A 10‐μm Ultrathin Lithium Metal Composite Anodes with Superior Electrochemical Kinetics and Cycling Stability

Lithium metal is a promising candidate for the promotion of the next generation high energy density batteries. The employment of ultrathin Li metal anode with controllable thickness could enable a higher efficiency of Li utilization. Herein, a simple method to fabricate free‐standing 10 μm ultrathin Li metal anode is developed in this work. A three‐dimensional MnOx‐coated CNT framework is constructed through a facile hydrothermal process, utilizing as a host for molten Li infusion, which could not only put forward a simple strategy to modulate the thickness of Li metal film but also restricts the volume expansion. The abundant MnOx nanoparticles acting as lithiophilic sites reduce the Li nucleation barrier and optimize the electrochemical kinetics at the anode/electrolyte interface. As a result, the ultrathin Li composite anode exhibits a superior lifespan expanded to 2000 cycles in a symmetric cell, as well as a better capacity and rate capability than that of bare Li anode in full cell, fulfilling the requirements of high energy density and stable cycling life. Furthermore, a wave‐shaped Li metal pouch cell based on the ultrathin Li composite anode is assembled that exhibits remarkable mechanical bending toleration and cyclic stability, demonstrating large potential application in the field of flexible wearable devices.

Lithium metal is a promising candidate for the promotion of the next generation high energy density batteries. The employment of ultrathin Li metal anode with controllable thickness could enable a higher efficiency of Li utilization. Herein, a simple method to fabricate free-standing 10 μm ultrathin Li metal anode is developed in this work. A three-dimensional MnO x -coated CNT framework is constructed through a facile hydrothermal process, utilizing as a host for molten Li infusion, which could not only put forward a simple strategy to modulate the thickness of Li metal film but also restricts the volume expansion. The abundant MnO x nanoparticles acting as lithiophilic sites reduce the Li nucleation barrier and optimize the electrochemical kinetics at the anode/electrolyte interface. As a result, the ultrathin Li composite anode exhibits a superior lifespan expanded to 2000 cycles in a symmetric cell, as well as a better capacity and rate capability than that of bare Li anode in full cell, fulfilling the requirements of high energy density and stable cycling life. Furthermore, a wave-shaped Li metal pouch cell based on the ultrathin Li composite anode is assembled that exhibits remarkable mechanical bending toleration and cyclic stability, demonstrating large potential application in the field of flexible wearable devices.
At the same time, the specific surface area of anode is increased, which reduces the local current density and promotes the uniform distribution of Li ion at the interface, thus lessening the side reactions caused by the contact between lithium dendrites and electrolyte. [27][28][29] The 3D host materials are based on conductive frameworks with high electrochemical stability, commonly including porous metal substrates, carbon fiber cloth and graphene films. [30][31][32] Nevertheless, the 3D host materials usually have high nucleation barriers and poor affinity for lithium ions. [33][34][35] Lithophilicity is an important factor to achieve uniform lithium deposition. [36][37][38] In previous reports, lithiophilic transition layers, such as metal oxides, metal-organic frameworks, amorphous silicon, and heteroatom doping, are utilized to decorate the surface of 3D frameworks, which can generally be alloyed with Li and reduce the nucleation barrier, thus improving the lithophilicity of 3D frameworks. [39][40][41][42][43][44][45][46] As a result, it is of great importance to design an ultrathin lithiophilic 3D host for Li metal anode with superior performance.
