Recent Advances in Lithiophilic Porous Framework toward Dendrite-Free Lithium Metal Anode

: Rechargeable lithium metal anode (LMA) based batteries have attracted great attention as next-generation high-energy-density storage systems to fuel the extensive practical applications in portable electronics and electric vehicles. However, the formation of unstable solid-electrolyte-interphase (SEI) and growth of lithium dendrite during plating / stripping cycles stimulate safety concern, poor coulombic e ﬃ ciency (CE), and short lifespan of the lithium metal batteries (LMBs). To address these issues, the rational design of micro / nanostructured Li hosts are widely adopted in LMBs. The high surface area of the interconnected conductive framework can homogenize the Li-ion ﬂux distribution, lower the e ﬀ ective current density, and provides su ﬃ cient space for Li accommodation. However, the poor lithiophilicity of the micro / nanostructure host cannot govern the initial lithium nucleation, which leads to the non-uniform / dendritic Li deposition and unstable SEI formation. As a result, the nucleation overpotential and voltage hysteresis increases, which eventually leads to poor battery cycling performance. Thus, it is imperative to decorate a micro / nanostructured Li host with lithiophilic coatings or seeds for serving as a homogeneous nucleation site to guide the uniform lithium deposition. In this review, we summarize research progress on porous metal and non-metal based lithiophilic micro / nanostructured Li hosts. We present the synthesis, structural properties, and the signiﬁcance of lithiophilic decorated micro / nanostructured Li host in the LMBs. Finally, the perspectives and critical challenges needed to address for the further improvement of LMBs are concluded.


Introduction
It has been challenging for lithium-ion batteries (LIBs) to meet the ever-increasing energy demands. The state-of-art LIBs, which are widely used in smart consumable electronics, energy-efficient plug-in-hybrid, and pure electric vehicles energy, are expected to reach the energy density of 300 Wh kg −1 at the cell level [1][2][3]. However, the current market of LIBs has not yet reached a new goal set of 500 Wh kg −1 or higher at the cell level for broad applications that require high energy density, high efficiency, and long life span. To reach such a high goal, Li metal has been regarded as a suitable alternative anode material, owing to its high theoretical capacity of 3860 mAh g −1 , low density of 0.59 g cm −3 , and low redox potential of −3.04 V versus standard hydrogen electrode (SHE) [4,5]. Besides, lithium can involve aggressive chemistries approach such as operating at high voltage and/or pairing of the high capacity for the conversion-reaction in intercalation cathodes. For example, the Li metal anode (LMA) coupled with cathodes such as sulfur (S) and oxygen (O 2 ), as next-generation lithium batteries, can offer splendidly high theoretical specific energy of 2567 Wh kg −1 and 3505 Wh kg −1 , respectively [6]. However, Li anode raises lots of issues such as highly reactive nature with organic electrolytes, relatively infinite volume expansion during Li plating/stripping, uncontrolled dendrite formation that leads to low coulombic efficiency (CE) and safety concerns [7,8]. Tremendous efforts have been made to realize the use of metallic Li as anode material in the battery, which includes (a) developing nano/microstructured Li host to accommodate the Li [8][9][10] and (b) ex-situ or in-situ derived solid electrolyte interphase (SEI) to shied the highly reactive Li metal from the liquid electrolyte [1,2,5,11], (c) use of suitable solid-state electrolytes [12,13], and (d) modification of separators [14,15], etc. However, these result-oriented strategies are mainly concentrated on the dendrite-free Li deposition but less discusses the initial Li nucleation. Recently, the advanced technologies of employing a nano/microstructured Li host for dendrite-free Li deposition have attracted great attention. Three dimensional (3D) nano/microstructured Li host has a high surface area that lowers the local current density and allows a very thin and regular lithium coating, accommodates the volume expansion of Li within the structure and offers lower interfacial resistance [16]. However, most of the metal-based or carbon-based 3D nano/microstructured Li hosts are lithiophobic to the Li that leads to the poor Li wettability. As a result, the nucleation overpotential increases leading to high voltage hysteresis and unfavorable Li deposition. Thus, to increase the lithiophilicity of the porous materials, and spread the molten Li evenly, it is imperative to decorate the nano/microstructured Li host with the lithiophilic coating and/or heterogeneous seed growth. The lithiophilic nano-seeds, nanoparticles (NPs), or coatings can provide the initial nucleation site for guiding the uniform Li deposition densely and reversibly. This leads to a reduction in the voltage hysteresis, stability, and long battery cycling performance.
In this review, the current fundamental understanding and key progress of lithiophilic nano/ microstructured Li hosts are summarized. Furthermore, the issues, challenges, and a comprehensive viewpoint on the future development of lithiophilic nano/microstructured Li hosts are provided, aiming to boost the practical applications of LMA in high-energy-density batteries. The organized understanding of the formation of stable and robust SEI in lithium metal batteries (LMBs) can motivate designing the effective and efficient Li host in other battery systems.

