Emerging Carbon Nanotube‐Based Nanomaterials for Stable and Dendrite‐Free Alkali Metal Anodes: Challenges, Strategies, and Perspectives

Alkali metals (Li, Na, and K) are promising candidates for high‐performance rechargeable alkali metal battery anodes due to their high theoretical specific capacity and low electrochemical potential. However, the actual application of alkali metal anodes is impeded by the challenges of alkali metals, including their high chemical reactivity, uncontrolled dendrite growth, unstable solid electrolyte interphase, and infinite volume expansion during cycling processes. Introducing carbon nanotube‐based nanomaterials in alkali metal anodesis an effective solution to these issues. These nanomaterials have attracted widespread attention owing to their unique properties, such as their high specific surface area, superior electronic conductivity, and excellent mechanical stability. Considering the rapidly growing research enthusiasm for this topic in the last several years, we review recent progress on the application of carbon nanotube‐based nanomaterials in stable and dendrite‐free alkali metal anodes. The merits and issues of alkali metal anodes, as well as their stabilizing strategies are summarized. Furthermore, the relationships among methods of synthesis, nano‐ or microstructures, and electrochemical properties of carbon nanotube‐based alkali metal anodes are systematically discussed. In addition, advanced characterization technologies on the reaction mechanism of carbon nanotube‐based nanomaterials in alkali metal anodes are also reviewed. Finally, the challenges and prospects for future study and applications of carbon nanotube‐based AMAs in high‐performance alkali metal batteries are discussed.

the practical applications of these alkali metal anodes (AMAs) are impeded by their high chemical reactivity, which could induce dendrites growth, low Coulombic efficiency, and safety problems during cycling. [17,20]In this regard, developing efficient and stabilization strategies for AMAs are urgently required.25][26][27] As a superior type of carbon materials, carbon nanotubes (CNTs) have received much attention in AMAs because of their unique 1D nanostructures, high specific surface area (SSA), excellent electronic conductivity, as well as outstanding flexibility and chemical stability. [28]CNT-based nanomaterials with excellent properties would enable AMAs to obtain outstanding cycling stability during repeated charging and discharging cycles. [29]In recent years, CNT-based AMAs have become popular research topics.For example, Pan et al. [30] prepared CNT/CNF composite layers to modify separators for stable lithium metal batteries.Sun et al. [31] reported CNT film decorated with Ag particles, which was used to modify Na metal surface for dendrite-free Na metal batteries.Tang et al. [32] synthesized nitrogen-containing MXene/CNT@K composites and used it as high-performance K metal battery anode materials.Moreover, Tang et al. [33] summarized recent progress on carbon-based materials for Li metal anodes, including carbon nanotube, graphene, graphitic-amorphous carbon, and their composites.Pang and co-workers [34] reviewed recent advances in 2D materials for stable AMAs.Nevertheless, to the best of our knowledge, a critical review exclusively focusing on the recent progress of CNT-based nanomaterials in stable and dendrite-free AMAs has not been reported.
In this review, we summarize recent progress on CNT-based nanomaterials for AMAs. Figure 1 shows a gradual increasing trend in publications on this topic.First, the research background of AMAs is reviewed, including the categories, merits, existing issues, and the corresponding stabilizing strategies.Furthermore, the structures, synthesis, and properties of carbon nanotubes are introduced.Moreover, the applications of CNT-based nanomaterials in AMAs are reviewed, including alkali metal (AM) host design, modifying the current collector, building artificial SEI on AM surface, and other strategies.The relationships among methods of synthesis, nano-or microstructures, and electrochemical properties of CNTbased AMAs are systematically discussed.In addition, advanced characterization technologies on the reaction mechanism of CNT-based nanomaterials in AMAs are also summarized.Finally, challenges and reasonable suggestions for future research directions in this field are proposed.

Alkali Metal Anodes for Metal Batteries
Alkali metal anodes usually include metallic Li anode, Na anode, and K anode, and they have many commonalities in their chemistry.In recent years, AMs have gained extensive attention as anode materials for highenergy-density AM batteries (AMBs). [35]ongxiu Liu received his B.Eng. degree from Henan University of Science and Technology in 2019 and continues to pursue a Master's degree under the supervision of Prof.

. Large Theoretical Specific Capacity
Alkali metal anodes have large theoretical specific capacities. [34]The theoretical specific capacities of the Li, Na, and K metals are ~3860, ~1166, and ~687 mAh g −1 , respectively (Figure 2a).When assembled into AMBs, they can provide a higher specific capacity than other AMion battery anode materials.

Low Electrochemical Potential
Metallic Li, Na, and K have an electrochemical potential of −3.04, −2.71, and −2.93 V, respectively (Figure 2b). [34]AMBs can operate at high voltages because of their ultralow electrochemical potentials.

Suitability for Batteries with High Specific Energy
The large theoretical specific capacity and low redox potential of AMAs are essential for building batteries with high specific energy (Figure 2c).Furthermore, lithium, sodium, and potassium metals can be directly utilized as anode materials, which can avoid the use of large-mass and inactive current collectors, thus increasing the batteries' specific energy. [20]

Issues of AMAs
Alkali metal anodes are facing many challenges which severely limit their practical application.Their high chemical reactivity and infinite volume expansion are the root of the formation and growth of dendrites. [17]The resulting crack of the solid electrolyte interphase (SEI), coupled with severe side reactions and dendrites growth, ultimately leads to safety issues and loss of capacity. [17]Figure 2d-f shows the main issues facing AMAs.

High Chemical Reactivity
Since there is only one electron in the outermost orbital of AMs, they can easily lose this electron and form cations. [17] Metallic Li, Na, and K are very sensitive to oxygen (O 2 ) and water (H 2 O).Therefore, AMs should be stored and handled in a hermetically sealed environment with low oxygen and water content, increasing their storage and handling cost.

Unstable Solid Electrolyte Interphase
The SEI formed by the side reactions of AMs with liquid electrolytes is fragile and unstable (Figure 2d). [35,36]The infinite volume expansions of AMAs during AM plating and stripping could lead to the crack of the fragile SEI, which exposes new AM below it and further induces the formation of new SEIs through undesirable side reactions, reducing the Coulombic efficiency. [37]

Infinite Volume Expansions
Because of the "hostless" property of AMAs, their volume expansion during cycling is infinite.The uncontrolled volume variation leads to severe damage to the SEI and the electrode's structure, resulting in the short lifespan and inferior cycling stability of AMAs. [38]The repeated destruction and rebuilding of the SEI and the formation of dead AM will constantly consume fresh AM and the liquid electrolyte, leading to rapid battery failure (Figure 2e).

Uncontrolled Dendritic Growth
Dendrite growth behaviors have been intensively studied in Li metal anodes and also proven in Na and K metal anodes. [39]The formation of dendrites is one of the greatest challenges, as it can induce significant safety hazards.Uncontrolled dendrites can easily puncture the separators and cause a short circuit of the battery, which could lead to safety problems such as overheating, fire, and even thermal runaway (Figure 2f).In addition, the growth of dendrite accelerates the side reactions between the AMs and the electrolyte, which can irreversibly consume the AMs and the electrolyte, thus inevitably further reducing the Coulombic efficiency.

AMA Host Design
Designing a host for a "hostless" AMA is a valid strategy to reduce volume change during the charging and discharging processes. [44]he host usually has a large SSA to decrease local current density (LCD) and suppress the AM's dendritic growth. [45]Besides, the host could reduce the AM's direct contact with the liquid electrolyte, alleviating corrosion of the active AMs by the liquid electrolyte and the incidental side reactions. [23]The following attributes should be present in the ideal host: 1) the host must be lightweight, have sufficient mechanical strength and have an abundance of pores accommodating the AMs; 2) the host ought to have excellent electrical conductivity for rapid electrochemical reactions kinetic; and 3) the host should have outstanding stability during repeated AM plating and stripping processes.

