The Progress of Cobalt-Based Anode Materials for Lithium Ion Batteries and Sodium Ion Batteries

Limited by the development of energy storage technology, the utilization ratio of renewable energy is still at a low level. Lithium/sodium ion batteries (LIBs/SIBs) with high-performance electrochemical performances, such as large-scale energy storage, low costs and high security, are expected to improve the above situation. Currently, developing anode materials with better electrochemical performances is the main obstacle to the development of LIBs/SIBs. Recently, a variety of studies have focused on cobalt-based anode materials applied for LIBs/SIBs, owing to their high theoretical specific capacity. This review systematically summarizes the recent status of cobalt-based anode materials in LIBs/SIBs, including Li+/Na+ storage mechanisms, preparation methods, applications and strategies to improve the electrochemical performance of cobalt-based anode materials. Furthermore, the current challenges and prospects are also discussed in this review. Benefitting from these results, cobalt-based materials can be the next-generation anode for LIBs/SIBs.


Introduction
With the increasing consumption of fossil energy, more and more attention has been paid to the development and utilization of renewable energy. However, new energy resources like wind, solar and tidal power are not sustainable by virtue of their uncertainty over time. In this scenario, the development of high-performance energy storage devices is highly necessary. To this end, rechargeable batteries (RBs) are of particular interest due to the multiple advantages of rational working voltage, high capacities and long cycling life [1][2][3][4]. Lithium ion batteries (LIBs), as one of the dominant RBs, have gradually penetrated into many aspects of our lives, such as portable electronic devices [5], consumable electronics [6] and electric vehicles [7]. However, the large-scale application of LIBs is hobbled by the rareness of lithium resources in the earth's crust [8,9]. Sodium ion batteries (SIBs) have captured great concerns as a complementary technology for LIBs by virtue of abundant raw materials. Furthermore, many electrode materials of LIBs can be applied as a drop-in replacement for SIBs because of a similar conversion reaction mechanism with LIBs.
There is a growing demand for the energy density of RBs. For LIBs, the mainstream anode materials-graphite materials are unable to meet the ever-increasing requirement of energy density because of the relatively lower theoretical specific capacity (~372 mA h g −1 ). As for SIBs, the larger radius of Na + (1.02 Å vs. 0.76 Å of Li + ) gives rise to severe volume fluctuations of electrode materials and sluggish diffusion kinetics, finally yielding poor electrochemical performances of SIBs. Up to now, there have been no right anode materials of SIBs for practical applications [10]. Therefore, a key requirement for both LIBs and SIBs is the development of anode materials with applicable working voltage, marvelous cycle performance and high energy density.

Li + /Na + Storage Mechanism in Co-Based Alloys
For cobalt-based alloys, lithium/sodium ion storage properties are decided by metal components that make up the alloy with cobalt [40][41][42][43]. Even so, the presence of cobalt components in the alloys is essential, which disperses into the alloys to improve the conductivity of alloys, as well as the retard volume expansion during cycles [44,45]. Many types of research on alloying storage mechanisms of cobalt-based materials have been published. For example, Yu et al. briefly pointed out the lithium ion storage mechanisms of Sn-Co@polymethylmethacrylate (PMMA) nanospheres [40]. In Figure 2a, an obvious slope between 1.2 and 0.4 V during the first discharge can be ascribed to the formation of SEI film. Then, a plat plateau at 0.3 V is attributed to the initial lithiation of the Sn component of the alloy, including the isolation of Co and the formation of Li x Sn. Subsequently, the platform centered at about 0.5 V during the charge can be ascribed to the delithiation of Li x Sn and the formation of Sn. The entire lithium storage process of the alloy can be described by following equations: Li x Sn + yCo → Sn + yCo + xLi + + xe − Charge (2) Figure 2. Li + and Na + storage in the Sn-Co alloy anode. (a) Charge-discharge profiles of Sn-Co@polymethyl methacrylate (PMMA) for the initial three cycles in lithium ion batteries (LIBs) [40]. (b) Cyclic voltammetry (CV) curves of SnCo@C in sodium ion batteries (SIBs) for the initial three cycles [41]. (c) and (d) Charge-discharge curves of Sn@C and SnCo@C at different cycles in SIBs, respectively [41].
For cobalt-based alloys, lithium/sodium ion storage properties are decided by metal components that make up the alloy with cobalt [40][41][42][43]. Even so, the presence of cobalt components in the alloys is essential, which disperses into the alloys to improve the conductivity of alloys, as well as the retard volume expansion during cycles [44,45]. Many types of research on alloying storage mechanisms of cobalt-based materials have been published. For example, Yu et al. briefly pointed out the lithium ion storage mechanisms of Sn-Co@polymethylmethacrylate (PMMA) nanospheres [40]. In Figure 2a, an obvious slope between 1.2 and 0.4 V during the first discharge can be ascribed to the formation of SEI film. Then, a plat plateau at 0.3 V is attributed to the initial lithiation of the Sn component of the alloy, including the isolation of Co and the formation of LixSn. Subsequently, the platform centered at about 0.5 V during the charge can be ascribed to the delithiation of LixSn and the formation of Sn. The entire lithium storage process of the alloy can be described by following equations: Co Sn + xLi + xe → Li Sn + yCo(0 < x < 4.4) Discharge (1) Li Sn + yCo → Sn + yCo + xLi + xe Charge (2) As displayed in Figure 2b, the sodium storage mechanism of cobalt alloy can be simply demonstrated by cyclic voltammetry (CV) curves of SnCo@C in SIBs for the initial three cycles [41]. During the first discharge process, peaks centered at around 0.75 and 0.15 V can be ascribed to the first sodiation of the alloy, which indicated the formation of NaxSn and Co. The peaks at 0.3, 0.6 and 0.7 V upon the charge process represented a desodiation process from NaxSn to Sn. Similar to the lithium storage mechanism, the sodium storage mechanism of SnCo@C can be concluded by following equations:  [40]. (b) Cyclic voltammetry (CV) curves of SnCo@C in sodium ion batteries (SIBs) for the initial three cycles [41]. (c) and (d) Charge-discharge curves of Sn@C and SnCo@C at different cycles in SIBs, respectively [41].
As displayed in Figure 2b, the sodium storage mechanism of cobalt alloy can be simply demonstrated by cyclic voltammetry (CV) curves of SnCo@C in SIBs for the initial three cycles [41]. During the first discharge process, peaks centered at around 0.75 and 0.15 V can be ascribed to the first sodiation of the alloy, which indicated the formation of Na x Sn and Co. The peaks at 0.3, 0.6 and 0.7 V upon the charge process represented a desodiation process from Na x Sn to Sn. Similar to the lithium storage mechanism, the sodium storage mechanism of SnCo@C can be concluded by following equations: Na x Sn + Co → Sn + Co + xNa + + xe − Charge (4) In Figure 2c,d, the charge-discharge profiles of Sn@C and SnCo@C displayed similar voltage platforms at different cycles, indicating the decisive role of the metal component alloyed with cobalt. Furthermore, the presence of cobalt improves the capacity retention ratio during cycles, suggesting the importance of cobalt.
2.2. Li + /Na + Storage Mechanism in Co x A y (A=O, S, P and Se) Co x A y electrode materials consist of cobaltous oxides, cobaltous sulfides, cobaltous phosphide and cobaltous selenide. Li + /Na + storage mechanism in Co x A y is the conversion reaction that eventually Appl. Sci. 2020, 10, 3098 5 of 39 forms Co and Li x A or Na x A [14,15,20,27,35,[46][47][48]. The typical reactions of Co x A y with Li + is displayed in Table 1. It should be noted that some temporary intermediate reactions may occur during the discharge process of Co x A y , as depicted in the following equations: or Ge et al. investigated the Na + storage mechanism of CoP by using ex-situ TEM and CV tests [27]. In Figure 3b, two peaks centered at 0.2 and 0.05 V can be attributed to the stepwise formation of Na x P and Na 3 P during the first cathodic scan. During the subsequent charge process, a stepwise extraction of Na + occurred along with the decomposition of Na 3 P, characterized by two anodic peaks located at 0.5 and 1.5 V. It should be noted that a partial Na 3 P transformed to P during the decomposition of Na 3 P. The corresponding sodiation/desodiation reactions are as follows: Fei and coworkers investigated the electrochemical reactions of CoS 2 /rGO by CV curves of the initial two cycles (Figure 3c) [49]. The peak located at 1.4 V can be ascribed to the intercalation of Na + into the CoS 2 crystals in the first discharge process. The resulting intermediate Na x CoS 2 suffered a further conversion reaction to form Na 2 S and metallic Co at 0.8 V. Upon desodiation, two main oxidation peaks at 1.6 and 2.2 V can be due to the reversible formation of Na x CoS 2 and CoS 2 , respectively. The corresponding electrochemical reactions are as follows: or Co A + Li + e → Co A + LiA (6) The sodium storage reaction mechanism of CoSe 2 was investigated by Qiu's group with a CV test, as shown in Figure 3d [48]. During the first anodic scan, two adjacent peaks at 1.7 and 1.87 V were attributed to stepwise conversion reactions from Co and Na 2 Se to an intermediate Na x CoSe and a final CoSe 2 , respectively. In the second anodic scan, three reduction peaks appeared at 1.4, 1.1 and 0.68 V, corresponding to the uptake of Na + and a stepwise sodiation of CoSe 2 . The corresponding reactions can be seen as follows: Discharge process: Charge process: The Li + /Na + storage mechanism is decided by components that composited with cobalt. Similarly, other cobalt-based composites like Co 2 SnO 4 , CoMn 2 O 4 and Sb-Co-P have more complex Li + /Na + storage mechanisms, due to more components in composites [39,[50][51][52]. Nevertheless, the Li + /Na + storage mechanisms are still composed of basic alloying and conversion reactions.