Metal-Based Li Host
The electrochemical stability of the metal must be taken into account before using it as the Li host in LMBs [16]. To host the Li anode, the challenge is with Li alloying at a low potential range of 0.01-2.0 V versus Li/Li + . Copper (Cu), nickel (Ni), and stainless steel are considered stable enough to be used as Li host materials. The initial plating of lithium on planar two dimensional (2D) substrate renders inhomogeneous Li deposition due to the inhomogeneous distribution of the electric field. The rough, cracks, and inhomogeneous Li nucleation act as hot spots, which favor the continuous deposition of Li leading to dangerous Li dendrites and mossy Li. The continuous growth of needle and whisker shaped Li dendrite growth not only challenges the safety concern of batteries but also leads to the formation of uncontrolled and undesired formation/deformation of SEI. Thus, a host with abundant Li storage sites is strongly recommended, which can provide the route for dendrite free Li deposition and accommodate the volume expansion of Li, during Li plating/stripping, even at higher current density.

Lithium
Lithium metal itself can be constructed as the Li host in the form of a nano/microstructured framework. Ryou et al. designed defects in the lithium foil using the microneedle technique shown in Figure 1a to introduce high surface areas that serve as preferred sites for Li plating [17]. The other methods of producing microstructured Li include soft lithography, where continuous surface relief (CSR) and discontinuous surface reliefs (DSR) were fabricated on the surface of the Li metal foil electrode [18]. The continuous and discontinuous 10 µm deep surface modifications on 20 µm thin Li electrode as shown in Figure 1b,c provides an effective way to replace the use of excess Li. Kong et al. pressed the Li powder with a particle size of fewer than 20 µm on stainless steel mesh Figure 1d to create a porous Li-powder anode [19]. The porous anode with 25% Li loading eliminates the growth of Li dendrites. Later, Park et al., introduced micropatterns on lithium metal using micrometer-scale pyramid reliefs (height-50 µm, width-50 µm, and ridge length-40 µm) [20]. During stripping, the original dimension was recovered draining the liquid-like/or granular like plated Li from such structures as shown in Figure 1e,f. The coating of the micrometer particle size Li powder (CLiP) on the copper foil as shown in Figure 1g can also be done to improve the battery performance and replace the excess Li foil electrode [21]. This study shows that the porous structure and high surface area of the coated/pressed Li powder significantly reduce the current density during plating/stripping. The CLiP introduced the safety and practical battery operation at a lean lithium condition.
To further enhance the self-healing electrostatic shield mechanism, Kim et al., employed an electrolyte solution containing cesium hexafluorophosphate CsPF 6 (0.05 M) into micropatterned LMA [22]. The cesium ions form a positively charged electrostatic shield around the initial growth of protrusions, which obliges the deposition of Li adjacent to the protrusions leading to dendrite-free Li deposition [23,24]. Figure 1h,i show the cross-section SEM image of micropatterned Li without and with CsPF 6 . The micropatterned Li metal shows the uncontrolled granular Li deposition. In contrast, the micropatterned Li with CsPF 6 modifications shows the dendrite-free and stabilized Li metal anode. image of the Li-powder electrode. Reproduced with permission from [19], Institute of Physics Publishing, 2012. (e,f) SEM images of Li plated and stripped. Reproduced with permission from [20], John Wiley and Sons, 2016. (g) SEM image of the coating of the micrometer particle size Li powder (CLiP) electrode. Reproduced with permission from [21], John Wiley and Sons 2013. (h,i) Cross-section SEM images of patterned Li without and with CsPF 6 . Reproduced with permission from [22], Elsevier, 2018. (j,k) Schematic illustration of the preparation and Li plating/stripping on the Li/Li-Sn composite electrode. Reproduced with permission from [25], Springer Nature, 2020.
The battery operation at a higher rate demands low effective current density and fast lithium-ion diffusion kinetics. The nanostructured electrode lower effective current density and the lithium-containing alloys can facilitate the faster lithium-ion diffusion. Recently, Wan et al. reported an interpenetrated 3D lithium metal/lithium tin alloy nanocomposite by a simple calendaring and folding route and lithium-tin alloying reaction mechanism as shown in Figure 1j [25]. The schematic of dendrite-free Li deposition during plating/stripping cycles achieved with such lithium and the lithium-tin alloy is shown in Figure 1k. The strong anchorage affinity between lithium and lithium-tin alloy leads to low interface impedance and dense structure [1]. The potential difference between lithium and lithium-tin alloy and abundant interfaces enable ultrafast Li-ion diffusion across the entire LMA electrode. As a result, the lithium metal/lithium tin alloy nanocomposite anode delivered stable lithium plating/stripping cycles at 30 mA cm −2 and 5 mAh cm −2 for 200 cycles. The full cell configuration with LiNi 0.6 Co 0.2 Mn 0.2 O 2 (NCM) cathode exhibited an outstanding rate capability of 74% at 6 C compared to its initial capacity (167 mAh g −1 at 0.5 C) and with LiFePO 4 (LFP) cathode demonstrated capacity retention of 91% at 5 C after 500 cycles, respectively.