Decorating Current Collectors
Ideal current collectors for AMA should have a high SSA and excellent mechanical stability. [46]However, there are still problems with their further applications, such as the problem of their poor affinity to AM, large mass density, and insufficient electrical conductivity, which can be solved by decorating current collectors for AMAs.A modified current collector could not only regulate ion distribution but also effectively reduce the LCD and alleviate volume expansion during charging and discharging cycles. [47]

Building Artificial SEIs
The SEI formed on AMAs is brittle and cannot withstand large volume fluctuations during AM plating and stripping. [35]Building an artificial SEI can stabilize AMAs.An excellent artificial SEI ought to have 1) thin and strong adhesion to AM surfaces; 2) high mechanical strength to effectively inhibit dendrite growth; and 3) enough flexibility to relieve the electrode's volume expansion during cycling.A dense and stable artificial SEI formed on the surface of AMAs can efficiently guide the deposition of AM and inhibit the reaction between the electrolyte and AMAs. [48,49]

Optimizing Liquid Electrolytes
Alkali metal anodes can be stabilized by modulating the composition of liquid electrolytes, which could be optimized through adjusting AM salts, and solvents and adding multifunctional additives. [50]Optimized liquid electrolytes for AMAs can not only regulate the ion distribution but also suppress the side reactions between the liquid electrolyte and the AMs. [51]

Modifying Separators
The separators are often modified by one side and double-sided through slurry coating method and magnetron sputtering approach. [43,52,53]The one-side modified separator can prevent side reactions and regulate AM ion transport behavior during plating and stripping, improving AM uniform deposition, which leads to enhanced Coulombic efficiency and cycling stability of AMAs. [43,53,54]Moreover, the double-sided modified separators with carbon-based materials can effectively improve the performance of both the AMA and the composite cathode, leading to greatly improved electrochemical performance of the corresponding AM full cells. [30,52]

Introducing Solid-State Electrolytes
Liquid electrolytes typically corrode AMAs and can bring potential safety hazards. [55]The intrinsic safety of batteries can be improved by introducing solid electrolytes.In addition, the strong mechanical properties of a solid electrolyte can inhibit the growth of dendrites.Solidstate AMBs is regarded to have superior safety and specific-energy density than AMBs with organic electrolyte. [56]However, the following issues should be given attention: 1) the inferior contact between SSEs and the AMs, which hinders their further application; 2) the relatively

Structure, Synthesis, and Properties of CNTs
Carbon nanotubes have been widely investigated since their discovery in 1991. [58]They have excellent mechanical and electrical properties compared with conventional engineering materials.CNTs are composed of hexagonal C atoms which form several to tens of layers of coaxial tubes.They are regarded as curled graphene sheets and mainly include single-walled CNTs (SWCNTs) and multi-walled CNTs (MWCNTs), which could be differentiated by the number of sheets. [59]uring the preparation of MWCNTs, the walls of MWCNTs are generally filled with holes and defects because the layer of CNTs can act as a trap to catch various defects. [60]Compared to MWCNTs, SWCNTs have a narrower diameter distribution range, fewer flaws, and better uniformity than MWCNTs.Carbon nanotubes could be thought of as 1D quantum wires with excellent electronic conductivity when their diameter is smaller than 6 nm, and have great potential as electrode materials for batteries. [28][62][63] Among them, the CVD method has great potential to produce quite pure CNTs on a large scale, and the reaction process is easy to control. [59]However, the CNTs produced using the CVD method usually have more structural defects than other methods. [64]Furthermore, the arc discharge method has been widely used for fabricating perfectly straight CNT with fewer structural defects compared with the CVD method. [64]Nevertheless, when CNTs are obtained by the arc discharge method, the alignment is difficult to control. [65,66]Moreover, the laser ablation method is used for producing CNT with relatively lower metallic impurities compared with the arc discharge method.But the CNTs obtained by the laser ablation method are not uniformly straight and contain some branching. [64]In addition, the CO gas-phase catalyzed growth method is promising to produce CNTs with a small diameter, and the size and diameter distribution of the nanotubes can be roughly controlled.However, the synthetic procedure of the gasphase catalyzed growth method is complicated and difficult to scale up. [62]s an excellent kind of carbon material, CNTs have been widely studied because of their large SSA, outstanding flexibility, excellent electronic conductivity, and good chemical stability. [28]Meanwhile, owing to the above-mentioned merits, CNT-based nanomaterials can regulate the uniform deposition of the AM and alleviate volume expansion during repeated AM plating and stripping processes, and they have great potential for AMAs.However, the fabrication processes of CNTs are usually expensive and complex, and the initial charging-discharging Coulombic efficiency is low when tested for electrode materials. [60]Besides, the impurities are hard to remove and the size of the CNTs is not easy to control.The methods of synthesis, structures, advantages, and disadvantages of CNTs are summarized in Figure 4.

The Application of CNT-Based Nanomaterials in AMAs
In recent years, CNT-based nanomaterials have been shown to have prospective applications in the field of AMAs, [67] and they typically have the following advantages: first, CNT-based nanomaterials have great potential to be fabricated into various porous scaffolds because of their 1D structure and excellent flexibility.Second, CNT-based nanomaterials have high electronic conductivity and ion diffusion coefficients, so modifying the host or current collector by them could enable fast electrochemical kinetics during the charge and discharge process of AMAs.Third, the large specific area and porous structures of CNTbased nanomaterials can relieve volume expansion during the cycling  [57] Copyright 2022, Elsevier.Top middle: reproduced with permission. [47]Copyright 2021, Elsevier.Top right: reproduced with permission. [48]Copyright 2021, Elsevier.Bottom left: reproduced with permission. [25]Copyright 2021, The Royal Society of Chemistry and the Chinese Chemical Society.Bottom middle: reproduced with permission. [26]opyright 2019, Royal Society of Chemistry.Bottom right: reproduced with permission. [27]Copyright 2020, WILEY-VCH.
Energy Environ.Mater.2023, 6, e12525 of AMBs.In recent years, various CNT-based nanomaterials have been successfully applied in various AMAs including lithium, sodium and potassium metal anodes (Scheme 1).The summary of the methods of synthesis and their battery performance are provided in Table 1.

The Application of CNT-Based Nanomaterials in Li Metal Anode
Lithium metals are a promising candidate for battery anode due to their high theoretical specific capacity and low electrochemical potential. [56,110,111]Nevertheless, the issues of "hostless" lithium metal include uncontrollable dendrite growth, an unstable SEI, and infinite volume expansion in charge and discharge cycles, limiting the further applications of lithium metal anodes (LMAs). [112]Because of the excellent conductivity and other unique features, CNTs have been extensively used to solve the above-mentioned issues and enhance the Coulombic efficiency and electrochemical performances of lithium metal batteries (LMBs).Typically, the strategies for using CNT-based nanomaterials in LMAs include CNT-based Li host design, [23,68,69,71,72,79,83,84,113] modifying current collectors, [41,46,47,[75][76][77]80,82,93] decorating the Li metal surface, [24,42,81] and other strategies. [30,114]

