Integrated Design of Hierarchical CoSnO3@NC@MnO@NC Nanobox as Anode Material for Enhanced Lithium Storage Performance.

Transition-metal oxides (TMOs) are potential candidates for anode materials of lithium-ion batteries (LIBs) due to their high theoretical capacity (∼1000 mA h/g) and enhanced safety from suppressing the formation of lithium dendrites. However, the poor electron conductivity and the large volume expansion during lithiation/delithiation processes are still the main hurdles for the practical usage of TMOs as anode materials. In this work, the CoSnO3@NC@MnO@NC hierarchical nanobox (CNMN) is then proposed and fabricated to solve those issues. The as-prepared nanobox contains hollow cubic CoSnO3 as a core and dual N-doped carbon-"sandwiched" MnO particles as a shell. As anode materials of LIBs, the hollow and carbon interlayer structures effectively accommodate the volume expansion while dual active TMOs of CoSnO3 and MnO efficiently increase the specific capacity. Notably, the dual-layer structure of N-doped carbons plays a critical functional role in the incorporated composites, where the inner layer serves as a reaction substrate and a spatial barrier and the outer layer offers electron conductivity, enabling more effective involvement of active anode materials in lithium storage, as well as maintaining their high activity during lithium cycling. Subsequently, the as-prepared CNMN exhibits a high specific capacity of 1195 mA h/g after the 200th cycle at 0.1C and an excellent stable reversible capacity of about 876 mA h/g after the 300th cycle at 0.5C with only 0.07 mA h/g fade per cycle after 300 cycles. Even after a 250 times fast charging/discharging cycle both at 5C, it still retains a reversible capacity of 422.6 mA h/g. We ascribe the enhanced lithium storage performances to the novel hierarchical architectures achieved from the rational design.

Introduction

With the continuous development of human society, the demand for clean and sustainable energy is proposed and constantly serves as a key to the development of energy storage systems. Lithium-ion batteries (LIBs) that own merits such as high energy density, long cycle life, and environmental friendliness have been widely used in various electronic apparatus. However, as an anode material of LIBs, market-applied graphite is far from meeting the requirement of high energy density due to its limited theoretical capacity (372 mA h/g), long diffusion pathways for lithium ions, and concomitant lithium dendrite during cycling.[1,2] In the field of LIBs, increasing interest is aroused in coping with the obstacle of high power densities in energy storage, which mainly focuses on the development of various materials to use as electrodes.[3,4] Thereinto, enormous efforts have been made to explore high-performance anode materials for LIBs. Nowadays, transition-metal oxides (TMOs) are being widely explored as anode material due to their features of much higher capacity (∼1000 mA h/g) and enhanced safety from suppressing the formation of lithium dendrites compared to that of commercial graphite, as well as their easy-to-obtain and low price.[5,6] Common TMOs, e.g., iron oxides,[6−9] cobalt oxides,[10−13] nickel oxide,[14−16] and manganese oxides,[17−21] have been frequently studied, as well as metal oxide composites derived from double transition-metal elements, e.g., CoFe2O4,[22,23] Ni2CoO4,[24−26] CoSnO3,[27−31] and MnSnO.[17,32] Among these TMOs, CoSnO3 featuring high capacity (1238 mA h/g) and facile synthesis and MnO featuring low conversion potential (1.032 V vs Li+/Li) and long cycle life emerge as promising alternative candidates to substitute commercial graphite as next-generation anode materials of LIBs.[18,27]

However, all of the TMOs have to face drastic capacity fading caused by intrinsic strain derived from volume expansion during lithiation/delithiation processes and poor electron conductivity, which greatly hinder the practical application of TMOs.[33] To chase these challenges, on one hand, various hollow structures of TMOs, such as hollow cube,[34] hollow sphere,[35] and hollow nanotube,[36] have been applied widely because its cavity can alleviate the volume expansion, boosting the cycling stability.[37] On the other hand, to minimize the negative effect of poor conductivity, some novel strategies about the nanosized incorporation of TMOs with carbonaceous materials have been proposed, where individual and synergetic properties of the compositions can function together to overcome the issues. For instance, the rationally structured carbonaceous composition can offer good electronic conductivity, as well as allow more sufficient utilization of active sites.[38] Furthermore, when carbon structures serve as a shell layer on TMOs, they are also able to release internal stress by offering a cushion effect, which is crucial to maintain the reversible capacity of LIBs.[27] Hence, combining hollow-structured TMOs with carbon layers becomes a promising approach to achieve high-performance LIBs.[39,40]