Copper
The 3D nanostructured copper with a high electroactive surface area is considered as an appealing Li host matrix. Such a framework can lower the current density and provide the room to accommodate the Li and manipulate the electric filed for guiding the Li deposition. To synthesize 3D Cu porous nanostructure, Yang et al. reported the simple solution-based approach with a submicron framework (median pore size 2.1 µm) by immersing the planar Cu foil in ammonia solution [9]. Figure 2a shows the schematic representation of preparing 3D porous Cu from the planar Cu. Figure 2b shows the lithium-ion flux in planar and 3D porous Cu. In planar Cu, the irregular, rough, and cracks act as a local hotspot for Li-ion accumulation. The Li metal deposition is amplified at those Li-ion accumulation centers with higher plating cycles, which leads to the growth of Li dendrites. The high surface area of Li dendrite consumes an excess amount of electrolyte and Li leading to poor CE and quick capacity fading. Besides, the dendritic Li can pierce the separator, which can cause a short circuit or fire caught, threatening the safety concern of batteries. In contrast, 3D porous Cu has uniformly distributed the Li-ion flux due to the numerous and uniform protuberant tips on the submicron fibers. As a result, the deposited Li fills the pores evenly without any Li dendrites. The high pore volume of the structure can provide sufficient room to accommodate Li that leads to the stable and reversible Li plating/stripping. As a result, the 3D porous Cu showed plating/stripping cycles for 600 h with low voltage hysteresis (<50 mV) at 0.2 mA cm −2 and exhibited a CE of 98.5% at 0.5 mA cm −2 . Further Li deposition behavior study on 3D Cu foam with an average pore size 170 µm showed that Cu foam does not have any effect of the Li space constraint. This led to the Li metal detachment from the backbones of the Cu foam causing electric disconnection, which induces Li dendrite growth. Lu et al. reported a free-standing 3D Cu nanowire (CuNW) network membrane via reducing copper nitrate in the NaOH aqueous solution [26]. The voltage hysteresis was further lowered to 40 mV and the CE, which reached 99.2% after 50 cycles remain stable to 98.6% after 200 cycles at 1 mA cm −2 . The full cell coupled with the LiCoO 2 cathode shows an outstanding 30% increase in the specific capacity at 5 C indicating the advantage of the Li-CuNWs nanocomposite in LMBs at a higher rate.
The dealloying mechanism such as the dissolution of zinc (Zn) out of brass (Cu-Zn alloy) using various chemical-solution and vacuum distillation processes results in the formation of voids and microstructures [27][28][29]. The porosity can be easily tuned by adjusting the etching condition (such as the concentration of the solution, distillation temperature, and time) depending upon the method used. The interconnected, uniform, and smooth porous Cu provides sufficient electrical conductivity, uniform Li-ion flux, and space for Li accommodation. Besides, 3D porous Cu exhibited the lower voltage hysteresis as shown in Figure 2c,d, and improved cycling performance with a capacity retention of 87.2% after 300 cycles of charge/discharge at 50 mA g −1 as shown in Figure 2e [28,29].
The vertically distributed electric field promotes the vertical growth of Li dendrites, which continues to grow towards the separator and cathode and eventually pierce the separator causing short-circuits. To address these issues, it is necessary to constrain the lithium dendrite growth within the porous scaffold and distribute the electric field in the lateral direction. For this, the compartmented and the vertically aligned (VA) 3D porous Cu have been designed by the laser microprocessing system, and a series of processes, including hot lamination, laser ablation, and alkaline etching treatments. Zou et al. developed compartmented 3D Cu (150 µm diameter) by an industry-adaptable technology following a series of processes, including hot lamination, laser ablation, and alkaline etching treatments [30]. The insulating polyimide (PI)-clad copper grid with lateral electric field distribution designated as E-Cu guides the Li deposition laterally within the porous scaffold of Cu (Figure 2f,g). In contrast, pristine copper (P-Cu) showed larger, thicker dendrites with an uneven surface as shown in Figure 2h,i. Wang et al. used a laser microprocessing system to design VA microchannels with optimized pore radius, pore depth, and pore spacing of 5 µm, 50 µm, and 12 µm, respectively [31]. The VA porous Cu provides a large surface area to deposit the lithium in the microchannels as shown in Figure 2j,k. The VA porous Cu demonstrated a stable CE of 98.5% for 200 cycles and voltage hysteresis of 30 mV after 50 cycles [31]. The lithium deposited porous Cu anode paired with the LiFePO 4 (LFP) cathode demonstrated higher capacity retention of 90% compared to the planar Cu (80%) after 100 cycles at 0.5 C.
The other techniques such as mechanical press or folding techniques have also been reported to make a 3D Cu/Li-metal composite anode that demonstrated a high CE and a longer and ultra-high plating/stripping capacity up to 50 mAh cm −2 at an ultrahigh current density of 20 mA cm −2 [32,33]. Cao et al., developed the vertically oriented lithium-copper-lithium arrays composite anode developed by mechanical rolling or repeated stacking [33]. Such composite anode showed voltage hysteresis of less than 55 mV and a lifetime up to 2000 h at 1 mA cm −2 to achieve a capacity of 1 mAh cm −2 . The poor wettability of lithium towards Cu does not spread the molten Li across the surface of lithiophobic substrates, promoting the non-uniform and dendritic Li deposition. To overcome this problem, the lithiophobic surface of Cu can be converted to lithiophilic by various coatings or layers of metal [35,36], oxide [37][38][39][40][41][42], nitride [43,44], oxynitride [45], sulfide [46], and phosphide [47], the addition of functional groups or doping [48][49][50][51], and by an increase in temperature [52].