Preparation Methods of Synthesizing Cobalt-based Active Materials
Up until now, many reports have been published for different synthetic strategies of various cobalt-based anode materials. In this section, we review some common methods.

Hydrothermal/Solvothermal Methods
As a typical synthetic method, hydrothermal/solvothermal reactions are mainly used to prepare materials by providing appropriate temperatures and times [53]. Under the circumstances of moderate temperatures and high pressures, various structured cobalt-based materials have been synthesized with high-phase purity and crystallinity. Moreover, hydrothermal/solvothermal methods are environmentally benign and cost-efficient due to their own characteristics [21,54]. Many cobalt-based anode materials with various morphologies have been synthesized by these methods, like nano-sized Co-Sn alloy, single-crystal intermetallic Co-Sn nanospheres and single-crystalline Co 3 O 4 nanobelts [55][56][57].
Zhang's group synthesized a facile structure with CoS 2 confined in the graphitic carbon walls of porous N-doped carbon spheres (CoS 2 -in-wall-NCSs) by the hydrothermal method associated with carbonization and sulfurization, as shown in Figure 4a-c [58]. During the process of hydrothermal, Co-containing melamine-phenolic resin spheres formed under the condition of high temperatures and high pressures. Followed with carbonization and sulfurization, CoS 2 nanodots were evenly distributed throughout the carbon sphere. Figure 4d,e reveals the worm-like sample with CuCo 2 S 4 nanocrystalline (5-20 nm) evenly anchored on the carbon nanotubes (CNTs) synthesized by Jin et al. through solvothermal treatment [59]. CoSe@ carbon spheres (denoted as CoSe@CSs) were prepared through a facile solvothermal treatment by adding Co(NO 3 ) 2 ·6H 2 O and H 3 O 2 Se to a moderate isopropanol solution [60]. The TEM image (Figure 4f) reveals that CoSe nanoparticles uniformly embedded in amorphous carbon spheres (100 nm). Zhu's group synthesized a pure orthorhombic CoP film of nanorod arrays through a hydrothermal method [61]. As displayed in Figure 4g, these CoP nanorods were evenly distributed and interconnected with others to form a 3D interwoven. Li and coworkers reported a facile solvothermal reaction to synthesize Co 3 O 4 /CNT nanocomposites [62]. In Figure 4h, the cyboidal Co 3 O 4 nanoparticles were evenly anchored on the surface of the CNT. The elemental mapping of the sample reveals the distribution of C, Co and O.
isopropanol solution [60]. The TEM image (Figure 4f) reveals that CoSe nanoparticles uniformly embedded in amorphous carbon spheres (100 nm). Zhu's group synthesized a pure orthorhombic CoP film of nanorod arrays through a hydrothermal method [61]. As displayed in Figure 4g, these CoP nanorods were evenly distributed and interconnected with others to form a 3D interwoven. Li and coworkers reported a facile solvothermal reaction to synthesize Co3O4/CNT nanocomposites [62]. In Figure 4h, the cyboidal Co3O4 nanoparticles were evenly anchored on the surface of the CNT. The elemental mapping of the sample reveals the distribution of C, Co and O.