Recently, bicomponent TMOs in hierarchical structure have been reported to improve electrochemical kinetics, such as CoFe2O4@Fe3O4[41] and NiCo2O4@NiO grown on carbon cloth.[42] However, the direct hybrid of two or more TMOs has the main drawback that the core–shell nanostructure cannot be well maintained for long and thus causes huge capacity variations during lithium cycling. For example, Yang et al. adopted a controllable pyrolysis method of the CoSn(OH)6/Mn0.5Co0.5CO3 precursor to synthesize multicomponent Co–Sn–Mn–O anode electrodes. It delivered 952.6 mA h/g at the 2nd cycle to 145.4 mA h/g after the 160th cycle, which shows extraordinary capacity fading, showing unstable cyclic ability caused by the reactive process during cycling.[43] In contrast, TMO@C@TMO composite with double active materials achieves large capacity as well as a long cycling life and stable cycling performance.[44,45] For instance, Guo et al. reported that SnO2@C@VO2 hollow nanospheres exhibited a stable specific capacity of 597.4 mA h/g at 0.5C after 100 cycles.[44] Zhang and his co-workers adopted Fe2O3@C@MnO2@C as the anode of LIBs, delivering near 900 mA h/g at 0.5C after 50 cycles with a gentle decreasing tendency.[46] We are then convinced that rational usage of carbon layers as a barrier and supporting scaffold is an effective means to maintain the hierarchical morphology of multicomponent TMO-based anode materials and thus promote the cycling performance of LIBs.

Inspired by these works, herein, we successfully fabricate a core–shell hierarchical nanobox of CoSnO3@NC@MnO@NC (CNMN) through facile hydrothermal treatments and one-step calcination (Scheme ). As mentioned above, the hollow structure of cubic CoSnO3 is employed to accommodate volume change and shorten the electron transfer distance. The inner N-doped carbon layer is mainly used as a reaction substrate to form another active MnO layer, as well as a barrier to separate two active TMO materials. The outer N-doped carbon layer is dedicated to the high electron conductivity and preferred revisable capacity. As expected, the LIB with as-prepared CNMN as anode material exhibits a quite high specific capacity of 1195 mA h/g after the 200th cycle at 0.1C and excellent stable reversible capacity of about 876 mA h/g after the 300th cycle at 0.5C with only 0.07 mA h/g fading per cycle after 300 cycles. Even after a 250 times fast charging/discharging cycle both at 5C, the CNMN electrode still retains a reversible capacity of 422.6 mA h/g. These are evidently benefited from the rational combination of CoSnO3 and MnO bicomponent through the carbon layers in a hierarchical hollow structure.

Scheme 1

Schematic Preparation Processing of CNMN Nanobox

Experimental Section

Chemicals and Materials

All of the reagents were adopted without further purification as received. Tin chloride pentahydrate (SnCl4·5H2O), cobalt chloride hexahydrate (CoCl2·6H2O), trisodium citrate dihydrate (C6H5Na3O7), sodium hydroxide (NaOH), potassium permanganate (KMnO4), and tris (hydroxymethyl) aminomethane (C4H11NO3) were obtained from Sinopharm Chemical Reagent Co., Ltd. (China). All of the dopamine hydrochloride (C8H11NO2·HCl) was purchased from Aladdin Industrial Inc.

Synthesis

Synthesis of CoSn(OH)6@PDA

The hollow cube CoSn(OH)6 was synthesized as reported previously.[31] Typically, 5 mL of ethanol solution containing SnCl4 (1 mmol) uniformly was slowly added into 30 mL of sodium citrate (1 mmol) and CoCl2 (1 mmol) aqueous solution under vigorous stirring. Next, 5 mL of 2 M NaOH aqueous solution was added into the above solution drop by drop. After 1 h of stirring, an as-prepared 8 M NaOH aqueous solution of 20 mL was slowly poured into the suspension along with the solution color changing from pink to purple and the reaction lasted for 15 min at ambient temperature. The resulting product was obtained by rinse-centrifugation cycles with deionized (DI) water and ethanol, and dried at 60 °C in the oven.