In this present work, we fabricated a lithiophilic manganese oxidecoated carbon nanotube (MnO x /CNT) framework film via a facile hydrothermal method, which acting as 3D host to prepare a freestanding ultrathin (~10 μm) lithium metal anode by a molten Li infusion process. The MnO x /CNT framework serves as Li host reducing the local current density and softening the anode deformation during cycling process. In this case, the amorphous manganese oxide adhered to the surface of the carbon nanotubes (CNT) significantly improves the lithophilicity of the framework through providing a lower Li nucleation potential barrier, thus inducing uniform lithium deposition. Moreover, the Li composite anode embedded in MnO x /CNT 3D host with an ultrathin thickness of 10 μm is generated by molten-infusion method, which exhibits a low loading and enables not only a high Li utilization but also a rather balanced N/P ratio. Under the synergistic effect of the above advantages, uniform Li deposition can be easily obtained on such prepared Li composite electrode with the volume expansion close to zero. Hence, Li||Li symmetric cell with such prepared Li composite ultrathin electrode displays stable striping/plating cycling of 400 cycles at a current density of 0.5 mA/cm 2 and a charge/discharge depth of 18%. To demonstrate the potential of Li composite anode in a practical battery application, full cells were assembled by pairing Li composite anode and LiFePO 4 (LFP), LiCoO 2 (LCO), and LiNi 1/3 Mn 1/3 Co 1/3 O 2 (NCM111) respectively and exhibited favorable capacity retention and Coulombic efficiency. Furthermore, a flexible wave-shaped pouch-cell is assembled that exhibits remarkable mechanical bending toleration and cyclic charge-discharge stability, demonstrating large potential application in the field of flexible wearable devices.

Results and Discussion
Carbon nanotubes were synthesized by a floating catalyst chemical vapor deposition (CVD) and then were transferred into films by a vacuum filtration method. Manganese oxides are conformal coated on carbon nanotube surface via a facile one-step hydrothermal process, which is then dipped into molten Li to fabricate freestanding ultrathin Li metal anode ( Figure 1). The CNT films were firstly treated using an ultraviolet light and ozone (UVO) device in order to produce oxygencontaining functional groups on the surface, thus improving the surface wettability. As previously reported [47][48][49] the CNT surface was decorated by manganese oxide nanoparticles spontaneously when soaked in the neutral KMnO 4 solution through the following redox Reactions (1): Logically, the carbonate ions from the Reaction (1) remain in the solution and will be washed off in the next step but MnO x precipitates settle on the nearest CNT surface. As shown in the scanning electron microscopy (SEM) images reflecting top view of the CNT and MnO x / CNT films (Figure 2a  which will result in a thorough coverage of the free CNT surface under a sufficient time condition. Besides, no long-range ordered lattice fringes can be observed as shown in the amorphous coated areas ( Figure 2e). Figure 2f shows two diffuse rings in the selected area electron diffraction (SAED) pattern, representing a typical amorphous diffraction pattern, which is in accordance with X-ray diffraction (XRD) patterns ( Figure S1). The MnO x /CNT was further texted by Raman spectroscopy, as shown in Figure S2. The intensity ratio of D band and G band (I D /I G ) indicates the disorder degree of graphite. In this work, compared with untreated CNT, the value of I D /I G ratio calculated from Raman spectrum of MnO x /CNT increases from 0.27 to 1.21 because the density of defects grows and the ordered degree decreases due to the coating of amorphous MnO x on the surface of CNT. From Figure 2g-i, the energy dispersive spectrometer (EDS) mapping confirms the coating nanoparticles are composed of Mn and O elements and loaded through the entire interior as well. X-ray photoelectron spectroscopy (XPS) analysis was performed to distinguish the oxidation states of amorphous MnO x nanoparticles. Generally, the Mn 3 s peak has two multiplet split components which are caused by coupling of non-ionized 3 s electron with 3d valence-band electrons. [50] In this work, the binding energy separation between Mn 3 s doublet peaks is The preparation of Li composite electrode was performed in an argon-filled glove box. At first the polished solid Li metal was heated to 300°C on a hotplate to form molten liquid Li, and then the edge of the pre-heated MnO x /CNT film was contacted with the molten Li. The molten Li could spontaneously spread from the edge to the entire surface in seconds due to the superior wetting action of MnO x .