CNT-Based Li Host Design
Designing a host for LMAs can effectively solve the above-mentioned issues.CNT-based materials can be a good choice as hosts and have received considerable attention for LMAs. [69,70]For example, in 2018, Sun et al. [69] first demonstrated the application of CNT paper in LMAs and proposed CNTs as a promising host for the anode of Li-based battery.This porous and conductive network in the CNT paper could reduce the LCD and suppress the formation of lithium dendrites (Figure 5a,b).The CNT paper exhibited a network of MWCNTs (Figure 5c).When fabricated into symmetric cells, the Li/CNT electrode depicted excellent rate performance under the range of 2-10 mA cm −2 , and showed outstanding cycling performance over 2000 h without short-circuiting at 10 mA cm −2 (Figure 5d).Later, Yang et al. [70] investigated the Li plating and stripping behaviors on a CNT sponge.The large SSA of CNT sponges increased the Li nucleation sites and decreased the LCD, ensuring uniform lithium deposition. [70]Furthermore, Wu and co-workers [74] prepared a double-layer CNT sponge microfilm (CSMF) wrapped with fused lithium (Li-CSMF) through a metallurgical technique (Figure 5e).In a CSMF, the evenly distributed and enmeshed CNTs have a mean diameter of 10.6 nm (Figure 5f).After the homogenous infusion of Li, the mean diameter of the Li-CSMF was increased to 136 nm (Figure 5g), which still preserved the uniform and dense winding structure of the CNTs.The diffraction peaks corresponding to (110), (200), and (211) of lithium are shown in Figure 5h, demonstrating the excellent crystalline quality of Li-CSMF.When fabricated into punch cells (Figure 5i inset), the resistance of the Li-CSMF¦¦S punch cell was smaller than that of the Li¦¦S pouch cell (Figure 5i).Besides, the Li-CSMF¦¦S pouch cell could supply stable power for a rotating electric fan in the kneaded or folded state, indicating the excellent structural stability and flexibility of the Li-CSMF electrode (Figure 5j).The superior performance is attributed to the unique structure of the Li-CSMF electrode, in which the high SSA of CNTs could effectively reduce the LCD in the cell, and the conductive CNT network can facilitate the lithium ions transportation during the charging/discharging process. [74]lthough CNTs are promising for LMAs, the poor lithiophilicity of CNTs hinders their further development. [79]Therefore, improving the lithophilicity of CNTs has become the focus of recent research. [23,71,84,113]For example, Wang et al. [113] grafted amidecontaining functional groups (A f s) on CNT films (A f -CNT) to improve the lithophilicity of CNTs (Figure 6a).The flexible A f -CNT film showed a dense structure, and the A f s were evenly distributed on the CNTs' surface (Figure 6b-d), which lowered the nucleation barriers and guided uniform Li deposition along or into the nanotubes. [113]When lithium was deposited at 2-8 mAh cm −2 , the A f -CNT@Li electrode depicted a uniform and smooth surface (Figure 6e).When fabricated into symmetric batteries, the A f -CNT electrodes exhibited a stable cycling performance over 880 h at 3 mA cm −2 and 1 mAh cm −2 (Figure 6f).When tested in A f -CNT@Li¦¦LFP cells, the A f -CNT electrodes delivered a reversible capacity of 125 mAh g −1 after 350 cycles at 1C (Figure 6g).The outstanding electrochemical performance can be attributed to the high electronic conductivity and 3D framework with a large interspace of the porous A f -CNT, which can efficiently lower the LCD and alleviate volume expansion during repeated cycling. [113]In addition, Mei et al. [73] prepared CNT/NiO@Li electrodes as LMAs via thermal infusion of CNT/NiO sponge (Figure 6h).NiO spheres with excellent lithiophilicity were chosen to anchor the CNT sponge to improve the wettability of the CNT with Li.After Li infusion, the NiO spheres and CNTs were still tightly crosslinked (Figure 6i).When lithium was plated at 3 mA cm −2 , the CNT/NiO@Li electrode displayed a dendrite-free surface (Figure 6j,k).The uniformly distributed nanostructured NiO could provide plenty of nucleation sites, regulating uniform Li + deposition and lowering Li nucleation barriers. [73]The CNT sponges can dissipate high current densities and prevent the generation of isolated Li during plating and stripping due to their superior conductivity and large SSA. [115]Because of their excellent properties, when fabricated into symmetric cells, the CNT/NiO@Li electrodes exhibited a stable cycling performance at 3 mA cm −2 and 3 mAh cm −2 (Figure 6l).
Moreover, CNTs are usually composited with other carbon-based materials as hosts for AMAs, including reduced graphene oxide (rGO), carbon nanofiber, or MXene. [23,44,100]For example, Zhou et al. [71] designed a 3D CNT/rGO skeleton with lithiophilic ZnO nanoparticles (G-ZGC) as a host for LMAs.First, zinc acetate was dissolved in GO/ CNT aqueous solution under stirring for 12 h and then conducted vacuum filtration.After lyophilizing for 48 h, the ZnOAc/GO/CNT precursor was obtained.Finally, after annealing under an Ar atmosphere, a gradient ZnO/rGO/CNT host was fabricated (Figure 7a).The CNT and ZnO nanoparticles were uniformly anchored on the rGO nanosheets (Figure 7b,c).When lithium metal is deposited on the G-ZGC electrode, it showed a dendrite-free morphology, indicating that the 3D host can effectively guide the deposition of lithium (Figure 7d).During the discharge process, the reaction equation is as follows: And during the charging process, the de-alloying of metallic Li occurs: As a result, the full LFP/Li@G-ZGC cell exhibited superior rate performance (Figure 7e).After rate cycling, no dendrite was formed on the Li@G-ZGC electrode (Figure 7f). [71]In addition, CNT-based materials modifying other hosts can also achieve excellent performance in LMAs. [23,44]For example, Zeng et al. [44] prepared CNT-modified Ndoped carbon nanofibers with Ni nanoparticles (CNTs-Ni@NCFs) as a host for LMAs (Figure 7g).First, the freestanding N-doped carbon nanofibers (NCFs) precursor was prepared by the electrospinning method (Figure 7h).Next, cross-linked Ni nanosheets were prepared on NCFs by a hydrothermal method (Figure 7i).Finally, after annealing under an N 2 condition, the CNTs-Ni@NCFs were obtained (Figure 7j).The interlayer spacing of 0.34 nm is corresponding to the CNT on the edge of the NCF, and in the core of the CNT, Ni nanoparticles can be observed on the outer layer, indicating that Ni catalyzed the growth of CNTs (Figure 7k,l).The CNT-Ni@NCF electrodes contain lithiophilic Ni nanoparticles and have a uniform distribution, which can lower the interfacial energy and reduce the volume expansion during repeated cycles. [44]When fabricated into symmetric cells, the CNT-Ni@NCF-Li electrodes showed a stable overpotential during 300 cycles at 30 mA cm −2 and 1 mAh cm −2 (Figure 7m).
According to the research summarized above, a CNT-based host design could effectively enhance the electrochemical performances of LMAs, especially rate performance. [44,72]CNT-based hosts can be directly electrodeposited Li as an electrode to replace the "hostless" lithium metal, and can also be thermally infused by melt lithium as an electrode because of their good lithiophilicity.When fabricated into an anode for LMBs, CNT-based nanomaterials can supply abundant spacing and active sites for Li plating, which could lower local current density and alleviate volume change during the Li plating and stripping process, further suppressing Li dendrite formation.

Decorating Current Collectors
So far, current collectors such as copper, titanium, nickel, and stainless steel mesh (SSM) have been applied in LMAs because of their unique properties. [46,88,116,117]Copper has high conductivity, and SSM has a certain mechanical property. [46,82]However, the electrochemical properties of the current collector without modification used for LMA are usually unsatisfactory.At this point, the combination of current collectors and CNT-based nanomaterials can effectively enhance their conductivity and mechanical stability, improving the electrochemical performances of LMA.
Cu is one of the most commonly used current collectors in LMAs. [118]Nevertheless, pristine Cu often has a large overpotential and an unsatisfactory Columbic efficiency during the Li plating/stripping process, which impedes its further application in LMAs. [82]To improve the stability and electrochemical performances of Cu-based LMAs, introducing CNT-based nanomaterials to modify Cu is an effective strategy. [41,75,77,80,82,93]For example, in 2018, Zuo et al. [75] modified Cu foil with a nano-SiO 2 microsphere/CNT composite and used it in LMAs.The modified electrode outperformed the pristine Cu foil in terms of Columbic efficiency.Moreover, Zhao et al. [76] developed Znand N-co-doped carbon sheets with CNTs to modify Cu foil as an electrode material for LMAs (Figure 8a).First, the Zn-Co-based metal-organic framework precursor was spin-coated onto Cu foil.After treatment in an argon-filled tube furnace at 800 °C for 2 h, the precursor was transformed into an N-doped carbon layer with Zn atoms (Zn-NC) (Figure 8b).Next, the CNTs were grown on Zn-NC via the CVD method (Zn-NC-CNT) (Figure 8c).In the interior of the CNTs, Zn and Co are uniformly distributed (Figure 8d,e).During the lithiation process, the protective layer of Zn-NC-CNT can guide the lithium-ion flux and induce uniform lithium deposition on the Cu foil, which was attributed to the strong affinity of the Zn atom with lithium. [76]When fabricated into symmetric cells, the Zn-NC-CNT-Cu electrode can cycle for 1000 h at 2 mA cm −2 and 2 mAh cm −2 (Figure 8f).When tested in full cells, the Zn-NC-CNT-Cu¦¦NMC811 exhibited superior rate capabilities than pristine Cu¦¦NMC811 (Figure 8g).Inspired by this method, Ma and co-workers developed an Al-based mesoporous carbon nanorods/CNT (Al-PCR/CNT) for modifying Cu foil (Figure 8h). [47]he Al-PCR sample was fabricated via a hydrothermal method, followed by a calcination process.Then, after the CVD process, an Al-PCR/CNT sample was obtained (Figure 8i,j).The Al-PCR storage units Ref.
Energy Environ.Mater.2023, 6, e12525 Al-PCR/CNT modification layer contains the atomically dispersed Al lithiophilic sites with strong Li affinity, meanwhile, the CNTs provide sufficient units to store lithium and improve the mechanical properties of the structure. [47]n addition to Cu foil, stainless steel mesh has also been employed as a current collector for Li metal anode due to its excellent mechanical property. [46,119]For example, Zhao et al. [46] synthesized nitrogendoped CNT arrays on a stainless steel mesh (N-CNT@SSM) via an in situ catalyze method.The N-doped CNTs can provide many N-functional groups, which can enhance the lithiophilic properties of stainless steel, improving uniform Li deposition.When evaluated in the symmetric cell, the N-CNT@SSM electrodes cycled for more than 600 h.The excellent performance can be ascribed to the N-CNT arrays, which regulate lithium-ion flux, resulting in homogeneous lithium deposition on the electrode. [46]he performance of LMBs could be enhanced through CNT-based nanomaterials modifying current collectors.According to the discussion above, the common ground could be summarized in these CNT-based hybrid current collectors: 1) CNTs have enough space to store lithium, reducing volume expansion during cycling; 2) they possess outstanding mechanical properties; and 3) they usually depict excellent cyclic stability under high current densities.