Wang et al., studied the tuning of Li wettability by reacting molten lithium with elemental additives and organic compounds containing functional groups [50]. The negative value of Gibbs formation energy (∆ r G) for the reaction between the molten Li and coated/doped compounds indicates the formation of a new chemical bond, which improves the Li wettability. The wettability of Li was improved by adding elements such as indium (In), tin (Sn), and magnesium (Mg), which reduces the surface tension of the material and spread the lithium. Figure  Zhang et al., reported thin lithiophilic CuO VA nanosheets grown on Cu substrate by a simple wet chemical reaction, which not only ensures the good electrical conductivity of the electrode but also facilitates the fast Li-ion transport and regulates the Li nucleation [41]. Figure 3b shows that the molten Li is uniformly spread out on the VA-CuO-Cu surface but the molten Li forms a droplet on the Cu surface confirming the lithiophilicity and lithiophobicity of the VA-CuO-Cu surface and Cu surface, respectively. To further enhance the lithiophilic property, the dual lithiophilic materials can be employed. The lithiophilic copper oxides were grown on Cu foil by liquid oxidation of bare Cu foil and the subsequent N 2 plasma treating to achieve a high CE of 99.6% for 500 cycles at 1 mA cm −2 for 1 mAh cm −2 capacity [39]. The symmetrical cell demonstrated long cycling of more than 600 h with a low voltage hysteresis of 23.1 mV. This can be attributed to the regulated Li nucleation and reduced overpotential for guiding dendrite-free Li deposition. Figure 3c-e clearly demonstrated that the plasma strengthened CuO/Cu 2 O decorated Cu showed the uniform morphology of deposited Li. For the dense nucleation of Li and steady deposition into the porous structure, the copper nanowire with a phosphidation gradient (CuNW-P) was prepared by the chemical solution process first to grow CuNW and the subsequent chemical vapor reaction to generate phosphine gas [47]. A high Li loaded (44%) CuNW-P demonstrated a significant decrease in the nucleation overpotential and voltage hysteresis as shown in Figure 3f.

Nickel
Similar to the 3D porous Cu, the 3D porous Ni has also been considered as a better alternative to any 2D Li host. The 3D structure incorporates the active anode material and provides efficient charge transport pathways. To improve the high rate capability, cyclability, and reduce the lithium amount, the foam lithium anode was prepared by Li electrodeposition on a Ni foam substrate [53]. Developing the honeycomb-like porous 3D Ni host on the Cu current collector via a hydrogen bubble dynamic template electrodeposition method has also been reported for stable Li and Na metal batteries [54]. However, the electrochemical deposition of lithium into the Ni foam host material for making Li/3D Ni composite anode, generally, leads to uneven Li deposition. Moreover, the batteries have to be dissembled, cleaned, and further assembling of a new battery is needed for pairing the resultant composite electrodes with the practical cathode materials. To address this complex and costly process, the development of a simple approach for encapsulating Li inside porous Ni or other scaffolds to develop Li-based composite anode is highly looked-for.
Chi et al., reported the thermal infusion strategy for pre-storing molten Li into stable 3D Ni foam to achieve a composite anode [55]. Figure 4a shows the schematic illustration of making the Li/3D Ni foam composite anode. Figure 4b,c shows the digital photos of Ni foam and the Li-Ni composite. To further address the issues of LMA and improve the electrochemical performance of LMBs, it is required to decorate the 3D Li host with lithiophilic coatings. The lithiophilic decoration in a porous Li host can greatly reduce the nucleation overpotential and guide the uniform Li deposition. To increase the affinity or Li wettability between the lithium and 3D porous host, various polymers [56], organic coating [50], elementary substances (such as Si nanoparticles [57], plated Ag [58]), oxides (such as NiO [59], CoO [60], Co 3 O 4 [61], ZnO nanoarrays (NAs) [62]), fluorides (NiF x ) [63], nitrides (Ni 3 N) [64], and sulfides (Ag 2 S) [65] have been utilized. Wu et al., prepared a lithiophilic Cu-Cuo-Ni hybrid structure by depositing Cu on Ni foam by radio frequency (R-F) sputtering followed by annealing at 400 • C and then finally, CuO-Ni was reduced by H 2 plasma [66]. Figure 4j-l shows the lithium plating mechanism on Cu-Ni, CuO-Ni, and Cu-CuO-Ni current collector. The copper has high overpotential for Li nucleation, which can be lowered by using the lithiophilic CuO nanowire (NW) arrays. Besides, the evenly distributed electric field on CuO-NW arrays leads to the uniform Li-ion flux and lowers the current density. The thin layer of the Cu layer on CuO still maintains the lithiophilicity, offers lower nucleation potential, and increases the electrical conductivity. The Cu film acts as a protective layer during electrochemical cycling. As a result, the hybrid Cu-CuO-Ni structure demonstrated a dendrite free Li plating with CE > 95% for more than 250 cycles at 1 mA cm −2 and long symmetrical plating/stripping cycling up to 580 h at 0.5 mA cm −2 with a capacity of 0.5 mAh cm −2 . Lu et al. designed nanostructured lithiophilic VA NiO hexagonal nanoplates on Ni foil (HP-NiO-Ni) by a facile hydrothermal process as shown in Figure 4m [59]. Such a HP-NiO-Ni nanostructure significantly reduces the current density and nucleation overpotential for efficient, homogeneous, and smooth Li deposition. Besides, the lithiophilic Li 2 O and electronic conductive Ni formed during the reversible conversion reaction of NiO to Ni/Li 2 O facilitates uniform Li-ion transport and fast electron conduction pathways, respectively. Thus, the synergy of reversible conversion reaction, and the uniform/reversible Li plating mechanism stores excess Li even at higher current density.