Galvanic Replacement
Galvanic replacement has been widely applied to synthesize nanostructured electrode materials due to its facile and low-temperature synthesis method. Zhou and coworkers synthesized a broccoli-like Co 3 O 4 @CNTs anode material for SIBs through a galvanic replacement route [63]. Firstly, the acidizing operation of carbon nanotubes (CNTs) can create numerous functional groups with a negative charge on the surface of CNTs. Owing to electrostatic attraction, lots of Co 2+ ions with a positive charge will attach to the surface of CNTs with carboxylic groups and serve as nucleation precursors. During the subsequent galvanic replacement reaction, a Co 2+ ion was oxidized to Co 3 O 4 at 140 • C. After that, Co 3 O 4 crystalline particles aggregated on the surface of CNTs owing to the electrostatic attraction and Van der Waals forces; thus, broccoli-like Co 3 O 4 @CNTs were finally obtained (Figure 5a). The TEM and HRTEM images displayed in Figure 5b-d clearly confirmed that the well-crystallized Co 3 O 4 nanocrystals were assembled on the surface of the CNTs. Ma and collogues successfully synthesized a novel Sn-Co@C alloy through galvanic replacement with metal-organic framework ZIF-67 as both the template and carbon source [64]. Figure 5e-g reveal that the Sn-Co nano-alloy particles (~10 nm) were evenly embedded in porous N-doped carbon, forming a novel hierarchical structure. In Figure 5h-l, the distribution of C, O, N, Co and Sn elements clearly confirms the uniform structure of the Sn-Co@C alloy. Galvanic replacement has been widely applied to synthesize nanostructured electrode materials due to its facile and low-temperature synthesis method. Zhou and coworkers synthesized a broccolilike Co3O4@CNTs anode material for SIBs through a galvanic replacement route [63]. Firstly, the acidizing operation of carbon nanotubes (CNTs) can create numerous functional groups with a negative charge on the surface of CNTs. Owing to electrostatic attraction, lots of Co 2+ ions with a positive charge will attach to the surface of CNTs with carboxylic groups and serve as nucleation precursors. During the subsequent galvanic replacement reaction, a Co 2+ ion was oxidized to Co3O4 at 140 °C. After that, Co3O4 crystalline particles aggregated on the surface of CNTs owing to the electrostatic attraction and Van der Waals forces; thus, broccoli-like Co3O4@CNTs were finally obtained ( Figure 5a). The TEM and HRTEM images displayed in Figure 5b-d clearly confirmed that the well-crystallized Co3O4 nanocrystals were assembled on the surface of the CNTs. Ma and Mesoporous Co 3 O 4 nanowire arrays, needlelike Co 3 O 4 nanotubes [65], porous carbon/CoS 2 [66] and some other cobalt-based anode materials with various morphologies and structures have also been synthesized by galvanic replacement [67].
Pan's group successfully distributed CoSe nanoparticles into the porous carbon polyhedral (CoSe@PCP) through a two-step heat treatment [68]. During the first heat treatment, the precursor ZIF-67 was firstly converted to Co@PCP, including the carbonization of the organic linkers and reduction of the Co ion. Subsequently, CoSe@PCP was finally synthesized by selenization of Co@PCP at a high temperature ( Figure 6a). As shown in Figure 6b, the resulting CoSe powders were uniformly embedded into porous carbon without aggregation. Zhang and coworkers prepared a 3D spongy CoS 2 /C anode material through a heating precursor accompanied by a hydrothermal process [12]. Figure 6c delivers a TEM image of the synthesized spongy CoS 2 /C with CoS 2 nanoparticles (~20 nm) homogeneously distributed in the porous carbon. Wang and colleagues realized accurate control over the synthesis of multi-shelled Co 3 O 4 hollow microspheres by adjusting the solution and subsequent heating in the air [69]. The TEM and HRTEM images displayed in Figure 6d,e clearly reveal the morphology of triple-shelled Co 3 O 4 hollow microspheres. Yin's group firstly synthesized a unique core-shell porous FeP@CoP phosphide microcube interconnected by reduced graphene oxide nanosheets (RGO@CoP@FeP) through a facile phosphorization process [70]. In Figure 6f, the SEM image of RGO@CoP@FeP reveals a typical hierarchical structure of wrinkled RGO interconnected microcubes. Qiu's group synthesized micro-scaled spherical CoSn 2 /Sn alloy through a sintering process, employing a stoichiometric amount of SnO 2 , Co 3 O 4 and C as raw materials [22]. SEM observation of CoSn 2 /Sn alloy reveals that the particles were composed of lots of small grains, as shown in the inset of Figure 6g. Zeng et al. developed a facile heat treatment, composed of pyrolysis and sulfurization, to synthesize hollow Co 9 S 8 /N-C composites [28]. The SEM image of hollow Co 9 S 8 /N-C composites displayed in Figure 6h reveals the thickness of the wall was approximately 70 nm. Co/(NiCo)Se 2 box-in-box structures were firstly fabricated by Kang's group via a selenization process at Ar/H 2 , applying well-known ZIF-67 as a template [71]. As shown in Figure 6i, the thickness of the outer Ni-Co selenide shell is about 60 nm, while the inner CoSe 2 shell shows a very uneven thickness.

High-Energy Mechanical Milling
As a facile preparation technology, high-energy mechanical milling (denoted as HEMM) has also been commonly applied to the fabrication of cobalt-based materials owing to its large-scale production [34,42,45,[80][81][82][83]. For HEMM, the products are synthesized under high temperatures and pressures generated by highly energetic collisions of tiny balls.
Carbon-coated CoP 3 (CoP 3 @C) nanocomposites were synthesized by HEMM methods with a certain proportion of phosphorus, cobalt powders and carbon black [80]. The synthesis of the schematic illustration is displayed in Figure 7a. HRTEM images shown in Figure 7b reveal the presence of CoP 3 nanoparticles and a carbon matrix. B. Scrosati's group successfully transformed a certain proportion of Sn, Co and graphite into a nanosized SnCoC-2 alloy through HEMM [34]. Figure 7c shows the SEM image of the obtained SnCoC-2 alloy. Li et al. prepared Sn 30 Co 30 C 40 electrode materials by mechanical attrition with CoSn 2 , Co and graphite as the precursors [81]. SEM image of the obtained Sn 30 [84]. (i) SEM image of Co3O4 thin films synthesized by electrochemical deposition [85]. (j) SEM image of the Sn-Co alloy deposited on a Ni nanocon-array [38].
As a facile preparation technology, high-energy mechanical milling (denoted as HEMM) has also been commonly applied to the fabrication of cobalt-based materials owing to its large-scale production [34,42,45,80,81,86,87]. For HEMM, the products are synthesized under high temperatures and pressures generated by highly energetic collisions of tiny balls.
Carbon-coated CoP3 (CoP3@C) nanocomposites were synthesized by HEMM methods with a certain proportion of phosphorus, cobalt powders and carbon black [80]. The synthesis of the schematic illustration is displayed in Figure 7a. HRTEM images shown in Figure 7b reveal the presence of CoP3 nanoparticles and a carbon matrix. B. Scrosati's group successfully transformed a certain proportion of Sn, Co and graphite into a nanosized SnCoC-2 alloy through HEMM [34]. Figure  [81]. (e) SEM image of the Co-Sn alloy after electrodeposition [84]. (f) Schematic illustration of the preparation of the Sn-Sb-Co alloy. (g) SEM images of the SnSbCo-3 electrode; the insets show a corresponding high-magnification image (top-right) and cross-sectional photos of the electrode (bottom-right) [85]. (h) SEM image of the porous SnCo alloy [86]. (i) SEM image of Co 3 O 4 thin films synthesized by electrochemical deposition [87]. (j) SEM image of the Sn-Co alloy deposited on a Ni nanocon-array [38].

