrate influence is gradually decayed during Li accumulation due to the fact that the original surface is gradually covered by deposited Li, and thus, the uncontrolled morphologies remain unavoidable in practical Li metal cells with large areal capacities and high current densities.
[35] For example, the alloying reaction is readily processed by casting Hg drops onto Li foil without further treatment at room temperature. [36]The Li-Hg composite anodes with Li amalgam on the top and Li metal at the bottom afford fast adsorption of Li deposition in electrochemical cells, that is, the elimination of dendrite morphologies at high current densities.In fact, the amalgam layer can be easily peeled off from Li foils and served as an alloy substrate due to the flexible, self-supported, and ultrathin characteristics.Unlike regular metal substrates, an Li-Hg interface is kinetically maintained even with a high plating capacity, which is a distinct amalgam from other metal substrates that can only form Li-M (M = metal) at the surface.The uniquely super-fast diffusion of Hg atoms in the Li-rich Li-Hg solid-solution phase affords fresh amalgam interfaces during electrochemical Li deposition.

Results and Discussion
To study different Li-M solid solution formation and their influence on Li deposition, we first compared binary Li-M phase diagrams with selected metals Hg, Ag, Au, and Zn, respectively.Phase diagrams in Figure 1a confirm solid-solution compositions in the range of Li atomic ratio near 100%. [37]With reduced Li concentration in the alloys, intermetallic compounds are formed such as Li 3 Hg, LiAg, Li 3 Au 4 , and LiZn.The solubility of Li in the metal substrates significantly improves the deposition interfaces and leads to much reduced overpotential upon Li nucleation. [30]On the other hand, the substrates can exhibit distinct properties after the initial lithiophilic interfaces are covered by deposited Li.A key factor is their different melting points.The much lower melting point (−38.8°C) of Hg is thermodynamically desirable for atom self-diffusion in alloys.The assembly of the cells comprising Li metal and various substrates is presented in Figure 1b, in which Li foil was used as the counter electrode and the substrate for Li deposition.The Ag, Au, and Zn thin foils were purchased and cut into desired size for use.LiHg film electrodes were synthesized by the simple drop-casting method. [36]5 μL Hg was coated onto 10 cm −2 pure Li foil to spontaneously form thin alloy film (about ~7 μm in thickness).LiHg film has a lightweight areal density of 7 mg cm −2 , and thus, the composition is calculated to be Li 0.46 Hg 0.54 (Figure S1, Supporting Information).
In Figure 1c, the LiHg film shows a two-step feature in the curve.The onset of voltage sharp drop is attributed to the vast Li deposition on the surface of the LiHg electrode at a large current density of 2 mA cm −2 .However, the voltage is rapidly recovered to above 0 V versus Li + /Li due to the formation of Li-Hg alloys.After Li is saturated in the amalgam, Li deposition appears as indicated by the flat curve below 0 V. Thus, LiHg substrate exhibits three distinct steps upon Li deposition: the initial Li nucleation, the subsequent alloying between Li and LiHg, and the Li deposition on the lithiated amalgam.This feature is not presented for other metal substrates, which only have the sharp dip and the following flat curve below 0 V.[40] The unique Li deposition process on the amalgam substrate is correlated to the fast Li-Hg alloying reaction that allows immediate Li diffusion into the amalgam phase above 0 V.This assumption is also verified by the violent alloying upon the contact between fresh Li and Hg even at room temperature when preparing the LiHg film substrate (Figure S2, Supporting Information).