The formation of the Li containing alloy is believed to lower the nucleation overpotential and guide the uniform Li deposition [1]. Sun et al. reported the lithiophilic ZnO NA modified Ni foam (ZMNF) by a facile seed-mediated hydrothermal reaction process [62]. The ZMNF current collector improved the affinity towards the Li, provided fast Li-ion kinetics, reduced the nucleation barrier, and lessened the Li volume expansion during plating/stripping cycles. The ZnO NA improves the Li affinity with molten Li by alloying and capillary effect as shown in Figure 4n. As a result, ZMNF exhibited a CE above 98% for over 300 cycles at 1 mA cm −2 and a symmetrical cell demonstrated a long plating/stripping cycles over 1200 h at 1 mA cm −2 to achieve 2 mAh cm −2 .
The graphene-based materials can provide scaffolds for Li electrodeposition and act as nucleation seeds. Graphitized carbon on Ni foam by heating Ni foam up to 1000 • C under Ar and H 2 and introducing methane (CH 4 ) [67] or acetylene (C 2 H 2 ) [68], few-layered graphene sheets attached to the surface of Ni foam following the acid-catalyzed hydrothermal process [69] have been reported for reduced overpotential and improved battery performance.
The other metal-based current collector such as titanium (Ti) and stainless steel has also been reported to guide the uniform Li deposition in LMBs. Zhang et al. proposed a porous lightweight corrosion-resistance 3D Ti current collector, which achieved outstanding CE of 99% at 1 mA cm −2 to achieve a capacity of 5 mAh cm −2 [70]. The metal-based porous framework including the methods of preparation and electrochemical performance is summarized in Table 1.  [70] To further enhance the lithiophilicity of Ti, the CuO nanoflower was grown on Ti mesh by the microwave-assisted solution reaction at 100 • C, which was further pressed on the Li wafer by a battery sealer with mechanical pressure of 800 psi [71]. The Li/Cuo@Ti-mesh (LCTM) composite anode exhibited a high CE of 94.2% at 10 mA cm −2 over 90 cycles with a low overpotential of 50 mV. The low overpotential of 50 mV and 250 mV were achieved at a high current density of 20 and 40 mA cm −2 . Lee et al., used the 3D conductive stainless steel fibrous metal felt (FMF) of a fiber diameter of 10 µm, porosity of 80%, and surface area or 0.05 m 2 g −1 to prepare the FMF/Li electrode by roll-pressing [72]. The FMF/Li symmetrical cell test showed a low overpotential of 30 mV at a high current density of 10 mA cm −2 .

Non-Metal-Based Li Host
Carbon-based materials are considered as an ideal candidate for a Li host. Employing the carbon-based Li host lowers the packing cost and increases the energy density of the battery due to its lower cost and lower density. Porous carbon materials have attracted intense attention due to their unique properties such as a high surface area with excellent mechanical strength and chemical stability, natural abundance, and are eco-friendly, which have great potential for practical energy storage applications. Using 3D porous carbon on top of the metal current collector shows the conductivity difference between them. The weak cohesion between the metal current collector and the porous carbon host provide space to deposit the Li underneath the porous carbon. The emerging 3D carbon such as hollow carbon spheres [73], carbon nanofiber (CNF) [74,75], carbon nanotubes (CNT) [76], porous carbon [10], graphene or reduced graphene oxide (rGO) [77][78][79][80][81], and carbonized MOFs [82] have made a great success by the pre-electrodeposition of Li. However, the lack of spatial control during the electrodeposition of Li leads to uneven Li deposition and incompetent production. The thermal infusion of molten Li into the scaffold of 3D porous carbon is one of the prominent approaches to make the Li/C composite anode. Moreover, most of the carbon materials possess a relatively weak affinity and lack a bonding interaction towards Li. The poor Li wettability of the porous carbon structure leads to the increase in nucleation overpotential, increases the voltage hysteresis, increases the cell's impedance, and cannot uniformly guide the Li deposition. Thus, it is imperative to decorate or modify the 3D porous carbon with the lithiophilic layer or seed growth. The various types of lithiophilic coatings and decoration of seeds for the guiding the lithium deposition will be discussed in the following section.