Electrodeposition
Electrodeposition, as a method for separating metals or alloys from their compounds, has been used in the preparation of cobalt-based electrode materials [18,38,43,[84][85][86][87][88][89]. Ricardo's group prepared the Co-Sn alloy by depositing Sn and Co atoms through the co-electroplating method, according to moderate time and current density [84]. Figure 7e indicates that the deposited Co-Sn alloy covered the nickel substrate with certain roughness and protuberances. Sun's group reported a novel nanoarchitectured Sn-Sb-Co alloy through direct electrodeposition on a Cu nanoribbon array [85].
A schematic diagram of the preparation of 3D nanoarchitectured Sn-Sb-Co alloy is shown in Figure 7f. SEM images (Figure 7g) clearly reveal the 3D array nanostructure and thickness (1500 nm) of the Sn-Sb-Co alloy.
A porous Sn-Co alloy was fabricated by Sun's group via electrodepositing the Sn-Co ally on the porous Cu film prepared by electroless plating [86]. SEM image (Figure 7h) indicates that the Sn-Co alloy was uniformly covered on the porous Cu film. Porous Co 3 O 4 nanostructured thin films were synthesized via a facile electrodepositing method (Figure 7i) [87]. SEM image reveals that the sample was composed of small nanoflakes of about 30-40 nm in thickness. Du et al. successfully a synthesized nanoarchitectured Sn-Co alloy electrode through a two-step electrodeposition [38]. The SEM image in Figure 7j evidently indicates that there is sufficient room between the Sn-Co cylinders
Electrospinning is a method to synthesize 1D nanomaterials by the breakdown of a high-polymer solution by a high-voltage electrostatic field [24,93,97]. Li's group distributed CoSe 2 nanoparticles uniformly into nitrogen-doped carbon nanofibers through two steps, including the electrospinning approach and subsequent selenization process (Figure 8a) [24]. TEM image, HRTEM image and corresponding elements mapping clearly reveal that CoSe 2 particles were well-wrapped in the N-doped nanofibers without aggregation, as shown in Figure 8b-d, respectively. Co-Sn/CNF composites were fabricated by Lee's group through electrospinning, followed with heat treatments [90]. SEM image of the Co-Sn/CNF surface shows that the average diameter of the fibers was 180 nm (Figure 8e). Wang et al. prepared carbon-encapsulated wire-in-tube Co 3 O 4 /MnO 2 heterostructure nanofibers through electrospinning, followed by calcination [91]. The TEM image in Figure 8f shows a tubular structure with three layers, including an MnO 2 wire, Co 3 O 4 tube and carbon layer, subsequently. X-ray EDS in Figure 8g clearly indicates the presence of Mn, Co, O, N and C elements. Kim's group prepared 1D carbon nanofibers embedded with uniformly SnCo nanoparticles via electrospinning, followed by calcination [92]. SEM image in Figure 8h shows that SnCo alloys with size distributions of 5-10 nm were homogeneously encapsulated in carbon nanofibers. Niu's group successfully synthesized a novel urchin-like sample composed of CoSe 2 nanofibers rooted into the carbon nanofibers (CoSe 2 @CNFs) [93]. In the preparation process, Niu's group firstly synthesized the precursor (Co@CNF) via electrospinning, followed by calcination, and subsequently prepared CoSe 2 @CNFs through a hydrothermal selenation. TEM image and corresponding HRTEM image (inset) in Figure 8i obviously show that urchin-like CoSe 2 nanorods with lengths of 50-100 nm and diameters of 20-30 nm are rooted into the electrospun CNFs. Qiu et al. employed the electrospinning method to synthesize Co/Co 3 O 4 -carbon nanofibers with porous structures [94]. The Co/Co 3 O 4 nanoparticles were distributed throughout the carbon nanofiber, as shown in Figure 8j.
synthesized via a facile electrodepositing method (Figure 7i) [85]. SEM image reveals that the sample was composed of small nanoflakes of about 30-40 nm in thickness. Du et al. successfully a synthesized nanoarchitectured Sn-Co alloy electrode through a two-step electrodeposition [38]. The SEM image in Figure 7j evidently indicates that there is sufficient room between the Sn-Co cylinders. Chemical vapor deposition (CVD) processes have also been applied to synthesize cobalt-based electrode materials with multifarious structures [44,[98][99][100]. He's group successfully prepared a novel structure via the CVD process [44]. The TEM image in Figure 8k clearly indicates a typical 0D-1D-3D hierarchical structure.

Other Methods
The magnetron sputtering method is also developed to prepare cobalt-based materials on substrates [19,101]. Chen et al. synthesized a Co 3 O 4 -C core-shell nanowire array via hydrothermal and magnetron sputtering. SEM image in Figure 8l shows that the Co 3 O 4 nanowires were well-coated by carbon layers with a thickness of 18 nm.
Furthermore, Yoon's group reported an electric arc-melting method to produce a Co 2 MnSi ingot [96]. Figure 8m indicates that the synthesized materials exhibited a certain degree of aggregation.
Various strategies for the preparation of cobalt-based electrode materials are summarized in Table 2. Obviously, hydrothermal/solvothermal, heat treatment, galvanic replacement and HEMM can be used to prepare different types of cobalt-based materials due to their economy and simplicity naturally. Other methods like electrospinning, electrodeposition and magnetron sputtering are more likely to synthesize cobalt-based materials with specific morphologies and structures, such as nanofibers and films. The diversity of the preparation process ensures the variety of synthesized cobalt-based electrode materials, thereby increasing the opportunities of cobalt-based materials for real batteries.