We performed in-depth studies to understand the Li-Hg alloying process.Figure 2a shows cross-sectional scanning electron microscope (SEM) images of LiHg film with different Li deposition capacities (0.0, 2.3, 3.1, 5.0, 9.0, 55.0 mAh cm −2 , respectively) at 2 mA cm −2 .A two-layer morphology is clearly shown with the Li capacity of 2.3 mAh cm −2 .The cylindrical morphology might correlate to the high Hg concentration that alters the nucleation environments of deposited Li. [41] In Figure 1c, the lithiation of LiHg film is completed above 1.5 mAh cm −2 , followed by Li deposition on the amalgam surface.The two-layer morphology of the electrode confirms the electrochemical process.Both layers gradually expanded upon the increase of Li deposition capacities, but the compact electrode feature is maintained even with a high volume of 55 mAh cm −2 .The growth of the bottom layer indicates that it was not simply a lithiation-deposition process.
Instead, the amalgam layer was actively involved in the Li deposition even after Li was saturated in the amalgam solid-solution phase to yield the compact Li layer with smooth surface (Figures S3-S5, Supporting Information).The structures of deposited electrodes were analyzed with ex situ XRD shown in Figure 2b and  Energy Environ.Mater.2024, 7, e12618 nucleation and accumulation of lithium atoms on the surface of the amalgam.Combined electrochemistry, SEM, and XRD results, the alloying process (stage I) is the phase transition from LiHg 3 to Li 3 Hg, while the subsequent Li deposition gives high compact Li layer with the involvement of Hg (stage II) (Figure 2c).
As discussed above, it is highly possible that the Li deposition process on the top of the amalgam layer is affected by Hg to obtain compacted growth of Li layer.To examine this assumption, energy-dispersive spectroscopy (EDS) mappings, HR-TEM, and Cs-corrected TEM were employed to explore the composition and structure of the Li layer (Figure 3).The electrodes with a Li capacity of 5 mAh cm −2 were selected for the analysis of the compositions of the Li layer.First, the electrode was examined by SEM and EDS mapping.The EDS mapping clearly confirms the existence of Hg atoms in the Li layer at micron scale (Figure 3a).Next, a sample was carefully prepared by FIB (Figure S7, Supporting Information) obtained from the top layer of the electrode.The average thickness of the sample is ~3 μm to avoid any contamination from the bottom Hgcontaining layer.It is not easy to probe Li by  EDS, but the electron energy loss spectroscopy (EELS) spectrum still shows strong Li signals at 56.0 eV in this sample (Figure 3b). [42,43]PS spectra confirm Hg 4f peaks at 101.9, 103.9 eV as well as Li peak at 54.9 eV.The peaks can be assigned to the metallic mercury and lithium (Figure 3c,d) with the atomic ratio of 1:99 for Hg:Li (Figure S8, Supporting Information). [44,45]In addition, HR-TEM images of the sample further confirm that Hg is distributed in the Li layer (Figure 3e-h).All the combined results suggest that Hg can diffuse upward along with the electrochemical deposition by forming a dynamic alloy interface to regulate Li deposition pattern, as shown in the SEM images in Figure 2.
In contrast, other selected substrates (Ag, Au and Zn foils) exhibit different deposited Li morphologies under the same electrochemical condition (2 mA cm −2 , 5 mAh cm −2 ).The porous and loosely packed dendrites are clearly observed on the surface of these substrates upon Li deposition, and there is little Ag, Au, and Zn being detected in the deposition layer (Figure 4a-c), a sharp contrast to observation on the LiHg substrate.The absence of Ag, Au, and Zn is also evidenced by the XPS analyses with the deposited Li layer (Figure S9, Supporting Information).These results imply the critical role of surface alloying process in avoiding dendric morphologies.Along with the coverage of the initial interface upon Li plating, the alloying function becomes weaker and dendric morphologies appear for Ag, Au, and Zn substrates (Figure 4d).It should be noted that Hg has negative charge in the Li-Hg structures, [36] and Li deposition pattern is also regulated by the surface Hg atoms despite the small Hg concentration of 1%.