After the molten-infusion treatment, the Li anode embedded in a 3D-MnO x /CNT film presents a surface-shining metallic luster. Compared with the MnO x /CNT film, the color of the Li-MnO x /CNT turns from black into yellowish-gray (Figure 3a,b), which means the Li spreads all over the MnO x /CNT film and combines with it to form a Li  Figure 3c, the surface morphology of Li composite anode formed after lithium infiltration is affected by the structure of MnO x /CNT framework. The molten Li solidifies and deposits between the MnO x /CNT fibers or constructs a new composite cladding layer, forming a rather rough surface compared with bare Li foils. The thickness of Li composite anode is around of 10 μm, which is much thinner than the current commercial Li metal anode (about 100-500 μm). And the internals seem well-distributed and compacted as the filled Li is closely combined with the MnO x /CNT composite bundles which play the role of "host" in it (Figure 3d). The accumulated Li is well embedded in the MnO x /CNT framework. The enlarged SEM image of Li composite shown in Figure S4 displays that Li is completely filled inside the MnO x /CNT framework without redundant Li layers on the surface.
The half cells were assembled with the as-prepared 10 μm Li composite anode used as working electrodes and commercial Li foil as anode and reference anode. The area-specific capacity measured at a current density of 0.1 mA/cm 2 was 1.67 mAh/cm 2 , as shown in Figure 4a. Nyquist plots of symmetric cells assembled with bare Li electrodes and Li composite electrodes were obtained by electrochemical impedance spectroscopy (EIS) test to characterize the interfacial impedances in the initial state and after 10 cycles at a current density of 1 mA/cm 2 . The semicircular radians of the Nyquist plots in the highfrequency region represent the interface resistance between Li/SEI/electrolyte. [51] In the initial state, the impedance of the Li composite electrode is around 70 Ω, while that of the bare Li electrode is close to 300 Ω. After 10 cycles at a current density of 1 mA/cm 2 , the semicircular radian in the high-frequency region was significantly reduced as the SEI was gradually formed. The impedance of the Li composite electrode dropped below 20 Ω while the corresponding resistance of the bare Li electrode still exceeds 120 Ω (Figure 4b,c).
The electrochemical performance of ultrathin Li composite electrodes at different current densities and different depths of discharge was tested with a symmetric cell configuration in ether-based electrolytes. From Figure 4, the battery was tested at a current density of 1 mA/cm 2 and a fixed capacity of 0.1 mAh/cm 2 , and the ultrathin Li composite electrode still exhibited a stable polarization voltage after 2000 cycles with an average of overpotential less than 20 mV. Tested at a current density of 0.5 mA/cm 2 and a fixed capacity of 0.3 mAh/cm 2 , the corresponding depth of discharge is about 18%, and the average overpotential is less than 50 mV after more than 350 cycles. Lithium metal is a kind of "hostless" electrode material. The major reaction occurring on the Li metal anode during the charging and discharging process is the deposition/dissolution of Li. The unstable factors in the electrochemical reaction process, such as current density, can easily interfere with the progress of the Li deposition and dissolution, thereby resulting in the formation of dendritic Li and the generation of irreversible "dead Li". The continuous dissolution and deposition of Li bring about large volume changes during charge and discharge, which could cause the solid electrolyte interphase (SEI) to repeat the burst-rebuild process, steadily expending the electrolyte. Meanwhile, the instability of SEI in turn accelerates the formation rate of Li dendrites. Further electrochemical measurements of symmetric Li composite anode and bare Li cells were conducted with a fixed areal capacity of 0.1 mAh/cm 2 at different rates of 0.2, 0.5, 1, and 2 mA/cm 2 . As shown in Figure S7, the 10 μm ultrathin Li composite anode exhibits a smoother voltage platform and lower average overpotential compared with the 30 μm bare Li electrode with increasing current density. As the current density increased to 2 mA/cm 2 , the voltage hysteresis of Li composite electrode is 49 mV, which is in sharp contrast to the 159 mV of Li electrode.