Cobalt-Based Alloys and Its Composites
As mentioned above, the electrochemical properties of cobalt-based alloys are mostly decided by metal components alloyed with cobalt, owing to the inactive nature with Li/Na of the cobalt ion.
As an inactive component of cobalt-based alloys, the presence of phase cobalt can alleviate the volume fluctuation as a protective matrix, thereby improving the electrochemical performances of cobalt-based alloys. Despite these features, the electrochemical properties of cobalt-based alloys are still disturbed by the pulverization and shedding of active materials during periodical cycles. Many cobalt-based alloys have been reported as anode materials for LIBs or SIBs with outstanding electrochemical properties owing to their unique structures and suitable compositions [40,55,67,81,83,92,104,105,171].
Shi et al. reported a novel structure with a Sn-Co alloy embedded in porous N-doped carbon microboxes that exhibited a superior electrochemical performance when used as anode materials for LIBs [64]. In Figure 9a, the Sn-Co@C-2 electrode exhibited the highest specific capacity at various current densities, from 0.1 to 2 A g −1 , indicating a superior rate performance. Furthermore, the Sn-Co@C-2 electrode delivered outstanding capacity retention when the current density went back to 0.1 A g −1 . Ex-situ XRD of the Sn-Co@C-2 electrode before and after cycles are shown in Figure 9b. CoSn2 disappeared after the first cycle, and the CoSn phase maintained well even after 100 cycles, suggesting a good reversibility of the Sn-Co@C-2 electrode.  Shi et al. reported a novel structure with a Sn-Co alloy embedded in porous N-doped carbon microboxes that exhibited a superior electrochemical performance when used as anode materials for LIBs [64]. In Figure 9a, the Sn-Co@C-2 electrode exhibited the highest specific capacity at various current densities, from 0.1 to 2 A g −1 , indicating a superior rate performance. Furthermore, the Sn-Co@C-2 electrode delivered outstanding capacity retention when the current density went back to He's group reported a novel method for the one-step synthesis of hierarchical SnCo@CNT-3DC [44]. When used as anode materials for LIBs, this unique structure can provide numerous channels for electron transfer, shorten the diffusion pathway of Li + and ensure the permeation of electrolytes. As shown in Figure 9c, the SnCo@CNT-3DC electrode displayed a high capacity of 826 mA h g −1 after 100 cycles at 0.1 A g −1 . Compared with the first discharge capacity, the capacity retention rate is 113%. During the cycles, there is a slight increase in capacity owing to the increasing electrochemical reaction of the SnCo nanoparticles caused by the activation of the electrode materials. As for the SnCo/C electrode, the relatively low capacity of 429.6 mA h g −1 was retained after 50 cycles at 0.1 A g −1 .
These results can be probably ascribed to the better transformation of lithium in the SnCo@CNT-3DC electrode. Figure 9d shows that the SnCo@CNT-3DC electrode delivered a high discharge capacity of 1154.4 mA h g −1 , with an initial coulombic efficiency (CE) of 61%. As the number of cycle increases, the capacity decreases slightly, suggesting good capacity retention.
Wang's group demonstrated an in-situ template technique for the preparation of the Sn-Co-CNT@CNT ternary tube-in-tube nanostructure [67]. When utilized as anode materials for LIBs, superior electrochemical properties were obtained owing to the confined volume change in the nanotube cavities and ensured a permanent electrical connectivity of the immobilized Sn-Co anodes. As shown in Figure 9e, the Sn-Co-CNT@CNT electrode with 45.4% Sn delivered a discharge capacity of 811 mA h g −1 after 200 cycles, with 91.1% of the initial discharge capacity retained. The TEM image of the Sn-Co-CNT@CNT anode after 200 cycles demonstrates that the 1D nanotube-like structure was retained, indicating its structural stability during the cycles (Figure 9f).

Cobalt Oxides and Its Composites
Cobalt oxides have been identified as hopeful anode materials for LIBs and SIBs due to their high theoretical capacities, natural abundances, and low costs. For instance, Co 3 O 4 , as the most common cobalt oxide, has a theoretical capacity of 890 mA h g −1 attributed to its eight-electron transfer reaction during cycling. However, the intrinsic defects of cobalt oxides, including low structural stability and poor electrical conductivity, lead to inferior electrochemical properties [57,63,77,78,111,113,116,118,120]. Yu's group reported hierarchically structured Co 3 O 4 @ carbon porous fibers to circumvent the volume fluctuation issue during lithiation/delithiation [11]. As shown in Figure 10a, the initial discharge and charge specific capacity of ES-CNCo 3 O 4 was 1824 and 1003 mA h g −1 , with an initial columbic efficiency of 55%. After five cycles, the charge/discharge profiles were almost overlapped, indicating the high stability of the composite. nanotube cavities and ensured a permanent electrical connectivity of the immobilized Sn-Co anodes. As shown in Figure 9e, the Sn-Co-CNT@CNT electrode with 45.4% Sn delivered a discharge capacity of 811 mA h g −1 after 200 cycles, with 91.1% of the initial discharge capacity retained. The TEM image of the Sn-Co-CNT@CNT anode after 200 cycles demonstrates that the 1D nanotube-like structure was retained, indicating its structural stability during the cycles (Figure 9f).  Wu et al. synthesized yolk-shell Co 3 O 4 /C dodecahedrons to solve the problem of volume change during lithiation or sodiation [129]. When utilized as an anode for LIBs, the yolk-shell Co 3 O 4 /C electrode delivered a high discharge capacity of 1100 mA h g −1 after 120 cycles. Regarding Co 3 O 4 -B, a lower specific capacity of 400 mA h g −1 after 50 cycles was obtained, suggesting a better lithium storage performance of yolk-shell Co 3 O 4 /C. Figure 10c shows the sodium storage performance of yolk-shell Co 3 O 4 /C at a current density of 1 A g −1 . The second discharge capacity of the sample was 395 mA h g −1 , and the specific capacity of 240 mA h g −1 was obtained after 200 cycles, suggesting the great electrochemical properties of yolk-shell Co 3 O 4 /C in SIBs.

Cobalt Oxides and its Composites
Xie's group reported a porous starfish-like nanocomposite of N-doped carbon with Co 3 O 4 coating [128]. The Co 3 O 4 @N-C electrode as a LIB anode delivered an initial specific capacity of 1235 mA h g −1 and a capacity of 795 mA h g −1 after 300 cycles, indicating a superior lithium storage performance to pure Co 3 O 4 (Figure 10d). To better understand the gap between the electrochemical performances of the two samples, Nyquist plots of Co 3 O 4 @N-C and pure Co 3 O 4 electrodes were fitted (Figure 10e). It is clear that the Co 3 O 4 @N-C composite exhibited smaller charge-transfer resistance, suggesting that the porous structure accelerates the diffusion of Li + , and the presence of N-doped carbon improves the electrical conductivity of the sample, finally resulting in the superior electrochemical properties of Co 3 O 4 @N-C. Yang's group reported the preparation of a Ni-doped Co/CoO/N-doped carbon hybrid, with Ni-Co-ZIF as the precursor [126]. When utilized as an anode material for SIBs, the Ni-doped Co/CoO/NC electrode exhibited a high discharge capacity of 218.7 mA h g −1 after 100 cycles, with 87.5% capacity retained (Figure 10f). Wang and his collogues reported a nanocomposite of N-doped carbon with Co 3 O 4 nanoparticles, exhibiting a superior sodium storage performance [124]. To further investigate the electrochemical mechanism of the electrode, five points that appeared at the first charge-discharge profile were tested by ex-situ XRD patterns (Figure 10g,h). Upon the discharge process, the phase of Co 3 O 4 gradually disappeared along with the sodiation. When the cell discharged to 0.6 V (B), a new phase of CoO came out owing to the reduction of Co 3 O 4 . When fully discharged (C), only one phase of the Cu collector was retained, attributed to the poor crystallization of the Na 2 O phase. During the following charge process, reverse reactions occurred accompanied by the reappearance of CoO and Co 3 O 4 at 1.4 V and only Co 3 O 4 when fully charged (E). These results can be attributed to a reversible reaction of Co 3 O 4 + 8Na + + 8e − ↔ 3Co + 4Na 2 O during cycling.