Metal Seeds
Yan et al., studied the nucleation overpotential of Li deposition on various materials such as Au, Ag, Zn, Mg, Al, Pt, Si, Sn, C, Cu, and Ni [83]. As shown in Figure 5a, for the materials with higher solubility (such as Ag and Mg), the slope of the voltage profile becomes flat before the onset of the Li deposition. For the materials with relatively lower solubility (such as Al and Pt), the Li nucleation overpotential can be observed. The nucleation overpotential increases as the solubility of the material decreases. Figure 5b shows a clear overpotential for Li nucleation on Si, Sn, C, Ni, and Cu substrate samples. As Cu and Ni do not form an alloy with Li, they have higher nucleation overpotential. Although C, Sn, and Si form an alloy with Li, still the overpotential can be observed but is lower than Cu and Ni. In their work, the carbon shell with embedded gold nanoparticles (AuNPs) showed a CE of 98% for more than 300 cycles higher than only the Au film or only the carbon shell, which implies that the necessity of lithiophilic porous scaffold for Li metal nucleation.
Wang et al., also studied the high Li wettability of elements such as Sn, Zn, Si, and Al, which form an alloy with Li [84]. This indicates the potential battery applications of these lithiophilic elements on any kinds of Li host. The chemical potential gradient inside the CNF was introduced using higher chemical potential elements (Ni or Co) in the inner layer and lower chemical potential (Ti or W) in the outer layer forming a hollow structure [85]. The lithiophilic particles inside the hollow structure obtained a high CE of 99.6% for more than 1400 cycles at a current density of 0.5 mA cm −2 and capacity 1 mAh cm −2 .
Liang et al., designed a Li/carbon composite anode by melt infusion of Li into the lithiophilic Si coated 3D porous carbon matrix [57]. The silicon coating and molten Li react to form a binary lithium silicide alloy that enhances the Li wettability of the entire surface and fills the porous structure. It was observed that the molten Li forms a droplet on the carbon framework. In contrast, the molten Li spread out on the surface of the Si coated carbon framework. This indicates the enhanced lithiophilicity of the porous carbon framework by Si coating. The Li/C composite anode exhibited stable Li plating/stripping cycles with an overpotential of <90 mV at 3 mA cm −2 over 80 cycles. Silver (Ag) has good solubility in Li and can form an Ag-Li alloy, it has been used to enhance the lithiophilicity of porous carbon materials for improving the plating/stripping cycles and full cell cycling performance. The anchoring of Ag nano-seeds on CNFs by Joule heating [86] and electroplating of coralloid-type Ag on carbon fiber [87] have shown low voltage overpotential and long stable cycling performance.
The conformally coated Sn on free-standing and highly flexible 3D CNFs (CNF-Sn) also provides a lithiophilic nucleation site and superior electric contact between the Li plating and conductive CNF [88]. The low energy barrier of Li-Sn alloying leads to the small nucleation overpotential of 28 mV and potential polarization of 14 mV during Li plating/stripping cycles. The schematic in Figure 5c shows the Li deposition with and without Sn coating on CNF. The poor lithiophilicity of CNF induces Li to nucleate on the hotspots, which grow like grains. The insufficient electric contact between the Li grains and CNF leads to the residual of dead Li. In contrast, the homogenous Li deposition can be attained with the CNF-Sn.

Graphene and Graphene Oxide
The graphene-based electrode gives rise to stable SEI due to their higher flexibility, lower-dimensional change, and a relatively larger surface area than 2D current collectors [77]. The infusion of Li inside the layered rGO films with uniform nanogaps can retain capacity up to 3390 mAh g −1 of capacity with a low overpotential of 80 mV at 3 mA cm −2 [78]. This can be attributed to the fast and uniform Li intake due to the lithiophilic nature of sparked rGO and the capillary force produced by the nanogaps. The uniformly spread lithium on layered rGO regulates the initial Li nucleation due to the good lithiophilicity of rGO and the low Li nucleation barrier. The functional group such as carbonyl (3.080 eV) and alkoxy (2.974 eV) on rGO have higher binding energy (BE) to Li than the bare graphene counterpart (1.983 eV), responsible for Li affinity.
Besides, the edge and high coordination sites in graphene and nitrogen-doped graphene with a functional group enhance the lithiophilic property [89][90][91][92]. Liu et al. employed crumpled paper ball-like ultrafine rGO particles with a specific surface area (SSA) of 382 m 2 g −1 and pore volume 1.8 cm 3 g −1 to accommodate high Li loading up to 10 mAh cm −2 [81]. The improvement can be attributed to the lithiophilic rGO due to a sufficient amount of surface oxygen functional groups. This eliminates the need for additional lithiophilic coating/seed growth.

Heteroatom Atom
The doping of polar heteroatoms (such as N, S, B, and O) into the nonpolar nanostructured carbon is another approach to trap the Li-ions. The lithiophilic oxygen and nitrogen co-doped porous carbon granules (ONPCGs) was coated on Cu foil to achieve a high CE of 99.83% after 350 cycles at 2 mA cm −2 , 96.4% after 130 cycles at 20 mA cm −2 and still stably work at a high capacity of 6 mAh cm −2 for 115 cycles [93]. The channel-like pore structure in porous carbon with a high specific area of 2396 m 2 g −1 containing oxygen and nitrogen functional groups lower the local electric field and enhances the lithiophilicity that leads to the uniform Li growth at an ultrahigh current density up to 30 mA cm −2 .