To understand the inherent different lithiation behaviors between Hg and other metals, the diffusion coefficient (D) of each metal was calculated in the Li-rich Li-M alloys.As shown in the phase diagrams, the Li-rich zone has a solid-solution phase for the selected metals M (Hg, Au, Ag, and Zn).The alloy diffusion theory can be explained with Darken equation in a binary A-B alloy: [46] where A and B refer to Li and M (Ag, Au, Zn and Hg), respectively.D Li and D M are intrinsic diffusion coefficients of Li and M, and x Li and x M are concentrations of Li and M in this study. [47,48]or the Li-rich solid-solution phase, it has x Li → 100% and x M → 0, thus the above equation can be simplified as D = D M (M = Ag, Au, Zn or Hg).D M can be calculated according to the diffusion equation: In which R is the ideal gas constant (R = 8.314 J mol −1 K −1 ), T is the sroom temperature (T = 298 K), D 0 is the frequency factor of the atom, and Q sd is the diffusion activation energy, respectively. [49,50]D 0 and Q sd of Ag, Au, Zn, and Hg atoms are shown in Table 1, respectively. [46]By Equation ( 2), the diffusion coefficients of Ag, Au, Zn, and Hg atom are calculated as 1.8 × 10 −31 , 1.0 × 10 −30 , 1.1 × 10 −15 , and 1.55 × 10 −3 mm 2 s −1 , respectively.It is clearly that D Hg is 28, 27, and 12 orders of magnitude higher than for D Ag , D Au , and D Zn , respectively.Meanwhile, the radial distance R (or net displacement) of an atom can be estimated by the following equation: where t represents diffusion time (s), D is the interdiffusion diffusion coefficient (mm 2 s −1 ). [51]Thus, the radial distance R of each metallic atom can be up to 4.2 × 10 −16 , 1.0 × 10 −15 , 3.2 × 10 −8 , 3.9 × 10 −2 mm s −1 , respectively.The results reveal that Hg can migrate as much as 3-14 orders of magnitude farther than for other metals (Figure 5a).Ideally, the plating capacity of 2 mAh (2 mA cm −2 current with 1 hr time and 1 cm 2 square condition) corresponds to a thickness of ~9.7 μm.According to the calculated radial distance R, the super-high Hg diffusion coefficient allows a migration distance of 39 μm s −1 , which indicates that Hg can easily migrate from the LiHg substrate to the top electrolyte/electrode interface to mitigate the formation of Li dendrites through the alloying process.Instead, Ag (Au or Zn) has very limited diffusion distance of 4.2 × 10 −13 μm (1.0 × 10 −12 or 3.2 × 10 −5 μm) within the same time.The significant difference on metal diffusion coefficient reveals that these metals can only regulate Li plating patterns with small current density and capacity, while Hg substrate is capable of evolving to a dynamic Li-Hg interface for uniform Li deposition (Figure 5b).Further studies on plating/stripping cycling were compared between Ag, Au, and Zn substrates and LiHg film at 2 mA cm −2 and areal capacity of 2 mAh cm −2 .The upper voltage was set 1.0 V when evaluating the Coulombic efficiencies.The cell with a LiHg film shows stable stripping/plating over 400 h, and the Li|LiHg film cell exhibits superior Coulombic efficiency (100%) over 160 cycles (Figure 6a-c).The stable cycling and Coulombic efficiency of LiHg film indicates efficient Li deposition and stripping process.In contrast, the Li|Zn cell has much reduced cycle life along and unstable Columbic efficiency in the initial cycles (Figure 6d).Similar results are also obtained for the cells assembled with Ag and Au substrates under the same test conditions (Figure S10, Supporting Information).Full cells were also assembled with a high LiFePO 4 (LFP) loading of 12 mg cm −2 .LiHg film|LFP full cell delivers a stable capacity about 100 mAh g −1 after 100 cycles at a high rate of 2 C (1 C = 170 mA g −1 ), while the Zn foil|LFP cell shows obvious capacity decay only in 40 cycles (Figure 6e-g).With the tailored electrolyte window (2.0-3.8V), the cycling of LFP full cell is further extended to 200 cycles (Figure S11, Supporting Information), indicating the great potential of the strategy for practical applications.Energy Environ.Mater.2024, 7, e12618