Cobalt Sulfide and Its Composites
Cobalt sulfide has been widely studied as an anode material for LIBs/SIBs owing to its high theoretical capacity, satisfactory electrical conductivities and fair thermal stabilities. Similar to cobalt oxides, the main obstacle to the development of cobalt sulfide is the severe volume fluctuation during periodical cycling [14,16,17,26,36,47,49,66,[172][173][174][175][176][177][178][179][180][181][182][183][184][185][186]. He and his colleagues synthesized self-assembled CoS 2 nanoparticles wrapped by CoS 2 -quantum-dots-anchored graphene nanosheets (denoted as CoS 2 NP@G-CoS 2 QD) and tested as an anode material for LIBs [187]. As displayed in Figure 11a, the CoS 2 NP@G-CoS 2 QD delivered a high specific capacity of 831 mA h g −1 at 1 A g −1 after 300 cycles. Meanwhile, the nearly 100% coulombic efficiency over all 300 cycles also indicates the excellent cycle performance of this material. The enhanced cycle performance can be ascribed to the presence of few-layer graphene, which can not only improve the electrical conductivity of this composite but also alleviate the volume fluctuation and aggregation of CoS 2 (Figure 11b).
Sun's group synthesized a sponge-like composite assembled with Co 9 S 8 quantum dots embedded in a carbon matrix, which was wrapped by rGO [172]. When utilized as an anode material for SIBs, the reaction mechanism was studied by an ex-situ XRD pattern in Figure 11c. Upon sodiation, the peaks of Co 9 S 8 gradually disappeared until the discharge was completed. At the same time, the peaks of the Na 2 S and Co phases increasingly appeared owing to the reduction reaction from Co 9 S 8 to Co and Na 2 S. During the subsequent charge process, an inverse reaction, including the disappearance of the Na 2 S and Co phases and the reappearance of Co 9 S 8 , occurred. This situation indicates the high reversibility of Co 9 S 8 quantum dots during the initial charge-discharge process. Peng's group built a CoS@rGO composite assembled by CoS nanoplates decorated on rGO and tested as an anode material for SIBs [173]. To investigate the industry application of this composite, a full cell was fabricated with Na 3 VPO 4 @C nanowires as the cathode (Figure 11d). As shown in Figure 11e, the reversible capacity of 290 mA h g −1 was obtained after 100 cycles, suggesting the superior cycle performance of a CoS@rGO// Na 3 VPO 4 (NVP)@C full battery. The alight red LED (inset of Figure 11e) fully demonstrates the practical application of a CoS@rGO//NVP@C full battery. This excellent electrochemical performance resulted from the novel structure and the presence of rGO.
Lou's group reported a CoS 2 nanobubble hollow prism and tested it as an anode material for LIBs [174]. This sample delivered remarkable specific capacities of 910, 778, 681 and 470 mA h g −1 at the current densities of 200, 500, 1000 and 5000 mA g −1 , respectively (Figure 11f). It is worth mentioning that the capacity recovered to 864 mA h g −1 when the current density was returned to 200 mA g −1 , indicating the outstanding robustness of the electrode. Qian's group synthesized hollow nanospheres assembled of mesoporous Co 9 S 8 [175]. When utilized as an anode material of LIBs, an excellent specific capacity was maintained as 896 mA h g −1 even after 800 cycles at 2 A g −1 (Figure 11g). The superior cycle performance resulted from the remarkable charge-transfer kinetics. Furthermore, Zhou and coworkers prepared a one-dimensional multiwalled carbon nanotube@a-C@Co 9 S 8 nanocomposites (MWCNT@a-C@Co 9 S 8 ) as an anode material for advanced LIBs [5]. In Figure 11h, a high reversible capacity of 1065 mA h g −1 was obtained after 700 cycles at a current density of 2 A g −1 . It is worth mentioning that a slight fluctuation was observed during the cycling, ascribed to the enhanced surface lithium storage, the electrode reactivation and so on.
Cobalt sulfide has been widely studied as an anode material for LIBs/SIBs owing to its high theoretical capacity, satisfactory electrical conductivities and fair thermal stabilities. Similar to cobalt oxides, the main obstacle to the development of cobalt sulfide is the severe volume fluctuation during periodical cycling [14,16,17,26,36,47,49,66,[173][174][175][176][177][178][179][180][181][182][183][184][185][186][187]. He and his colleagues synthesized self-assembled CoS2 nanoparticles wrapped by CoS2-quantum-dots-anchored graphene nanosheets (denoted as CoS2 NP@G-CoS2 QD) and tested as an anode material for LIBs [172]. As displayed in Figure 11a, the CoS2 NP@G-CoS2 QD delivered a high specific capacity of 831 mA h g −1 at 1 A g −1 after 300 cycles. Meanwhile, the nearly 100% coulombic efficiency over all 300 cycles also indicates the excellent cycle performance of this material. The enhanced cycle performance can be ascribed to the presence of fewlayer graphene, which can not only improve the electrical conductivity of this composite but also alleviate the volume fluctuation and aggregation of CoS2 (Figure 11b).
Sun's group synthesized a sponge-like composite assembled with Co9S8 quantum dots embedded in a carbon matrix, which was wrapped by rGO [173]. When utilized as an anode material for SIBs, the reaction mechanism was studied by an ex-situ XRD pattern in Figure 11c. Upon sodiation, the peaks of Co9S8 gradually disappeared until the discharge was completed. At the same time, the peaks of the Na2S and Co phases increasingly appeared owing to the reduction reaction from Co9S8 to Co and Na2S. During the subsequent charge process, an inverse reaction, including the disappearance of the Na2S and Co phases and the reappearance of Co9S8, occurred. This situation indicates the high reversibility of Co9S8 quantum dots during the initial charge-discharge process. Peng's group built a CoS@rGO composite assembled by CoS nanoplates decorated on rGO and tested as an anode material for SIBs [174]. To investigate the industry application of this composite, a full  [187]. (c) Ex-situ XRD of the cobalt sulfides quantum dots@mesoporous hollow carbon polyhedral@reduced graphene oxide (Co 9 S 8 QD@HCP)@rGO electrode tested at different voltage states [172]. (d) Schematic illustration and (e) cycle performance of CoS@rGO//Na 3 VPO 4 (NVP)@C sodium ion full cell [173]. (f) Rate performance of CoS 2 bubble-like hollow prisms [174]. (g) Cycle performance of Co 9 S 8 -650@C in LIBs [175]. (h) Cycle performance of Multiwalled carbon nanotube@a-C@Co 9 S 8 nanocomposites (MWCNT@a-C@Co 9 S 8 ) anode for LIBs [5].