The edge and high coordination sites in graphene and nitrogen-doped graphene with functional groups are suitable for Li adsorption. Huang et al., presented nitrogen (N)-doped nanoporous graphene with a high lithium loading amount of 95.3 wt % [90]. The melted lithium can be infused into the nanopore channels within 1 min due to the doping of the lithiophilic functional group and capillary forces of nanopores. The N-doping offers well-distributed anchoring sites for uniform Li deposition.
The N-doped graphene matrix is full of lithiophilic functional groups, such as pyridinic, pyrrolic, and quaternary nitrogen atoms [91]. Figure 5d shows the fitted results of N 1s X-ray photoelectron spectroscopy (XPS) spectra with thee peaks-pyridinic nitrogen (pnN) pyrrolic nitrogen (prN) and quaternary nitrogen (qN). Figure 5e displays the BE of Li atom with Cu, graphene, and different functional groups of nitrogen-doped graphene. The higher binding energy of prN (4.46 eV) and pnN (4.26 eV), which increases the local charge density, provides strong Li affinity. The extra pair of electrons in pyrrolic or pyridinic nitrogen acts as an electron-rich donor with filled p orbitals behaving itself as a Lewis base site. Such a Lewis base site has a strong affinity to adsorb Lewis acidic Li-ions through acid-base interactions, which brings a lower nucleation overpotential.

Metal Oxides or Nitrides
The overpotential during Li plating comes from (1) mass transfer, charge transfer, and ohmic resistance and (2) Li nucleation barrier. The mass transfer overpotential is related to the applied current density and the mobility of Li-ions, charge transfer and ohmic resistance depends on the electronic and/or ionic conductivity of the electrode and the ionic conductivity of the electrolyte. In contrast, the Li nucleation overpotential, which is the voltage difference between the bottom of the voltage dip and the latter voltage plateau depends on the energy barrier of Li nucleation.
During the initial Li plating, the lithiophilic metal oxides (such as ZnO and CuO) provide the Li nucleation side. The conversion and/or alloying reaction is beneficial for Li nucleation and excess storage of Li, which eventually lowers the nucleation overpotential and lowers the ultimate voltage hysteresis. For example, Jin et al., decorated the bamboo-derived 3D hierarchical porous carbon (HPC) with ZnO quantum dots to confine up to 131 mAh cm −2 lithium within its scaffold for stable LMA [10]. In the initial plating cycles, Li nucleates on the ZnO seeds with the conversion reaction.
After higher plating cycles, the Li and Zn form the LiZn alloy. The Li nucleation is effectively controlled by both the lithiophilic ZnO seeds and LiZn alloy, resulting in smooth Li deposition as shown in Figure 6a,b. Moreover, lithiophilic gradient distributed ZnO particles on porous carbon against the Li-ion concentration gradient are supposed to reduce the lithium metal nucleation barrier [94]. The other form of lithiophilic material can be film or layer-type. The coating of the ZnO thin layer on the channel walls of 3D porous carbon by a simple and low-cost solution process method demonstrated a low overpotential of 90 mV and stable cycling performance 150 h at 3 mA cm −2 [95]. The amorphous oxides show the superlithiophilic features owing to its higher binding energy (BE) compared to the crystallized oxides. Xue et al., showed that the amorphous TiO 2 (−6.73 eV) and SiO 2 (−2.20 eV) have higher BE compared to the crystallized TiO 2 (−3.66 eV) and SiO 2 (−1.97 eV) [96]. The higher BE indicates the strong anchoring for Li adsorption and dense Li deposition. The SEM images after 100 cycles at 4 mA cm −2 with a capacity of 1 mAh cm −2 on Cu, 3D carbon fiber (CF), and 3D porous core-shell carbon fiber with amorphous SiO 2 and TiO 2 (PCSF) are shown in Figure 6c,d, Figure 6e,f, and Figure 6g,h, respectively. Only PCSF showed the uniform and smooth surface morphology of Li deposition indicating the strong anchoring behavior of lithiophilic SiO 2 and TiO 2 . The PCSF demonstrated controlled Li plating, low nucleation overpotential, and low hysteresis. As a result, the symmetric cell test exhibited a stable Li plating/stripping of more than 350 cycles with a low overpotential of 35 mV at 4 mA cm −2 .
The use of metallic nitrides as a lithiophilic seed has been demonstrated in the 3D porous carbon for significant improvement in the CE, stable symmetric cell, and full cell testing [97]. During initial Li nucleation, Mo 2 N converted to Mo and Li 3 N due to the highly reductive nature of Li. The in-situ formed Mo further serves as a preferred nucleation site. Besides, Li 3 N has promising ionic conductivity and high Young's modulus [7]. As a result, Mo 2 N seeds modified CNF demonstrated an ultra-low Li nucleation potential of 10 mV, CE of 99.2% over 150 cycles at 4 mA cm −2 , and stable symmetrical cell operation of 1500 h at 6 mA cm −2 to achieve a capacity of 6 mAh cm −2 [98]. The use of lithiophilic and pseudocapacitive behavior of TiN also favors ultrafast Li storage and charge transport [97].