Forty milligrams of as-prepared CoSn(OH)6 was dispersed in 80 mL of Tris buffer solution with pH ∼ 8.5, and the suspension was sonicated for 20 min to obtain a uniformly dispersed solution. Then, 40 mg of C8H11NO2·HCl (dopamine hydrochloride) was slowly added into the above solution, followed by magnetically stirring for 1 h at ambient temperature, during which the dopamine self-assembled and self-polymerized on the surface of CoSn(OH)6 nanocomposites and a layer of polydopamine (PDA) was formed. The resulting brown powder CoSn(OH)6@PDA was obtained after rinse-centrifugation and dried at 60 °C in an oven in air.

Synthesis of CoSn(OH)6@PDA@MnOOH

Sixty milligrams of the as-synthesized CoSn(OH)6@PDA was added into 80 mL of 0.1M KMnO4 aqueous solution and the mixture was sonicated for 30 min into an evenly dispersed solution and then stirred slowly for 1 h, after which the solution was sealed into a Teflon-lined autoclave with a capacity of 100 mL. The autoclave was heated to 160 °C for 1h and maintained for 6 h. After the reaction, the Teflon-lined autoclave was cooled to room temperature. The resulting powder CoSn(OH)6@PDA@MnOOH was obtained after rinse-centrifugation by ethanol and DI water, and dried at 60 °C in the oven.

Synthesis of CoSnO3@NC@MnO@NC (CNMN)

Following the similar experimental recipe for CoSn(OH)6@PDA, CoSn(OH)6@PDA@MnOOH was further coated to give CoSn(OH)6@PDA@MnOOH@PDA. Forty milligrams of CoSn(OH)6@PDA@MnOOH and forty milligrams of C8H11NO2·HCl were used in the coating experiment.

The as-obtained CoSn(OH)6@PDA@MnOOH@PDA was then annealed under a N2 atmosphere using a three-stage method at 150, 300, and 500 °C, respectively, with a slow ramp rate of 1 °C/min, and each stage was kept for 2 h to get the final product CoSnO3@NC@MnO@NC (CNMN). For comparison, CoSnO3@NC was also synthesized by annealing CoSn(OH)6@PDA at the same temperature as that for CNMN.

Characterization

X-ray diffraction (XRD, Japan D/MAX-2500, Cu Kα, 40 kV, 40 mA) was carried out with the 2θ degree from 10 to 80° at a scanning rate of 8°/min. The morphology information of CNMN and its precursors was characterized by transmission electron microscopy (TEM, 200CX, 200 kV) and field emission scanning electron microscopy (FE-SEM, JSM-6700F, 5 kW), respectively. High-resolution transmission electron microscopy (HRTEM) observations were performed on a JEOL JEM-2100F electron microscope operating at 200 kV. The surface analysis of samples were performed by X-ray photoelectron spectroscopy (XPS) using an X-ray photoelectron spectrometer with a monochromatic Al Kα radiation. Raman spectra were recorded on a Raman spectrometer with 630 nm laser excitation. The nitrogen sorption measurement was carried out on a Micrometric Tristar 3020 analyzer at liquid-nitrogen temperature. The pore volume and specific surface area of products were estimated by Barett–Joyner–Halenda (BJH) and Brunauer–Emmett–Teller (BET) analyses, respectively. Inductively coupled plasma (ICP) analysis was performed on an inductively coupled plasma optical emission spectrometer (Thermo Fisher, Icapq6300). Thermogravimetric analyses (TGA) are conducted on TG-209C with a heating rate of 10 °C/min in air atmosphere.