To further investigate the relationship between long-term cycling stability and electrode volume change, the morphology was checked (Figure 5a-d) to show the microstructures after 50 cycles at a fixed current density of 0.5 mA/cm 2 with a fixed areal capacity of 0.3 mAh/cm 2 . After 50 cycles, the thickness of the Li composite electrode did not change significantly and remained at about 10 μm while the bare Li electrode expanding to 1.5 times its original thickness as shown in the cross-section view (Figure 5b,d). showing the stable electrochemical Li deposition/dissolution process. It is proved that exploiting the MnO x /CNT host could provide space for the dissolution and deposition of lithium in the process of charge and discharge, in which way the volume expansion of anode is effectively softened. Besides, there is no obvious Li dendrite formation on the surface, and the MnO x /CNT skeleton structure is basically intact. In order to compare the reaction kinetics at the interface of Li composite electrode and bare Li electrode, the exchange current densities of Li composite electrode and bare Li electrode were measured. As the concentration polarization is insignificant, Tafel equation was employed to calculate the exchange current density according to the Formula (2): where i 0 is the exchange current density, μ is the overpotential, and A is a dynamic constant. [52] The values of exchange current density are acquired at the point where the overpotential equals 0 mV by extending the linear parts of the Log i-Voltage curve shown in Figure 5d. The Li composite anode exhibits an exchange current density of 76 μA/cm 2 which is nearly 2.5 times higher than the 29.2 μA/cm 2 of bare Li. The higher exchange current density signifies a faster transfer at the electrode/electrolyte interfaces, meaning a reduced electrochemical polarization of the Li composite electrode, which is consistent with the subsequent experimental results. The electrochemical reaction kinetics can be significantly ameliorated by the faster charge transfer at the electrode/electrolyte interface. According to the Butler-Volmer electrode kinetics relationship, the Li composite anode with an enhanced exchange current density possesses a smaller chargetransfer resistance, which represents a better electrochemical kinetics behavior. The enhanced exchange current density can be attributed to the uniform distribution of the MnO x /CNT framework, and the amorphous manganese oxide coating reduces the Li nucleation barrier and induces uniform Li deposition with a lower overpotential. To investigate the lithiophilic/lithophobic difference between MnO x /CNT and CNT, the binding energy between CNT or MnO x /CNT and Li was calculated based on density functional theory (DFT) calculation as follows: As shown in Figure 5e,f, the binding energy value of MnO x /CNT (−2.78 eV) and lithium is more negative than the uncoated CNT (1.18 eV), indicating a significant surface chemistry conversion from lithophobic into lithiophilic. It is demonstrated that the Li deposition prefers to occur on the MnO x coatings according to the DFT simulation and then spread along the host, which is also confirmed by the wetting behavior of molten lithium that is characterized by observing the contact angle between the molten lithium and the sample ( Figure S6). Apparently, the contact angle of MnO x /CNT is about 35°, which is much smaller than that of CNT (~130°), indicating that amorphous manganese oxide coating can greatly improve the lithophilicity of CNT framework and reduce Li nucleation barrier which is beneficial to Li growth. In order to evaluate the electrochemical performance of the 10 μm ultrathin Li composite electrode, the cycling performances of full cells assembled with LiCO 2 (LCO), LiNi 1/3 Co 1/3 Mn 1/3 (NCM), and LiFe-PO 4 (LFP) as cathode and 10 μm Li composite or 30 μm bare Li as anode were tested. As shown in Figure 6a-c, the rate performances of LCO||Li, NCM||Li, and LFP||Li cells are compared from 0.5 C to 5 C. The discharge-specific capacities of the LCO||Li composite anode cells are 149.1, 143.8, 139.1, 131.9, and 120.5 mAh/g at 0.5, 1, 2, 3, and 5 C respectively. In comparison, the specific capacities of the LCO||bare Li cell are 130.7, 100.0, 78.2, 54.6, and 29.9 mAh/g at 0.5, 1, 2, 3, and 5 C respectively. Similarly, the NCM||Li composite cell offers a specific capacity of 130.6 mAh/g at 0.5 C, which retains 94.2 mAh/g when the current density increases to 5 C. However, the NCM||bare Li cell offers specific capacities of 114.0, 110.4, 100.9, 93.7, and 78.1 mAh/ g at 0.5, 1, 2, 3, and 5 C respectively. In the meanwhile, the LFP||Li cell also exhibits similar results. It is apparently observed that at all tested current densities with the same kind of cathode, the Li composite anode exhibits considerably higher reversible specific capacities than the bare Li electrode, especially presenting the largest capacity difference at higher current densities, which could be attributed to the faster kinetics of the Li composite electrode. The galvanostatic cycling performance of LCO||Li, NCM||Li, and LFP||Li cells are tested and compared in  Figure S8. The voltage hysteresis of Li composite anode full cells is obviously smaller than that of bare Li anode full cells, exhibiting the improved cycling stability, which could be attributed to the limited electrolyte consumption caused by the limited volume expansion of anode and the reduced dead Li resulted from the abundant lithiophilic sites.