All in One
Most of the previous reports are focused on developing the 3D porous carbon first and then modifying the porous matrix by using heteroatom doping, seed growth, or other lithiophilic coatings. Recently, 'all in one' porous carbonized metal-organic framework (MOF) as a Li host has attracted great attention. The doping of polar heteroatoms and decorating porous carbon or graphene with lithiophilic metal/metal oxides seeds not only enhance the electronic conductivity but also allows fast Li-ion diffusion due to the coordination between metal/metal oxides or nitrides seeds and polar heteroatoms.
Wang et al., reported one-step carbonization of MOF to obtain the mixed ionic/electronic conductive porous carbon with evenly distributed Co nanoparticles in N-doped graphene [99]. As a result, high CE of 91.5% after 130 cycles at 10 mA cm −2 and a full cell exhibited long cycle life with a capacity retention of 92% at 1 C. The carbonized MOF improves the electronic conductivity, the uniformly dispersed Co nanoparticles in high surface area N-doped porous graphene enhances the lithiophilicity and avoids charge concentration. The preparation of an 'all in one' Co@N-G porous Li host is shown in Figure 6i. The design of Co 4 N-doped and Co nanoparticles embedded into the N-doped carbon nanocubes (Co/Co 4 N-NC) by pyrolysis of MOF as shown in the schematic Figure 6j has also stabilized the LMA in Li-air battery [100]. The formation of stable and robust SEI on the surface of lithiated Co/CoN-N-doped porous carbon anode prevents the side reaction between the LMA and air/electrolyte. The functionalization of MOF with amine groups [101] and the use of metal oxides NA with N-containing functional groups have also shown improved battery cycling performances [102]. Thus, the 'all in one' Li host can be a promising Li host candidate for the high energy density LMBs. Table 2 summarizes the representative carbon-based porous framework that has been reported for improving the lithium metal battery performance.  The current research on LMBs is advancing towards the discovery and exploitation of materials, which have high Li anchorage affinity. There are numerous reports on the use of lithiophilic materials on porous Li host framework, separators, and artificial SEIs. Table 3 summarizes the lithiophilic materials employed in porous metal or carbon framework for dendrite-free, smooth Li deposition that leads to the enhanced cycling performance in LMBs. In our previous work, we deposited the thin layer of lithiophilic copper clad on the flexible porous carbon as the backbone [8]. The improved lithiophilicity, electronic conductivity, and mechanical flexibility of the Li host exhibited a CE of 99.5% even after 300 cycles (1200 h) at 0.5 mA cm −2 .

Conclusions and Viewpoints
Lithium metal-based anodes are considered a potential candidate to meet the ever-increasing energy demands. However, the serious safety and poor cycling performance of the lithium metal anode have impeded the practical application of next-generation energy storage systems. The well-designed micro/nanostructured Li host materials could effectively improve the safety concerns and efficiently boost cycling performance. The high surface area and the electronic conductivity of such a porous matrix renders a uniform electric field, buffers the Li expansion during plating/stripping cycles, and lowers the effective current density. As a result, the uniform Li-ion flux and dendrite-free Li deposition not only ensures the safety issues but also addressed the low cycling performance. However, the micro/nanostructured Li host is still not up to the challenges to thoroughly address all the issues associated with lithium metal anodes. Firstly, the lithiophobic host materials increase the Li nucleation overpotential, cannot guide the uniform Li deposition, and shows increased voltage hysteresis on continuous cycling. Secondly, the micro/nanostructured host materials occupy a larger volume compromising of the volumetric energy density, and thirdly, the high surface area needs excess electrolyte to wet the surface and expose the high area of the lithium metal to the electrolyte. This consumes both the lithium and electrolyte leading to quick capacity decay and poor CE.
The efficient use of a 3D porous Li host can only be done with the use of lithiophilic decorations or coatings. The design of lithiophilic micro/nanostructured composite Li anode by infusing the molten lithium metal inside a porous structure can effectively address the safety and challenges in high energy-density lithium metal batteries. The use of 3D scaffolds with mechanical and chemical stability and mixed ionic conductivity can generate a robust, stable SEI, suppress the Li dendrite growth, and inhibit the side reactions between the lithium metal anode and the electrolyte. More recently, the development of 'all in one', i.e., the lithiophilic and porous Li host, is the simple and low-cost technique for enhancing the Li nucleation behavior and reducing the voltage hysteresis. The decoration of a porous framework with lithiophilic coatings or seeds allows high Li loading and dendrite-free Li deposition into such porous structures. Although it is tricky to obtain the fresh Li during the critical initial plating mechanism, further understanding of the initial Li nuclei using advanced characterization tools is necessary. The ex-situ or in-situ derived stable SEI and the use of a solid-state electrolyte along with the lithiophilic materials decorated nano/microstructured Li host can further enhance the battery cycling performance.