Conclusion
In summary, we propose a new concept of dynamic alloying interfaces to regulate Li deposition pattern in this research.The exceptionally high diffusion coefficient of Hg allows its rapid migration from the substrate to the top electrolyte/electrode interfaces, thus effectively mitigating the formation of Li dendrites at high plating current (2 mA cm −2 ) and capacities (5 mAh cm −2 ).This unique feature is distinct from other metals such as Au, Ag, and Zn, which only work with the initial Li plating onto the substrates.This study provides new insight into the rational design of the plating substrates.Metals with low melting points such as Hg, Ga, and In may play a critical role in tailoring the substrate compositions.Further progress is expected with these metals for the design of binary or ternary alloy substrates or accurate control of coating species on regular Cu substrates.

Experimental Section
Materials: The pristine self-supported LiHg film was obtained by coating method in an argon-filled glovebox with <0.1 ppm O 2 and H 2 O.The preparation of LiHg film was as follows: 1) mercury droplet was spread onto the pure lithium metal to spontaneously form LiHg film by a brush; 2) this film was peeled off the pure lithium metal, the thickness of the film is about 7 μm; 3) the obtained thin film was cut into squares of 1 × 1 cm 2 by a scissor as electrode.After weighing of this self-supported film, the surface load of LiHg film is 7 mg cm −2 , and the surface load of Hg was 6.8 mg cm −2 (5 μL liquid phase Hg was coated on 10 cm −2 Li foil), thus the surface load of Li was 0.2 mg cm −2 .The commercial Ag, Au, and Zn foils have the same 1 × 1 cm 2 squares by a scissor with LiHg film, which possess thickness of 10, 2, and 10 μm, respectively.LiFePO 4 (LFP) cathode material was provided by Lishen company with mass loading of 12 mg cm −2 .Physical characterization: Scanning Electron Microscope (SEM) studies were carried out on Environment Scanning Electron Microscope with a field emission gun (Quanta FEG 250).X-ray diffraction (XRD) patterns were collected on a Rigaku Miniflex 600 desktop at 40 kV and 20 mA (Cu K α radiation).Transmission Electron Microscope (TEM) analysis was used on a High-Resolution Transmission Electron Microscope with FEG (Talos F200 X) at 200 kV (HR-TEM) and Transmission Electron Microscope with A Probe Corrector (Titan Themis Cubed G2 60-300) (Cs-TEM).Focused Ion Beam (FIB) sample was synthesized on FIB-SEM system (Helios Nanolab 460HP) at 5 and 30 kV (Ga ion beam etching).X-Ray Photoelectron Spectroscopy (XPS) characterization was measured with an X-Ray Photoelectron Spectroscopy (Escalab 250Xi) spectrometer.Optical images were conducted with a Keyence VHX-950F microscope and cell phone.Electrochemical evaluation: Coin half cells and full cells were assembled in an Ar-filled glove box (O 2 and H 2 O level below 0.1 ppm).In half cells, Ag, Au, and Zn foils and LiHg film acted as working electrodes and pure lithium metal (thickness of 600 μm) as counter electrode.In full cells, LFP electrode worked as cathode (12 mg cm −2 of mass load), the LiHg film and Zn foil anodes had predeposited Li capacity of 4 mAh cm −2 .The electrolyte was 1.0 M lithium bis-(trifluoromethane-sulphonyl)imide (LiTFSI) in 1,3-dioxolane (DOL) and 1,2dimethoxyethane (DME) (1:1 vol/vol) without any additives.About 100 μL of electrolyte was used in each coin cell to standardize the experiment.The three piece of separators with Al 2 O 3 coating layer (Celgard 2400) were placed in the middle of two electrodes to protect from short circuit in the cells.All electrochemical tests were carried out on LAND-CT 2001A multichannel battery tester (Wuhan, China) at room temperature.Electrochemical impedance spectroscopy (EIS) measurements were conducted on a Princeton PARSTAT 2273.The frequency ranged from 1 MHz to 0.01 mHz with an alternating voltage signal amplitude of 5 mV.
Figure S6, Supporting Information.The asprepared LiHg film consisted of intermetallic compounds of LiHg 3 , Li 3 Hg, and trace amounts of metallic Li or Li-rich Li-Hg solid solution.The initial lithiation converts LiHg 3 into Li 3 Hg within 10 min deposition (Li capacity of 0.34 mAh cm −2 ) and has completely evolved sole Li 3 Hg phase after 20 min (~0.67 mAh cm −2 ).This fast phase transformation is crucial to guide the initial Li diffusion into the inner phase at a high current density.With increased Li plating time to 80 min (~2.70 mAh cm −2 ), Li 3 Hg is the main phase in the XRD pattern as indicated by the 23.520°, 27.220°, and 38.860°peaks, corresponding to the (111), (200), and (220) diffractions.The growth of the characteristic diffraction peaks indicates that the alloying reactions were dominant process in electrodes.It is interesting that the the (111) diffraction is increased with extended lithiation, while the (200) and (220) diffractions become weaker.It might be related to the saturation of the Li 3 Hg phase and the formation of Li 3 + xHg compositions.The diffraction peak at 35.920°is emerged and strengthened in subsequent process (>80 min), suggesting the