Considering the ultrathin Li composite anode shows better mechanical deformation capability, we employed an imprint method to construct a wave-shaped flexible Li metal battery which contains many small arcs supplying sufficient deformation space to accommodate compression strain generated during the bending process. The wave-shaped Li pouch cell involving LFP cathode, PVDF-HFP gel electrolyte film, and ultra-thin Li composite anode exhibits excellent elastic tolerance, as schemed in Figure 7a. The capacitance retention of the wave-shaped pouch cell at 1 C as the bending angles increasing from 0°to 90°is present in Figure 7b,c. Under various bending angles the capacitance retention showed negligible changes, which demonstrates a superior mechanical flexibility. It is shown in Figure 7d that as the wave-shaped pouch cell bent to 90°for 1000 times and executed a charge-discharge process at 1 C after every 20 times of deformation, the capacitance retention is well maintained as high as 94.6%. From EIS tests in Figure S9, the Ohmic resistance is significantly decreased after folding 1000 times which further confirm the stable condition of the bendingtested cell. The superior bending toleration of Li metal battery is attributed to two aspects. On the one hand, the wave-shaped device structure provides a space to accommodate the stress produced in the bending process. [53] On the other hand, the ultrathin Li composite anode shows remarkable mechanical deformation capability due to the 3D host and reduced thickness. Based on the above results, it is convincing that the ultrathin Li composite anode shows great promise in the field of flexible wearable devices.

Conclusion
In summary, a 10-μm ultrathin lithium composite anode based on manganese oxide-coated carbon nanotube film as lithium host is fabricated via a lithium infusion process. The MnO x /CNT host prepared by a hydrothermal process provides space for the dissolution/deposition and accommodates volume expansion of Li metal anode in the charging/discharging process. Meanwhile, density functional theoretical (DFT) calculation exhibits abundant MnO x nanoparticles acting as lithiophilic sites to reduce the Li nucleation barrier and optimize the electrochemical kinetics at the anode/electrolyte interface. As a result, the ultrathin Li composite electrodes in the symmetric cells experienced 2000 stable cycles at 1 mA/cm 2 and 0.1 mAh/cm 2 and more than 400 stable cycles at 0.5 mA/cm 2 and 0.3 mAh/cm 2 , exhibiting extraordinary cycling stability. Besides, the full cells with LCO, NCM, LFP as cathode respectively combined with the 10 μm ultrathin Li composite anode exhibit higher capacity, better rate performance, and improved cycling stability than those with the 30 μm bare Li anode. Furthermore, a wave-shaped Li metal pouch cell based on ultrathin Li composite anode is assembled that exhibits superior capacity and excellent mechanical bending toleration, demonstrating that this ultrathin Li composite anode not only fulfilling the requirements of high energy density and stable cycling life, but also shows great promising in wearable electronics.

Experimental Section
Preparation of MnO x /CNT framework: Carbon nanotubes (CNTs) were synthesized via a floating catalyst chemical vapor deposition (CVD) approach and then were transferred into films by a vacuum filtration method. The original CNTs were sheared and then immersed into dilute hydrochloric acid (5 mol/L) for 24 h. The as-purified CNTs were dispersed into deionized water by ultrasonic treatment with aid of cetyltrimethylammonium bromide (CTAB) dispersant. The CNTs solution was vacuum filtered through a cellulose membrane and washed using deionized water repeatedly. After drying, the cellulose membrane was peeled off easily, and the freestanding CNT film was obtained. The preparation of MnO x /CNT framework is realized via a facile hydrothermal reaction method. Preparation of PVDF-HFP film: At first, 3 g PVDF-HFP pellet and 0.3 g glycerol were added in 10 mL NMP, kept in an airtight volumetric flask and stirred at 50°C for 12 h. Then we acquired a uniform transparent solution and let it stand for 30 min to remove bubbles. After that, the transparent solution was cast on the glass plate with a thickness of 200 μm and dried in the vacuum oven at 100°C for 24 h. The PVDF-HFP film was transferred to the glove box and then soaked in 1 M LiTFSI-EC/PC/EMC (1:1:1, v/v/v) electrolyte for 30 min.