Building Artificial SEI on Li Metal Surface
Li metals are very active because of low redox potential, and the high reactivity of Li metals and its infinite volume expansion during cycling would cause unstable SEI and dendritic growth, leading to short-circuiting of the batteries or even safety hazards. [120,121]Therefore, building artificial SEI on Li metal surface is a promising strategy to achieve dendrite-free LMAs and realize large-scale fabrication. [112,122]Different CNTs-based composites have been used to decorate lithium metal surfaces, which have been proved to be capable of regulating uniform lithium deposition and inhibiting dendrite growth. [123]For example, Zhang et al. [42] prepared a highly stable LMA by dripping a composite solution of CNT and ZnO onto Li foil (GZCNT) (Figure 9a,b).After 520 cycles at 1 mA cm −2 and e) The rate capability of the LFP/Li@G-ZGC cells.f) Optical images of the Li@G-ZGC after rate cycling; scale bar: 50 μm.Reproduced with permission. [71]Copyright 2021, Elsevier.g) Schematic diagram of the synthesis of CNT-Ni@NCFs.SEM image of h) the NCFs precursor, i) the Ni@NCFs precursor, and j) the CNT-Ni@NCFs.k) Highresolution transmission electron microscopy (HRTEM) and l) elemental mapping images of CNT-Ni@NCFs.m) Galvanostatic voltage profiles of CNT-Ni@NCFs-Li electrodes.Reproduced with permission. [44]Copyright 2021, Elsevier.
Energy Environ.Mater.2023, 6, e12525 1 mAh cm −2 , the dense lithium deposition was formed and no cracks or dendrites could be observed (Figure 9c).The GZCNT layer has a 3D structure inside, which provides abundant deposition sites.The topmost lithiophobic CNT layer has a porous structure, which facilitates Li diffusion and hinders the formation of dendrites, whereas the bottom lithiophilic ZnO/CNT layer is fixed on the Li foil, which effectively ensures uniform lithium plating through regulating deposition behavior and prevents the formation of corrosion layer between the CNTs and the lithium foil (Figure 9d).When investigated in symmetric cells, the interface resistance (R SEI ) of the cells made of GZCNT-coated Li electrodes decreased to 30 Ω, whereas that of CNT-coated electrodes increased to 158 Ω after 500 cycles at 5 mA cm −2 (Figure 9e).Furthermore, when fabricated into pouch cells, the GZCNT-Li pouch cell showed a stable cycling performance for 210 cycles at 1 mA cm −2 and 1 mAh cm −2 , which is superior to the performance of the Li pouch cell (Figure 9f). [42]n addition, Lin and co-workers [29] prepared a silk fiber-derived carbon and CNTs composite (SFC/CNT) layer on the lithium metal surface  [76] Copyright 2021, Royal Society of Chemistry.i) HRTEM image and j) the selected area electron diffraction (SAED) patterns of Al-PCR/CNT.k) TEM image and HRTEM image (inset) and l) HAADF-STEM image with the corresponding EDS elemental mapping of the Al-PCR/CNT.m) Galvanostatic cycling performance of Al-PCR/CNT electrodes.n) The discharge capacity and Coulombic efficiency of the full Al-PCR/CNT¦¦NCM-811 cells.Reproduced with permission. [47]Copyright 2021, Elsevier.
Energy Environ.Mater.2023, 6, e12525 by a mechanical transfer method (Figure 9g).First, the SFC/CNTs were mixed with polyvinylidene difluoride (PVDF) to form a homogeneous slurry.Second, the SFC/CNTs slurry was cast on polyethylene glycol terephthalate (PET) film and then dried under vacuum at 80 °C for 12 h.Third, the SFC/CNT-cast PET films were placed on the Li chips with direct contact.Finally, under a pressure of 350 kPa and heating at 140 °C for 1 h, an SFC/CNT layer on the Li chips was obtained after the PET film was removed.This novel method not only avoided solvent corrosion during the fabrication of the SFC/CNT layer but also constructed a high-quality buffer layer. [29]Because the interlaced CNTs inside carbon flakes can provide sufficient pores (Figure 9h,i), a symmetric cell with SFC/CNTs-Li exhibited superior cyclic performance for 600 h at 2 mA cm −2 and 2 mAh cm −2 (Figure 9j).Furthermore, LFP¦SFC/CNTs pouch cells have excellent cyclic stability of 390 cycles with 80% capacity retention at 0.5C, whereas LFP¦Li pouch cells are short-circuited after 180 cycles (Figure 9k).This excellent performance is attributed to the lithiophilic buffer layer with rich N/S heteroatoms, which not only facilitated homogeneous Li nucleation but also guided the subsequent uniform Li deposition. [29]ccording to the discussion above, the currently used building of artificial SEI on Li surface with CNT-based nanomaterials is mainly a mechanical pressing method. [29]Choosing the appropriate pressure is one of the most important factors: too small pressure could lead to a bad contact between the CNT-based nanomaterials and lithium, which is unable to effectively protect LMAs.However, too much pressure could damage the composite electrodes, which is destructive to uniform Li deposition. [35]Apart from the mechanical pressing methods, other strategies need to be developed for further studies.For instance, drip coating CNT-based solutions on the lithium metal surface to fabricate CNT-decorated composite LMAs may be one promising direction. [42]

Others
In addition to the above-mentioned strategies, other strategies have also introduced CNT-based nanomaterials to improve the LMAs, including modifying separators and introducing SSEs. [30,114]Modifying separators is an effective strategy to regulate uniform Li deposition. [124]For instance, Pan et al. [30] prepared a double-sided conductive separator for LMBs.The double-sided conductive separators have a sandwich structure with CNT/CNF layers on each side of the glass-fiber/CNF membrane.The porous and conductive CNT/CNF layer can enhance the Li deposition surface area compared to a pristine lithium anode, which can reduce LCD during Li deposition and stripping. [30]Symmetric pouch cells with a double-sided conductive separator depicted superior cyclic stability over 280 h at 1 mA cm −2 and 1 mAh cm −2 .In addition, Liu et al. [96] prepared one side ZIF-67@CNTs conductive separator for LMBs.The CNTs film can suppress dendritic growth as the 3D framework, and the microporous structure of ZIF-67 nanoparticles can regulate the distribution of the transport of Li + . [96]Due to the synergistic effect, when evaluated in the symmetrical cell, the Li@PCF/Z-67 electrode exhibited a stable polarization voltage for 1000 h at 1 mA cm −2 and 1 mAh cm −2 .Moreover, solid-state batteries with inorganic garnet (Li 7 La 3 Zr 2 O 12 ) solid electrolytes are promising due to their excellent safety. [125]However, the poor contact between the solid electrolytes and the Li metals impedes their further application. [126]To deal with this issue, Fuchs et al. [114] reported a composite anode that contained lithium metal and CNTs paired with a garnet-type solid electrolyte.The CNT framework could help to maintain tight contact between Li 7 La 3 Zr 2 O 12 and Li during plating and stripping, and improve the mechanical properties of lithium anodes, leading to enhanced electrochemical performance of LMAs. [114]ections 4.1.1-4.1.4discussed the strategies of application of CNT-based nanomaterials in Li metal anode, which includes CNT-based Li host design, decorating the current collectors, building artificial SEI on lithium metal surface, and so on.Among them, the CNT-based host design is the most widely used strategy for LMAs, because it could effectively lower LCD and accommodate volume expansion during Li plating and stripping, and exhibited excellent rate performance. [83]Furthermore, compared with pristine current collectors, the modified current collectors with CNT-based nanomaterials used for composite LMAs showed superior mechanical stability and cycling performance. [46,93]Moreover, Building artificial SEI on lithium metal surface with CNT-based nanomaterials is facile to operate, and it is one of the most promising strategies due to its low cost and simplicity. [42]In addition, applying double-sided modified separators for LMAs could improve the performance of both LMA and the composite cathode, leading to greatly improved electrochemical performance of the corresponding Li metal full cells. [30]And introducing SSEs for LMAs can improve the safety and mechanical performance of AMBs. [114]