Figure 1 .
Figure 1.a) Phase diagrams of Li-M (M=Hg, Ag, Au, and Zn, respectively).The regions where Hg, Ag, Au, and Zn dissolved are denoted as (Li), δ(Li), (Li) and β(Li); b) a schematic of Li|substrate cell and the formation of Li-rich solid solution alloys upon Li deposition; c) Li plating profiles of four cells at 2 mA cm −2 and 5 mAh cm −2 .

Figure 2 .
Figure 2. Morphology and structure evolution with different Li plating capacities at 2 mA cm −2 .a) The cross-sectional SEM images of LiHg electrodes with areal capacity of 0.0, 2.3, 3.1, 5.0, 9.0, and 55.0 mAh cm −2 ; b) the electrochemical profiles and corresponding ex-situ XRD patterns with controlled Li capacities; c) the schematic illustration of morphology evolution of LiHg substrate upon Li deposition.

Figure 3 .
Figure 3. Composition analyses for the sample with a plating capacity of 5 mAh cm −2 .a) The SEM image and corresponding EDS mapping of Hg; b) EELS spectrum of Li signal for the FIB sample; c, d) XPS spectra of Hg and Li of the FIB sample; e, f) HAADF image and corresponding EDS mapping image from HR-TEM; g, h) the low, high HAADF images and corresponding EDS mapping from the Cs-corrected TEM model, where Hg agglomeration is clarified as bright domains in the dark Li zones due to the super energy exposure of Cs-corrected TEM and significantly different atomic weight of the two elements.

Figure 4 .
Figure 4. a-c) The cross-and top-sectional SEM images and corresponding EDS mapping images with Li deposition capacity of 0 and 5 mAh cm −2 : a) Ag foil, b) Au foil, and c) Zn foil.d) The schematic diagram of the morphology evolution from pristine electrode to the subsequent plating process for Ag, Au, and Zn.

Figure 5 .
Figure 5. a) The histogram of theoretical diffusion equal distance R and diffusion coefficient D of Ag, Au, Zn, and Hg in the Li-rich solid solution phases; b) the schematic upon Li atoms deposition on different substrate foils (Ag, Au, Zn foil and LiHg film).

Figure 6 .
Figure 6.The electrochemical performance of cells assembled with Zn foil and LiHg film electrodes with a prelithiation capacity of 4 mAh cm −2 .a) Stripping/plating-time curves at 2 mA cm −2 and 2 mAh cm −2 ; b-d) Coulombic efficiencies and corresponding stripping/plating curves at 2 mA cm −2 and 2 mAh cm −2 ; e-g) LiHg film|LFP and Zn foil|LFP full cell performance: e) cycling performance, f, g) capacity-voltage curves at various cycles.

Table 1 .
The frequency factor (D 0 ) and diffusion activation energy (Q sd ) of Ag, Au, Zn, and Hg atoms.