Cobalt Phosphide and Its Composites
Cobalt phosphides, as promising electrode materials, have been studied as anode materials for LIBs/SIBs. Compared with cobalt oxides and sulfides, cobalt phosphides possess higher gravimetric/volumetric specific capacities. The main drawbacks that restrict their applications are the severe volume expansion during lithiation or sodiation and poor electrical conductivities. Many efforts have been made to circumvent these issues [18,31,46,82,134,136,137,188,189]. Sun's group hybridized the Co 2 P nanoparticles with N-doped carbon matrices and tested the composite as an anode material for SIBs [143]. Except for the initial discharge profile, the profiles of the galvanostatic charge-discharge process overlapped well with each other, indicating a good reversibility of Co 2 P@N-C@rGO (Figure 12a). Yan's group successfully fabricated Co x P nanostructures with controlled phases, sizes and shapes. When applied as the anode material of LIBs, the hollow CoP particles delivered a capacity of 630 mA h g −1 after 100 cycles, as shown in Figure 12b. With regard to solid CoP nanoparticles, a capacity of 480 mA h g −1 was retained after 100 cycles. Better cycling performance of the hollow CoP electrode can be ascribed to the hollow structure, which ensures effective lithiation/delithiation and alleviates the volume change during periodical cycling.

Cobalt Phosphide and its Composites
Xiong's group firstly synthesized a novel composite with core-like CoP embedded in N, P co-doped porous carbon sheets (CoP⊂NPPCS) [141]. When utilized as an anode material of LIBs, the CoP⊂NPPCS electrode delivered reversible capacities of about 709, 500, 501, 410 and 357 mA h g −1 at the current densities of 0.1, 0.2, 0.5 and 1.0 A g −1 , respectively. It is worth noting that the capacity retention was over 50% when the current density was expanded by 20 times, indicating the excellent rate capability. Peng's group prepared Co 2 P-Co (Co-P) hollow nanospheres decorated with graphene sheets and tested them as anode materials for LIBs [139]. When the current densities increased from 0.2 C, 0.4 C, 0.6 C, 0.8 C and 1.0 C to 2.0 C (1C = 540 mA g −1 ), the capacity of Co-P/graphene correspondingly decreased from 855, 674, 628, 608 and 610 to 567 mA h g −1 . When the current density returned to 0.2 C, a high capacity of 1039 mA h g −1 was retained even after 100 cycles (Figure 12d). Clearly, the rate performance of Co-P/graphene was better than their Co 2 P-Co counterparts, attributed to the presence of graphene, which improved the electrical conductivity of the electrode (Figure 12e).
Yang and co-workers synthesized a composite with cobalt phosphide nanowires and reduced the graphene oxide (CoP/RGO) [135]. As shown in Figure 12f, the cycle performances of CoP and CoP/RGO as anode materials of LIBs were displayed. The CoP/RGO electrode exhibited a higher capacity of 967 mA h g −1 after 200 cycles. With regard to the CoP electrode, a capacity of only 429 mA h g −1 was retained after 100 cycles. This situation can be ascribed to the framework of CoP/RGO, which possesses more free space in favor of the volume change during cycling. Zhang's group synthesized a carbon-coated CoP 3 nanocomposite and tested it as an anode material for SIBs [80]. The cycling performances of CoP 3 @C and CoP 3 at 100 mA g −1 are displayed in Figure 12g. The CoP 3 @C electrode exhibited a higher capacity than CoP 3 after 80 cycles, suggesting that the presence of the carbon layer contributed to the electrochemical performance of the electrode. Furthermore, Zhu's group successfully embedded ultrafine CoP nanoparticles into the carbon nanorod, which delivered a superior cycling performance when tested as an anode material of LIBs at 1 A g −1 (Figure 12h) [142]. Well-utilized CoP was the main reason for this result.