The Application of CNT-Based Nanomaterials in Na Metal Anodes
Sodium metal anodes (SMAs) are considered a sustainable alternative to LMAs because of their high theoretical specific capacity (~1166 mAh g −1 ), low electrochemical potential (~2.714 V), and abundant reserves on the Earth. [31]However, the application of sodium metal batteries has been impeded by several difficult challenges such as poor cyclic stability, low Coulombic efficiency, and safety issues resulting from the uncontrollable growth of Na metal dendrite. [97,127]The strategies involving CNT-based nanomaterials that have been investigated for use in SMAs mainly include CNT-based Na host design and building artificial SEI on Na metal surface. [31,57,97,100,102,105,107]

CNT-Based Na Host Design
To regulate Na deposition and relieve the uncontrolled volume expansion during Na plating and stripping, improving the sodiophilicity of CNT-based materials as host materials for SMAs is one of the most important solutions. [98,99,104,106]For instance, Ye et al. [98] synthesized an oxygen-functionalized CNT network (O f -CNT) as a host for SMAs.The O f -CNT was obtained after treating pristine CNT networks with oxygen microwave plasma at 100 W for tunable oxygen functionalization.During the plating process, the O f -CNT networks could guide Na to homogeneously nucleate around the functional groups and fill the voids within the O f -CNT skeleton (Figure 10a).The pristine O f -CNT exhibited an interconnected structure, and with the increasing Na deposition, Na can be deposited uniformly along the O f -CNT (Figure 10b).After 200 cycles at 5 mA cm −2 , the structure of the Na@O f -CNT electrode can be still maintained without breaking when the Na has been completely stripped, indicating its excellent flexibility and reversibility (Figure 10b).The Na@O f -CNT electrode in a symmetrical cell depicted an outstanding long cycling performance over 4500 h at Energy Environ.Mater.2023, 6, e12525 1 mA cm −2 and 1 mAh cm −2 (Figure 10c).Even during cycling at a large current density of 10 mA cm −2 , the Na@O f -CNT electrodes can maintain a stable overpotential without short-circuiting (Figure 10d).The excellent cyclic and rate performance of the Na@O f -CNT electrodes is attributed to the sodiophilic and robust oxygen-functionalized CNT networks, which improve uniform sodium nucleation and plating behaviors to obtain dendrite-free SMAs. [98]Later, Lin et al. [99] prepared an ionic-electronic dualconducting host with Na 3 P and CNTs for SMAs.This in situ generated Na 3 P has excellent sodiophilicity, and the Na 3 P participates with CNTs to form a highly electronic conductivity path, which can also reinforce SEI.When full Na 3 V 2 (PO 4 ) 3 (NVP) cells were fabricated, the full ionic-electronic dual-conductor-based cells exhibited superior rate performance. [99]urthermore, CNTs have been composited with other carbon-based materials as hosts for SMAs, such as rGO and MXene, which can also effectively improve their rate performance. [100,101,105]For example, Wang et al. [100] fabricated a V 2 CT x /rGO-CNT microgrid aerogel via 3D printing technology (Figure 10e).First, V 2 AlC was changed into multilayer V 2 CT x MXene through etching with HF acid.Subsequently, V 2 CT x MXene nanoflakes were incorporated into a GO-CNT aqueous slurry to fabricate the 3D-printing ink.After 3D printing the V 2 CT x /GO-CNT ink with the designed pattern, the V 2 CT x /GO-CNT hydrogel was freezedried to form an aerogel, followed by annealing under argon gas conditions to convert the GO into rGO.Finally, a 3D-printed V 2 CT x /rGO-CNT microgrid aerogel was obtained, which is composed of randomly distributed CNT, rGO, and V 2 CT x nanoflakes (Figure 10f,g).The CNTs in V 2 CT x /rGO-CNT could increase the surface area of the electrode to lower the LCD, and enhance the mechanical robustness of the electrode's structure to enhance the long-term cycling stability. [101]When tested in symmetric cells, the V 2 CT x /rGO-CNT electrodes exhibited excellent cycling performance over 900 h without short-circuiting at a capacity of 50 mAh cm −2 under 5 mA cm −2 (Figure 10h).The artificial hierarchical 3D-printed structure can promote electrolyte  and c) after cycling.Scale bar, 1 μm.d) Schematic illustration of Li deposition onto GZCNT-coated Li electrode.e) Summary of the R SEI fitting results of pristine cells and after 500 cycles.f) Galvanostatic cycling performance of pouch cell fabricated from GZCNT-coated Li electrode.Reproduced with permission. [42]Copyright 2018, the author(s).g) The schematic diagram for fabrication of the SFC/CNTs-Li electrode.h) SEM and i) TEM images of SFC/ CNTs.j) Cycling performance of symmetric cells with SFC/CNTs-Li.k) Cycling performance of LFP¦SFC/CNTs-Li pouch-cell.Reproduced with permission. [29]opyright 2022, Wiley-VCH.
Energy Environ.Mater.2023, 6, e12525 penetration to ensure tight contact between the electrolyte and the active components, and it also significantly increased the ionic transport rate and reduce electrode's volume change, which could be a research direction in the future. [100]n addition, CNT-based materials modifying other hosts can also obtain excellent performance for SMAs. [57,97,107]For example, Wang's group [97] developed gold nanoparticles (AuNPs) and CNTmodified carbon cloth (Au-CNT/CC) as a sodiophilic host for SMAs.The interconnected CNTs were grown on the carbon cloth through the CVD method, and then AuNPs were deposited on the CNT/CC via the magnetron sputter process (Figure 10i).The AuNPs were uniformly distributed on the CNTs' surface (Figure 10j,k) and effectively guided the uniform Na deposition.When fabricated into symmetric batteries, the Au-CNT/CC electrode exhibited excellent cyclic performance for 1600 h at 1 mA cm −2 and 2 mAh cm −2 (Figure 10l).Furthermore, the flexible Na@Au-CNT/CC¦¦NVP@C pouch cell could power a LED at bending angles ranging from 0°to 180°, indicating that the Au-CNT/CC electrode has great potential for flexible electronics (Figure 10m,n).The outstanding performance of Na@Au-CNT/CC electrodes can be ascribed to the following advantages: 1) the flexible carbon cloth provides a scaffold with excellent mechanical properties to support the entire electrode; 2) the interconnected CNTs on the carbon cloth have a large surface area, which could effectively decrease the LCD and provide abundant space to buffer the volume change during charge-discharge cycles; and 3) the sodiophilic AuNPs can reduce the Na metal's nucleation energy barrier, regulating the Na + flux and guiding the Na uniform deposition. [97]2.2.Building Artificial SEI on Na Metal Surface Sodium metals have higher reactivity than lithium metals.Similar to LMAs, the liquid electrolyte of SMAs can corrode the Na metal surface, which could result in SEI unstable and Na dendrites growth, even leading to short-circuiting of the Na metal batteries.[128] Therefore, decorating the Na metal surface by constructing an artificial SEI also has enormous potential to solve the above-mentioned issues.The use of CNT-based nanomaterials to build artificial SEI on SMA has been reported.[31,102] For example, Sun et al. [102] introduced an N, S co-doped CNT (NSCNT) interlayer to protect SMAs (Figure 11a).The NSCNT was fabricated by a thermal pyrolysis method, and the NSCNT maintained the nanotube morphology covered by an amorphous carbon coating layer (Figure 11b,c).The Na/NSCNT electrodes in symmetric cells still showed a low overpotential after cycling for 500 h at 1 mA cm −2 and 1 mAh cm −2 (Figure 11d).The N and S functional groups on the CNTs enhanced the sodiophilicity of NSCNTs, which guided the sodium nucleation and directed the Na to uniformly deposit on the NSCNT papers without dendrites formation (Figure 11e,f).Moreover, Sun et al. used a gradient CNT film decorated with Ag particles (grad-Ag@CNT) to modify the Na metal surface.[31] The CNT film was fabricated and immersed in an AgNO 3 solution, and after the magnetic stirring of the mixture for 1 h, Ag nanoparticles were distributed along the direction of the thickness in the CNT framework and the grad-Ag@CNT sample was obtained (Figure 11g).When grad-Ag@CNT/Na is used as an electrode for SMAs, the sodiophilic Ag as nucleation sites can induce uniform Na deposition (Figure 11h).With an increased capacity of Na deposition at a current density of 8 mA cm −2 , the grad-Ag@CNT/Na anodes depicted a flat surface without dendrite growth (Figure 11i). When bricated into symmetric cells, the grad-Ag@CNT/Na electrodes showed outstanding cyclic capability for 1000 h at 8 mA cm −2 and 8 mAh cm −2 (Figure 11j).This grad-Ag@CNT layer is lightweight and has good mechanical strength, which can regulate Na deposition in the Ag-rich region and effectively prevent the formation of Na dendrites.[31] 4.3.The Application of CNT-Based Nanomaterials in K Metal Anodes Compared with Li and Na metals, K metal have even higher reactivity, so preservation and reversible plating and stripping in electrolytes of K metal remain key difficulties, which hinder its development as an anode for potassium batteries.[129] Due to the "hostless" nature of K metal, the SEI film formed when K metal spontaneously reacts with electrolytes, could be easily broken during the repeated K plating and stripping processes, which induces nonuniform K + flux and results in the formation of K dendrite.[130] The research and development of K metal anodes (PMAs) are still in an early stage. To date,8,109] For example, Tang et al. [32] fabricated K@DN-MXene/CNT electrode via etching and vacuum-filtering processes, followed by molten K infusion (Figure 12a-c).After plating K at a capacity of 5 mAh cm −2 , the surface of K@DN-MXene/CNT electrode remained uniform (Figure 12d).When investigated by in situ optical microscopy, K dendrites appeared on the Cu foil, whereas the surface of DN-MXene/CNT was still flat and smooth (Figure 12e).When fabricated into K-S batteries, the K@DN-MXene/CNT¦¦SPAN cell showed better rate capability and cycling performance than bare K foil (Figure 12f,g).These superior electrochemical performances of the K@DN-MXene/CNT electrode could be attributed to the high SSA of the DN-MXene/CNT host, which not only reduced the LCD but also improved the electronic conductivity.[32] In addition to host design, Wang et al. [108] designed and fabricated a CNT film as an artificial SEI on PMAs.The freestanding CNT film was immersed in an ester-based electrolyte and kept in tight contact with K foil via mechanical processing (Figure 12h).The CNT layer was tightly anchored to the K foil, even after cycling for about 2000 h, the homogeneous amorphous SEI maintained its thickness without any breaks (Figure 12i,j), and the inorganic and organic components were randomly distributed in the SEI (Figure 12k).When investigated in a symmetrical cell, the K/CNT electrode could sustain 1100 cycles without short-circuiting at 1 mA cm −2 and 1 mAh cm −2 (Figure 12m).The outstanding cycling stability can be ascribed to a stable SEI layer on CNT framework, improving the K affinity to induce K nucleation and regulating the K deposition behaviors (Figure 12l).[108]