Material characterization: Scanning electron microscopy images and EDS mapping were recorded with a Zeiss SIGMA field emission scanning electron microscope attached with an energy-dispersive EDX. TEM images were obtained by a JEM-2100 plus transmission electron microscopy. X-ray diffraction patterns were collected on a D8 Advance instrument while Raman spectroscopy was conducted using a Renishaw inVia system with an excitation wavelength of 532 nm. XPS analysis was carried out on a K-Alpha + scanning XPS microprobe. Digital images were recorded using a Canon EOS 80D camera.
Electrochemical characterization: In order to explore the areal specific capacity of Li composite, a constant current density of 0.1 mAh/cm 2 was applied on half cells assembled with Li composite cathode and Li foil as the anode. To investigate the cycling stability of prepared ultrathin Li metal electrodes, symmetric cells were assembled with Li composite anode or 30 μm Li foil respectively. The electrolyte for symmetrical cells and half cells was 1.0 M lithium bis (trifluoromethanesulfonyl) imide (LiTFSI) in 1,3-dioxolane (DOL)/1,2dimethoxyethane (DME) (v/v = 1:1) with 1 wt% LiNO 3 additive. The amount of electrolyte for each cell was 60 μL. Electrochemical impedance spectroscopy (EIS) was carried out in the range from 100 kHz to 0.1 Hz at the open-circuit potential to study the interfacial transport behavior of the Li composite anode compared with commercial Li foil anode after different cycles of the symmetric cell. Tafel curves were acquired by analyzing the linear sweep voltammetry (LSV) testing results between −0.2 and 0.2 V at a scan rate of 1 mV/s. EIS and LSV testing were conducted on a BT2000 Arbin electrochemistry workstation. The voltage profiles of symmetrical cells were tested at the current density of 0.5 and 1.0 mA/cm 2 . Full cells utilized Li composite anode or 30 μm Li foil directly as anode paired with LFP, LCO, and NCM111 electrodes correspondingly as cathodes. The working electrodes were fabricated by mixing active materials, carbon black and polyvinylidene fluoride (PVDF) in N-methyl pyrrolidinone (NMP) solvent with a mass ratio of 8:1:1 followed by a 12-h drying at 100°C and attached to a carbon-coated Al foil. The active mass loading of LFP, LCO, and NCM111 electrodes was ≈2.3, 2.5, and 2 mg/cm 2 respectively. Commercial 1 M lithium hexafluorophosphate (LiPF 6 ) in dimethyl carbonate (DMC): ethylene carbonate (EC): ethyl methyl carbonate (EMC) (v:v:v = 1:1:1) was used as the electrolyte. The amount of electrolyte for each cell was 50 μL. The galvanostatic discharge/ charge measurements were conducted on a BTS Neware Battery system at a cut-off voltage of 2.3-3.8 V for Li||LFP cells, 3.0-4.35 V for Li||LCO cells and 2.8-4.3 V for Li||NCM111 cells correspondingly. All of the coin cells (CR2032, Kejing) mentioned above were assembled inside an argon-filled glove box with H 2 O and O 2 both less than 0.01 ppm. In order to assemble the pouch cells, the MnO x /CNT films were dipped into molten Li and then cut into the size of 40 mm × 20 mm as same as the single-faced LFP-coated cathode. PVDF-HFP electrolyte film (45 mm × 25 mm) was employed to assemble the pouch cells. At last, the pouch cells were modeled into wave-shaped by using a cold-compacting device.