Cobalt Selenide and Its Composites
Cobalt selenide, one of the transition metal chalcogenides, has attracted enormous attention in the field of anode materials for LIBs or SIBs owing to its good electrical conductivity and high theoretical capacity. Nevertheless, its poor structure stability during cycling is the main drawback of cobalt selenide [20,37,71,[168][169][170][190][191][192][193][194][195]. As an anode material of LIBs, the Co 0.85 Se NSs/G composite synthesized by Zhou's group delivered a high specific capacity of 730 mA h g −1 after 300 cycles, as shown in Figure 13a [160]. With regard to Co 0.85 Se NSs and Co 0.85 Se NSs without a binder, the specific capabilities rapidly decreased to 50 mA h g −1 after 50 cycles at 0.5 A g −1 . It should be noticed that the capacity of Co 0.85 Se NSs/G has a slight increase during cycling, ascribed to the presence of graphene and the formation of a polymetric gel-like film. Moreover, the Na-ion storage cyclic performances of Co 0.85 Se NSs/G, Co 0.85 Se NSs and Co 0.85 Se NSs without additives are displayed in Figure 13b [160]. Similar to lithium storage performances, the Co 0.85 Se NSs/G electrode exhibited a higher reversible capacity of 193.8 mA h g −1 after 100 cycles. The better lithium/sodium storage performance of the Co 0.85 Se NSs/G composite can be attributed to its unique morphology and composition.
Pan's group successfully fabricated a composite with CoSe nanoparticles uniformly distributed in porous carbon polyhedral (CoSe@PCP) and tested as an anode material for LIBs and SIBs [68]. In Figure 13c, the CoSe@PCP electrode exhibited reversible capabilities of 701.2, 645.6, 590 and 457.5 mA h g −1 when tested at 0.1, 0.5, 1 and 2 A g −1 , respectively. Moreover, the capacity could recover to 441.8, 452.7 and 524.1 mA h g −1 when the current density was back to 1, 0.5 and 0.1 A g −1 , respectively. When utilized as an anode material for SIBs, the electrode delivered the capacities of 360.3, 315.6, 278.9, 247.1 and 207.7 mA h g −1 tested at 0.05, 0.25, 1, 2 and 4 A g −1 , respectively. Furthermore, the capacities could recover to the original levels when the current densities returned to 0.5, 0.25 and 0.05 A g −1 (Figure 13d). Clearly, the CoSe@PCP electrode shows excellent rate performances when applied as anode materials for LIBs and SIBs, attributed to the novel structure of CoSe@PCP.
Chen's group synthesized a novel urchin-like CoSe 2 and firstly used it as an anode material for SIBs [190]. As shown in Figure 13e, the CoSe 2 electrode delivered a high capacity of 0.410 A h g −1 after 1800 cycles at 1 A g −1 , suggesting a superior cyclic performance. This result can be ascribed to the unique structure of the sample. Na-ion storage cyclic performance of the Co0.85Se nanosheets/graphene (NSs/G) at 0.5 A g −1 , in comparison with Co0.85Se NSs and Co0.85Se NSs without additives [160]. Rate performance of CoSe@PCP for (c) LIBs and (d) SIBs [68]. (e) Cyclic properties of CoSe2 at 1 A g −1 [190].
Cobalt selenide, one of the transition metal chalcogenides, has attracted enormous attention in the field of anode materials for LIBs or SIBs owing to its good electrical conductivity and high theoretical capacity. Nevertheless, its poor structure stability during cycling is the main drawback of cobalt selenide [20,37,71,[168][169][170][190][191][192][193][194][195]. As an anode material of LIBs, the Co0.85Se NSs/G composite synthesized by Zhou's group delivered a high specific capacity of 730 mA h g −1 after 300 cycles, as shown in Figure 13a [160]. With regard to Co0.85Se NSs and Co0.85Se NSs without a binder, the specific capabilities rapidly decreased to 50 mA h g −1 after 50 cycles at 0.5 A g −1 . It should be noticed that the capacity of Co0.85Se NSs/G has a slight increase during cycling, ascribed to the presence of graphene and the formation of a polymetric gel-like film. Moreover, the Na-ion storage  [160]. Rate performance of CoSe@PCP for (c) LIBs and (d) SIBs [68]. (e) Cyclic properties of CoSe 2 at 1 A g −1 [190].

Other Cobalt-Based Anode Materials
In addition to cobalt alloys, cobalt oxides, cobalt phosphide and cobalt chalcogenides, other cobalt-based anode materials like CoMn 2 O 4 have also been utilized in LIBs and SIBs [39,[50][51][52]72,96,108,109]. Zhang's group synthesized CoMn 2 O 4 microspheres and evaluated them as anode materials for LIBs [51]. As displayed in Figure 14a, the CoMn 2 O 4 electrode exhibited high initial charge and discharge capacities of 1860 and 2296 mA h g −1 at 100 mA g −1 , respectively. During the first 200 cycles, the capacity gradually declined owing to a severe structure degradation caused by the conversion reaction.
Chen's group synthesized a novel urchin-like CoSe2 and firstly used it as an anode material for SIBs [190]. As shown in Figure 13e, the CoSe2 electrode delivered a high capacity of 0.410 A h g −1 after 1800 cycles at 1 A g −1 , suggesting a superior cyclic performance. This result can be ascribed to the unique structure of the sample.  After 250 cycles, the reversible capacities increased, which can be ascribed to the gradual formation of the SEI and the degradation of the electrolyte. After 500 cycles, the CoMn 2 O 4 electrode delivered a capacity of 722 mA h g −1 , indicating the enhanced performance of the CoMn 2 O 4 electrode as an anode for LIBs. Wang et al. prepared CoSnO 3 /GN/CNTs composite papers that delivered excellent electrochemical properties [109]. The rate performances of the CoSnO 3 /GN/CNTs composite with different structures are displayed in Figure 14b. The CoSnO 3 nanobox/GN/CNTs composite delivered the capacities of 1233.8, 1106.4, 894.7, 780.3 and 676.7 mA h g −1 at the current densities of 0.1, 0.25, 0.5, 1 and 2 A g −1 , respectively. Moreover, a high reversible capability of 992.8 mA h g −1 could be retained when the current density returned to 0.1 A g −1 . The rate properties of CoSnO 3 nanocube/GN/CNTs were poorer than CoSnO 3 nanobox/GN/CNTs. Ji's group synthesized CoSnO 3 -NCs nanoboxes and employed them as anode materials for SIBs [52]. The cyclic performance of this sample was tested at a current density of 1 A g −1 , as shown in Figure 14c. The CoSnO 3 -NCs electrode exhibited a high capacity of 273.8 mA h g −1 after 1000 cycles owing to the stable structure derived from the presence of the carbon layer.

Other Cobalt-Based Anode Materials
The electrochemical properties of typical cobalt-based anode materials for LIBs and SIBs are summarized in Table 3. Clearly, this cobalt-based active material delivered different levels of rate performances and cyclic stabilities owing to the differences between morphology and composition. Hence, a rational structure is vitally important for the electrochemical properties of cobalt-based anode materials. Furthermore, the combination with carbonaceous materials is also an effective method to enhance the electrochemical properties of cobalt-based anode materials for LIBs and SIBs.

Conclusions and Perspectives
The increasing consumption of fossil energy makes the development of clean energy necessary. Due to time and geographical constraints, high-performance RBs are urgent for the development of clean energy. Designing and synthesizing suitable electrode materials is critical for the development of RBs. This review systematically demonstrates that cobalt-based electrode materials are expected to be a new generation of anode materials for LIBs/SIBs.
In this review, we sequentially summarized the Li + /Na + storage mechanism, typical preparation methods and applications of various cobalt-based anode materials, including cobalt-based alloy, cobalt oxide, cobalt sulfides and cobalt phosphides. For cobalt-based anode materials, their large volume changes and poor conductivity during charges and discharges are the main obstacles to the further development in LIBs/SIBs. In this regard, future research can focus on the following aspects: (1) Developing a simple and feasible synthetic process for the preparation of cobalt-based anode materials with specific morphology and sizes, which can ensure satisfactory contact between the electrolyte and active material, and cycle performances of cobalt-based anode materials. In addition, the combination with the conductive material can improve the conductivity of the electrode material and improve the rate performances of the cobalt-based anode materials. (2) More advanced characterization and calculation methods should be used to further study the Li + /Na + storage mechanisms of cobalt-based anode materials, which will be meaningful for designing a suitable morphology. (3) Considering commercial applications, the whole battery system, including cathode materials, binders, conductive agents, electrolytes and additives, should also be optimized. Appropriate matching materials can fully exploit the advantages of the high storage capacities of cobalt-based anode materials in LIBs/SIBs.