Summary
In Sections 4.1-4.3, the application of CNT-based nanomaterials in AMAs has been reviewed.Based on the discussions above, the rate capability and cyclic stability performance of AMBs could be efficiently enhanced by introducing CNT-based nanomaterials, which is due to the following aspects: First, CNT-based nanomaterials could enlarge the    [102] Copyright 2018, WILEY-VCH.g) Optical photos and top-view SEM images of grad-Ag@CNT.h) Schematic illustration of the plating process.i) In situ optical images of Na deposition behavior on grad-Ag@CNT/Na.j) Cycling performance of the grad-Ag@CNT/Na electrode.Reproduced with permission. [31]Copyright 2021, American Chemical Society.
Figure 10.a) Schematic illustration of the Na deposition process on the O f -CNT host.b) SEM images of the Na@O f -CNT electrode with deposited Na at 0, 5 and10 mAh cm −2 , and after 200 cycles of plating and stripping; scale bars: 500 nm.c) Cycling performance and d) rate performance of Na@O f -CNT symmetric cells.Reproduced with permission. [98]Copyright 2019, Wiley-VCH.e) Schematic illustration of the preparation process of V 2 CT x /rGO-CNT electrodes.f, g) SEM images of the V 2 CT x /rGO-CNT microgrid aerogel.h) Cycling performance of the V 2 CT x /rGO-CNT electrode.Reproduced with permission. [100]Copyright 2022, American Chemical Society.i) Schematic illustration of the fabrication process of the Au-CNT/CC host.j) SEM images and k) TEM images of Au-CC/CNT host.l) Cycling performance of Au-CNT/CC electrodes.The Na@Au-CNT/CC¦¦NVP@C pouch cell: m) bending at various angles with a light-emitting diode light and n) schematic diagram and actual wristband with the pouch cell.Reproduced with permission. [97]Copyright 2021, Elsevier.
Energy Environ.Mater.2023, 6, e12525 can make our research more accurate and reliable, and have been utilized in the research of AMAs.Here, advanced characterizations of the reaction mechanism of CNT-based nanomaterials in AMAs are discussed.
In situ transmission electron microscopy (TEM) is an effective technique used to observe dendritic growth and visualize the dynamic electrochemical process of AM on a CNT-based electrode.For example, Wang's research group [131] conducted in situ TEM to reveal the growth and dissolution kinetics of confined Li (Figure 13a,b).Once contact was established between the nanotube and the reverse Li/Li 2 O electrode, a positive bias of 3 V was applied to initiate the Li plating process.First, as a result of Li + insertion, the nanotube expanded from 5.8 to 9.9 nm, then the Li metal nucleated at the bottom of the nanotube and grew upward during the plating process (Figure 13b).After Li reached the top of the nanotube, the nanotube caused an electrical breakdown (Figure 13b), which can be used to simulate a lithium dendrite crossing a solid electrolyte, causing an internal short circuit. [131]n addition, Yang's research group explored the deposition behavior of Na metal on the V 2 CT x /rGO-CNT aerogel by in situ TEM (Figure 13c,  d). [100]Figure 13d shows that the Na particles are gradually nucleated and grew in the center of the V 2 CT x nanoflakes and the Na metals had a uniform distribution on the surface of the V 2 CT x /rGO-CNT nanoflakes.
In situ X-ray diffraction (XRD) techniques are widely utilized to investigate the structural evolution, capacity decay, and energy storage mechanisms of batteries during the electrochemical process, as these have important implications for battery manufacturing and optimization. [132]In situ XRD technique has also been used to investigate the AM deposition behavior on CNT-based nanomaterials for AMAs. [32,99]or instance, Tang et al. [32] investigated K deposition behavior on a DN-MXene/CNT scaffold via in situ XRD.In the first plating process, the peak appeared at 29.6°corresponding to K 2 CO 3 (112), which represents the products of side reactions in the electrolytes.The intensity of the peak gradually increased in the cell with a copper current collector but one remained low without any obvious change with the DN-MXene/CNT host, indicating that the DN-MXene/CNT host could effectively alleviate the side-reaction (Figure 13e).
In addition to in situ TEM and in situ XRD, other advanced characterizations, such as cryogenic TEM (cryo-TEM), have also been used for analyzing the reaction mechanism of CNT-based nanomaterials for f) Rate performances and g) cycling performance of K@DN-MXene/CNT¦¦SPAN K-S battery.Reproduced with permission. [32]Copyright 2019, WILEY-VCH.h) Schematic illustration of the synthesis procedure of the K/CNT electrode.TEM image of i) CNT film and j) K/CNT electrode after 2000 h cycles at 5 mA cm −2 .k) Schematic diagram of the structure of SEI; scale bar, 5 nm.l) Schematic diagram of the K deposition behavior of the K/CNT electrode.m) Cycling performance of the symmetric K/CNT¦K/CNT cells.Reproduced with permission. [108]Copyright 2019, WILEY-VCH.
Energy Environ.Mater.2023, 6, e12525 AMAs.Cryo-TEM is another advanced technique employed to visualize the structural evolution of AM dendrite, SEI, and products of side reactions.Different from conventional TEM, cryo-TEM could maintain high resolution at low temperatures, making it suitable for characterizing fragile samples. [133]Yang et al. [133] directly observed lithium metal growing in the cavities of CNT by the cryo-TEM technique.They confirmed that the iron carbide can induce the Li + to pass through the defective CNT walls and enter the cavity, and the Fe 2 C was detected within the cavity of the multi-walled CNT rather than on the surface, indicating Li nucleation and growth in the CNT cavities (Figure 13f-h).
These studies highlighted the importance of using advanced characterization techniques to elucidate the electrochemical behaviors of AMAs during the plating and stripping process and guide further research into dendrite-free AMAs, especially in situ characterizations. [34]owever, each technique has its limitations, and we need to explore more advanced techniques to characterize the electrochemical behavior of AMAs.

Conclusions and Outlook
In conclusion, this review summarizes recent progress on the applications of CNT-based nanomaterials in AMAs.First, the categories, merits, existing issues, and the corresponding stabilizing strategies of AMAs are summarized.Furthermore, the structures, synthesis, and properties of CNTs are introduced.The synthetic methods of CNTs for AMAs are including the arc discharge method, laser ablation method, CO gasphase catalyzed growth method, and the CVD method.Among them, the CVD method is the most commonly used method.The CNT-based nanomaterials for AMAs have multiple advantages: 1) CNT-based nanomaterials can enlarge the SSA of the composite AMAs to relieve the volume change during the charging and discharging processes of AMAs; 2) CNT-based nanomaterials have high electronic conductivity and possess an excellent affinity to AM can effectively regulate AM ion transport and deposition behavior, enhancing the rate capability of AMBs; and 3) CNTs have good mechanical property enables the excellent flexibility and cycling stability of CNTs-based composite AMAs.In addition, various strategies have been used for CNT-based nanomaterials in AMAs, including AM host design, modifying the current collector, building artificial SEI on AM surface, modifying separators, and introducing SSEs.Among them, the AM host design is the most widely used, and the building of artificial SEI on AM surface has great promise because of its low cost and simplicity.Although some progress has been made on CNT-based nanomaterials in AMAs, there are still many challenges to be solved.In our opinion, to further improve the application of CNT-based AMBs, researchers should focus on the following aspects (Figure 14):  [131] Copyright 2021, Wiley-VCH.c) Schematic illustration of the in situ nanobattery setup.d) In situ TEM images of the V 2 CT x /rGO-CNT during the sodium plating process.Reproduced with permission. [100]Copyright 2022, American Chemical Society.e) In situ XRD patterns of the Cu foil and DN-MXene/CNT scaffold deposited K. Reproduced with permission. [32]Copyright 2019, WILEY-VCH.Cryo-TEM images of the morphology of f) the microstructure of the CNTs containing iron carbides and the magnified views of the g) white and h) yellow squares.Reproduced with permission. [133]Copyright 2021, Elsevier.
Energy Environ.Mater.2023, 6, e12525 few studies on CNT-based materials for SSAMBs have been reported.The application of CNTs-based nanomaterials to modify SSEs should also be explored in the future.It is also necessary for researchers to explore the mechanism of the CNT-based nanomaterials used in SSAMBs.The following aspects need to be paid attention to 1) the problem of high interfacial impedance between SSEs and AMAs; 2) more advanced characterizations need to explore the reaction mechanism of CNT-based nanomaterials in SSAMBs; and 3) the difficulties of synthesizing CNT-based AMA and their pairing with various SSEs for SSAMBs.

The practical and large-scale application of CNT-based AMAs in
AMBs should be further investigated.In terms of fabricating CNT-based AMAs, there are several challenges to be considered and addressed: for example, AMs are very sensitive to water and air due to their high reactivity, hence, AMs should be stored and handled in a hermetically sealed environment with low oxygen and water content.In this regard, one of the possible solutions is to explore a durable and waterproof protective layer with CNT-based nanomaterials.The composite protective layer should have high flexibility and tightness to resist oxygen and water, and relieve the volume expansion of AMAs.For the CNT-based host materials for AMA, although their abundant nucleation sites and high conductivity can induce AM to uniformly plate in the framework, the limited area of lab-scale coin cell will hamper the large-scale production of CNT-based AMA.
Compared with a lab-scale coin cell, pouch cells have a larger area and higher capacity, and are closer to our daily lives.Hence, in this context, proper molds and corresponding equipment are very important for fabricating large-area CNT-based AMA for practical AMBs.
Overall, this review provides a fundamental understanding and recent advances in the application of CNT-based nanomaterials in AMAs.We hope this review could shed light on high-performance CNT-based AMAs and further promote their practical applications.
Fengzhang Ren and Prof. Yong Liu at Henan University of Science and Technology.His research interests include electrochemical energy storage materials.Yong Liu received his B.Eng. degree from Tianjin University in 2003.Before he received his Ph.D. from Tianjin University in 2012, he was a joint Ph.D. student under the supervision of Prof. Meilin Liu at Georgia Institute of Technology.He is currently an associate professor at School of Materials Science and Engineering at Henan University of Science and Technology.His research interests include electrochemical energy storage and conversion with batteries and nanostructured energy materials.Guangxin Wang received Ph.D. degree from Universität Bremen, Germany.He is currently a distinguished professor at School of Materials Science and Engineering at Henan University of Science and Technology.His current research is focused on the design and synthesis of novel nanostructured materials and their applications in advanced metallic materials and energy-relating materials.Jianmin Ma is a professor at the University of Electronic Science and Technology of China, Chengdu, China.He received his B.S. degree in chemistry from Shanxi Normal University in 2003 and a Ph.D. degree in materials physics and chemistry from Nankai University in 2011.During 2011-2015, he also conducted research in several overseas universities as a postdoctoral research associate.His research interest focuses on energy storage devices and components including metal anodes and electrolytes, and theoretical calculations from density functional theory and molecular dynamics to finite element analysis.Energy Environ.Mater.2023, 6, e12525

Figure 1 .
Figure 1.a) Schematic illustration of the application of carbon nanotube (CNT)-based nanomaterials in alkali metal anodes.b) Bar chart of research publications on the application of CNT-based nanomaterials in alkali metal anodes in recent years.c) Pie chart of the proportion of different alkali metal anodes with CNT-based nanomaterials in the published research articles.

Figure 2 .
Figure 2. The advantages and issues of alkali metal anodes.a) Theoretical specific capacity and b) voltage of alkali metal anodes compared with different anode materials.G, graphite; MCMB, mesophase carbon microbeads; HC, hard C; SC, soft C; LTO, Li 4 Ti 5 O 12 .c) Schematic diagram of the compositions and working mechanisms of alkali metal batteries.d) Uneven SEI film, e) volume expansion and the formation of dead alkali, and f) severe metal dendrite growth.

Figure 4 .
Figure 4.The methods of synthesis, structure, advantages, and disadvantages of CNTs.

Scheme 1 .
Scheme 1.The merits and applications of CNT-based nanomaterials in AMAs.

Figure 5 .
Figure 5. Scanning electron microscopy (SEM) images of the top surface and cross section of the a) CNT paper and b) Li/CNT composite electrode.c) Transmission electron microscopy (TEM) image of a multiwall CNT.d) Rate performances of Li/CNT symmetric cell at different current densities.Reproduced with permission.[70]Copyright 2018, WILEY-VCH.e) Schematic showing the fabrication process of Li-CSMFs composite anode.SEM images of f) CSMF film and g) Li-CSMF composite anode.h) In situ X-ray diffraction patterns of CSMF and LiCSMF in coin cell.i) Electrochemical impedance spectra of LiCSMF¦¦S and Li¦¦S pouch cells.j) The rotating electrical fan with LiCSMF¦¦S pouch cells.Reproduced with permission.[74]Copyright 2020, Wiley-VCH.

Figure 6 .
Figure 6.a) Schematic illustration of the synthesis process of the A f -CNT film.b) Digital photos of the A f -CNT film.c) SEM image of the A f -CNT film.d) TEM and the corresponding elemental mapping images of A f -CNT.e) SEM images of after depositing 2, 4, 6, or 8 mAh cm −2 of Li on A f -CNT film.f)Galvanostatic voltage profiles of A f -CNT@Li symmetric batteries.g) Voltage profiles of A f -CNT@Li¦¦LFP cell at particular cycles.Reproduced with permission.[113]Copyright 2021, Elsevier.h) Schematic representation showing the formation process of the CNT/NiO host and the CNT/NiO@Li composite electrode.i) SEM images of the CNT/NiO@Li composite anodes and the corresponding elemental mapping images.j) Optical microscopic images and k) optical surface profilometry image of the CNT/NiO@Li electrode.l) Cycling performance of CNT/NiO@Li electrodes.Reproduced with permission.[73]Copyright 2021, Wiley-VCH.

Figure 7 .
Figure 7. a) Schematic diagram of the formation of the G-ZGC host.b) SEM image of the G-ZGC host.c) Cross-sectional SEM image of the G-ZGC host.d) Density functional theory (DFT) simulation of the Li-ion concentration profiles and plating morphology of the G-ZGC electrode.e)The rate capability of the LFP/Li@G-ZGC cells.f) Optical images of the Li@G-ZGC after rate cycling; scale bar: 50 μm.Reproduced with permission.[71]Copyright 2021, Elsevier.g) Schematic diagram of the synthesis of CNT-Ni@NCFs.SEM image of h) the NCFs precursor, i) the Ni@NCFs precursor, and j) the CNT-Ni@NCFs.k) Highresolution transmission electron microscopy (HRTEM) and l) elemental mapping images of CNT-Ni@NCFs.m) Galvanostatic voltage profiles of CNT-Ni@NCFs-Li electrodes.Reproduced with permission.[44]Copyright 2021, Elsevier.

Figure 8 .
Figure 8. a) Schematic diagram of the fabrication and lithiation process of the Zn-NC-CNT-Cu electrode.b) SEM image and HAADF-STEM image (inset) of the Zn-NC electrode.c) SEM image of the Zn-NC-CNT composite.d) TEM and HRTEM images (inset) of the Zn-NC-CNT composite.e) High-angle annular dark field-scanning transmission electron microscopy (HAADF-STEM) image of the Zn-NC-CNT composite (the red circles indicate Zn).f) Cycling performance of the symmetrical cells with Zn-NC-CNT-Cu electrodes.g) The rate capability of full Zn-NC-CNT-Cu¦¦NMC-811 cells.h) Schematic diagram of the synthesis process of the Al-PCR/CNT composite.Reproduced with permission.[76]Copyright 2021, Royal Society of Chemistry.i) HRTEM image and j) the selected area electron diffraction (SAED) patterns of Al-PCR/CNT.k) TEM image and HRTEM image (inset) and l) HAADF-STEM image with the corresponding EDS elemental mapping of the Al-PCR/CNT.m) Galvanostatic cycling performance of Al-PCR/CNT electrodes.n) The discharge capacity and Coulombic efficiency of the full Al-PCR/CNT¦¦NCM-811 cells.Reproduced with permission.[47]Copyright 2021, Elsevier.

Figure 9 .
Figure 9. a) Schematic illustration of the synthesis procedure of Li foil coated with a GZCNT layer.SEM images of the GZCNT-coated Li electrode b) before and c) after cycling.Scale bar, 1 μm.d) Schematic illustration of Li deposition onto GZCNT-coated Li electrode.e) Summary of the R SEI fitting results of pristine cells and after 500 cycles.f) Galvanostatic cycling performance of pouch cell fabricated from GZCNT-coated Li electrode.Reproduced with permission.[42]Copyright 2018, the author(s).g) The schematic diagram for fabrication of the SFC/CNTs-Li electrode.h) SEM and i) TEM images of SFC/ CNTs.j) Cycling performance of symmetric cells with SFC/CNTs-Li.k) Cycling performance of LFP¦SFC/CNTs-Li pouch-cell.Reproduced with permission.[29]Copyright 2022, Wiley-VCH.
Energy Environ.Mater.2023, 6, e12525 SSA of the composite AMAs to buffer volume change during the charging and discharging processes of AMAs.Second, the high electronic conductivity and excellent affinity to AM of CNT-based nanomaterials can effectively regulate AM ion transport and deposition behavior, enhancing the rate capability of AMBs.Third, the good mechanical property of CNTs enables the excellent flexibility and cycling stability of CNTs-based composite AMAs.

5 .
Advanced Characterizations on the Reaction Mechanism of CNT-Based Nanomaterials in AMAs Advanced characterizations are essential to study the nucleation and growth behavior of AMs on different substrates, which can effectively guide the exploration of novel strategies to regulate the uniform deposition of AMs.Recently, in situ techniques have been developed, which

Figure 11 .
Figure 11.a) Schematic diagram of the cycling stability of Na/NSCNT electrode.b) TEM and c) HRTEM images of NSCNTs.d) Galvanostatic voltage profiles of Na/NSCNT electrodes in symmetrical cells.SEM images of the Na/NSCNT electrode e) before cycling and f) after 20 cycles.Reproduced with permission.[102]Copyright 2018, WILEY-VCH.g) Optical photos and top-view SEM images of grad-Ag@CNT.h) Schematic illustration of the plating process.i) In situ optical images of Na deposition behavior on grad-Ag@CNT/Na.j) Cycling performance of the grad-Ag@CNT/Na electrode.Reproduced with permission.[31]Copyright 2021, American Chemical Society.

Figure 12 .
Figure 12. a) Schematic showing the fabrication process of the K@DN-MXene/CNT electrode.SEM images of the b) DN-MXene/CNT host and c) the K@DN-MXene/CNT electrode.d) K@DN-MXene/CNT electrode after deposition of K metal.e) In situ optical images of K deposited on the DN-MXene/CNT.f)Rate performances and g) cycling performance of K@DN-MXene/CNT¦¦SPAN K-S battery.Reproduced with permission.[32]Copyright 2019, WILEY-VCH.h) Schematic illustration of the synthesis procedure of the K/CNT electrode.TEM image of i) CNT film and j) K/CNT electrode after 2000 h cycles at 5 mA cm −2 .k) Schematic diagram of the structure of SEI; scale bar, 5 nm.l) Schematic diagram of the K deposition behavior of the K/CNT electrode.m) Cycling performance of the symmetric K/CNT¦K/CNT cells.Reproduced with permission.[108]Copyright 2019, WILEY-VCH.

Figure 13 .
Figure 13.a) Schematic illustration of the in situ TEM experimental setup.b) In situ TEM images of the upward growth of lithiation and Li filling in a CNT and its final breakdown.Reproduced with permission.[131]Copyright 2021, Wiley-VCH.c) Schematic illustration of the in situ nanobattery setup.d) In situ TEM images of the V 2 CT x /rGO-CNT during the sodium plating process.Reproduced with permission.[100]Copyright 2022, American Chemical Society.e) In situ XRD patterns of the Cu foil and DN-MXene/CNT scaffold deposited K. Reproduced with permission.[32]Copyright 2019, WILEY-VCH.Cryo-TEM images of the morphology of f) the microstructure of the CNTs containing iron carbides and the magnified views of the g) white and h) yellow squares.Reproduced with permission.[133]Copyright 2021, Elsevier.

1 .
More CNT-based nanomaterial synthesis methods and different advanced CNT-based host structures for AMAs should be explored.To date, most of the current CNTs for AMAs are synthesized by the CVD method.CNTs fabricated via different methods have diverse tube diameters, which influenced the deposition behavior of AMAs.Hence, synthesizing CNT-based nanomaterials by various methods for AMAs should be extensively studied.The fabrication of CNT-based nanomaterials usually involves complicated reactions in a harsh environment, which is inappropriate for practical production.Therefore, it is critical to investigate environmentally friendly, mass-produced and controllable synthesis methods.Moreover, the current CNTbased host materials are usually composited with other carbonbased materials such as rGO, carbon nanofibers, and MXene.Hence, developing pure CNT-based host materials (such as CNT aerogel) would be an important but challenging direction.2. Research into Na metal anode and K metal anode should be further developed.So far, most studies have focused on the application of CNT-based nanomaterials in LMAs.However, there are still few studies on the application of CNT-based nanomaterials in SMAs and PMAs.With the increasing consumption of lithium metal, Na and K are candidates for replacing Li due to their abundance and low cost.Similar to LMBs, we should pay more attention to sodium metal batteries and potassium metal batteries, using various characterization methods to analyze their working mechanisms.In addition, CNT-based materials could be developed and applied to other metal anodes like Mg, Ca, Al, and Fe metal anodes to deal with their inherent problems.3. The mechanism of CNT-based nanomaterials for AMAs still needs to be explored.In situ characterization techniques are required to obtain insights into the time-dependent electrochemical behaviors of CNT-based AMAs during the charging and discharging process, including in operando optical microscopy, in situ XRD, in situ SEM, in situ TEM, and in situ NMR.The use of various advanced characterization tools to investigate the reaction mechanism of CNT-based nanomaterials in AMAs is expected in the future.Besides, a combination of theoretical calculation (e.g., DFT calculation, molecular dynamic simulation, COMSOL simulation, etc.) and in situ characterization are essential to illustrate further understanding of the deposition behaviors of AMs.In addition, other advanced technologies including artificial intelligence and machine learning can be coupled with theoretical calculation to promote the development of CNTbased nanomaterials in AMAs. 4. Solid-state AM batteries (SSAMBs) have huge potential.Whereas,

Figure 14 .
Figure 14.Schematic illustration of the future direction of CNT-based nanomaterials in alkali metal anodes.

Table 1 .
Summary of the methods of synthesis and the battery performance of AMAs improved by CNT-based nanomaterials.