Electrochemical Measurements

The working anode was prepared by mixing the active material (70 wt %), poly(vinylidene fluoride) (10 wt %), and Super P (20 wt %) into N-methyl pyrrolidone and stirring for 4 h to give a homogeneous slurry, after which it was left on the copper foil uniformly and then dried at 80 °C in a vacuum oven. The mass loading of active materials on the copper foil was ∼1.0 mg. The electrolyte contains LiPF6 (1 M) in a mixture of ethylene methyl carbonate, dimethyl carbonate, and ethylene carbonate with the same volume. The coin cell (type 2032) was assembled in an argon-filled glove box for electrical performance testing, using Li circular foil for counterelectrodes and microporous polypropylene film Celgard 2400 for the separator. The as-assembled coin cell was tested on a LAND CT2001A cell test system under current density ranging from 0.1C to 5C as well as the charge/discharge voltage from 0.005 to 3.0 V. An electrochemical workstation of CHI 660C was applied to investigate the cyclic voltammetry (CV) performance at a scan rate of 0.1 mV/s and further tested from 0.2 to 1.0 mV/s. Electrochemical impedance spectroscopy (EIS) was recorded at a frequency ranging from 0.01 to 100 kHz at an amplitude of 5 mV.

Results and Discussion

Scheme depicts the process of preparing CNMN with a hollow cube core and hierarchical shell structure. First, the hollow cube CoSn(OH)6 was prepared by co-precipitation of Sn4+ and Co2+ with OH– and then etching the interior by concentrated alkaline.[47] Second, the PDA layer serving as a precursor of the N-doped carbon layer was condensed on the surface of the hollow cube, which was also applied for the further reaction substrate and participant.[48] Next, after the facile hydrolysis reactions, MnOOH nanoflakes readily grew on the inner PDA layer. The main reaction between CoSnO3@PDA and MnOOH includes a two-step redox reaction of OH–.[49] The first redox reaction belongs to PDA (serving as electronic donor) and H2O with KMnO4, which is attributed to eq Then, the as-formed MnO2 transforms into MnOOH, mainly attributed to eq , during the hydrothermal process under 160 °CAfter assembling MnOOH, a further layer of PDA was coated on the product as prepared. Thus, CNMN was finally obtained using a three-step annealing approach under a N2 atmosphere, giving a hollow cube-shell structure with double N-doped carbon layers.

Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) (Figure a–h) were applied to carefully investigate the structures and morphologies of CoSn(OH)6, CoSn(OH)6@PDA, CoSn(OH)6@PDA@MnOOH, and the final product CNMN. As shown in Figure a,b, the hollow cubic CoSn(OH)6 nanoboxes present quite a uniform volume with ∼200 nm edge length, a cavity ∼80 nm in length, and outer wall thickness ∼60 nm.[47] As shown in SEM and TEM of CoSn(OH)6@PDA (Figure c,d), the PDA layer is condensed on the surface of the precursor nanoboxes successfully with the thickness of ∼22 nm. As for the further coating of the product of CoSn(OH)6@PDA@MnOOH (Figure e,f), a MnOOH shell of ∼78 nm is generated on the surface of CoSn(OH)6@PDA by chemical absorption, redox, and hydrolysis reactions as mentioned above. In the final product of CNMN (Figure g,h), the thickness of the outer N-doped carbon layer is about ∼23 nm, which is annealed from further outer PDA coating that has a strong attachment and can maintain the morphology of the nanobox. Notably, after annealing, the MnOOH shell evolves into MnO nanoparticles; then, certain free volumes are reserved in the carbon interlayer, which may potentially alleviate the volume expansion of MnO during lithium storage (Figure h). HRTEM was further applied to determine the fine structure of the obtained CNMN (Figure i–k). It can see that CNMN has a uniform amorphous carbon layer coating on the surface of MnO (marked by the red curve in Figure i) and presents two lattice fringes with d-spacings of 0.26 and 0.22 nm (Figure j,k) that are ascribed to the (111) and (200) planes of the MnO layer, respectively. The structures of the products were also characterized by XRD measurements (Figure l). As for CoSn(OH)6, no additional diffraction peaks of impurities were detected, confirming the high purity of the synthesized precursor. The diffraction of CoSn(OH)6@PDA@MnOOH shows three main planes of (111̅), (002), and (113̅) corresponding to MnOOH (JCPDS card no. 41-1379), which demarcated KMnO4 reducing into MnOOH in the hydrothermal reaction.[49] Evidently, five planes (labeled purple #) in the XRD diffraction of CNMN demonstrated the presence of MnO, which is in accordance with the above HRTEM results (Figure j,k). One broad band and two peaks of diffraction observed at around θ = 26, 34, and 52° (labeled blue *) are found identical to the XRD patterns of CoSnO3 previously reported.[50] The element mapping images (Figure S1a–f, in the supporting information (SI)) confirm that the C element and dopant N derived from PDA and Co Sn Mn elements are uniformly distributed within hollow nanoboxes of CNMN. Moreover, further elemental content of CNMN is characterized by both EDS and ICP analyses. As displayed in Figure S2 and Table S1 in the SI, the Mn/Co/Sn molar ratio is about 2:1:1, which implies that the molar ratio of MnO/CoSnO3 is about 2:1. Raman spectra (Figure S3, in the SI) were also used to determine the N-doped carbon layers. The two peaks at 1330 and 1600 cm–1 are associated with the characteristics of the D band (the disordered band) and G band (the graphene band), respectively. The calculated ID/IG = 0.978 attests that the amorphous carbon layer is derived from the PDA composite with a relatively high degree of graphitization.[48] Thermogravimetric analysis (TGA) was carried out to further determine the carbon content in CNMN with the temperature ranging from 30 to 950 °C (Figure S4a, in the SI). Under an air atmosphere, the slight weight loss before reaching 250 °C is attributed to the removal of adsorbed water and the water contained in the birnessite phase.[39,51] In the following 250 to 450 °C stage, the 33.1% weight loss is approximately equal to the content of the carbon layer derived from PDA.[52] Indeed, the N2 sorption isotherms (Figure S4b, in the SI) demonstrate that CNMN shows a BET surface aera of 39.44 m2/g. The calculated pore-size distribution displays that there are many nanopores less than 20 nm (inset of Figure S4b, in the SI), which promotes Li+ transportation and thus improves the lithium storage performance of CNMN.[53]

Figure 1

SEM image (a) and TEM image (b) of CoSn(OH)6; SEM image (c) and TEM image (d) of CoSn(OH)6@PDA; SEM image (e) and TEM image (f) of CoSn(OH)6@PDA@MnOOH; SEM image (g) and TEM image (h) of CoSnO3@NC@MnO@NC; HRTEM image (i–k) of CoSnO3@NC@MnO@NC; XRD patterns (l) of CoSn(OH)6, CoSn(OH)6@PDA@MnOOH, and CNMN.

XPS measurement was then performed to determine the elements as well as valence states of Co, Sn, and Mn in the CNMN product. The survey spectrum in Figure a indicates the presence of Co, Sn, Mn, C, O, and N elements. The high-resolution XPS spectrum of individual elements is presented in Figure b–f. As shown in Figure b, two peaks located at 780.8 and 796.7 eV, along with their satellite peaks (labeled as Sat. Peak), are assigned to Co 2p3/2 and Co 2p1/2, respectively, which agree with Co2+. As for Sn 4d (Figure c), two peaks centered at 486.6 eV (Sn 3d5/2) and 495 eV (Sn 3d3/2) represent the presence of Sn4+ in CNMN. Two peaks at 641.7 and 653.5 eV determined for Mn 2p match well with Mn 2p1/2 and Mn 2p3/2 with the feature of a satellite peak at ∼646 eV (Figure d), and these three peaks are determined to be Mn2+.[54] As Figure e illustrates, N 1s can be deconvoluted into four different peaks that contribute to pyridinic N, pyrrolic N, graphitic N, and oxidized N (chemisorbed nitrogen oxides), confirming the dopant of N in carbon layers in the as-prepared CNMN. Figure f depicts the high-resolution XPS spectrum of O 1s, the intensive peak centered at 530 eV, as well as the peak at 532.1 eV assigned to the oxygen deriving from CoSnO3 and MnO and the surface adsorbed oxygen species such as surface contaminates, respectively.[27] The XPS analysis further demonstrates the successful fabrication of the as-proposed CNMN nanocomposite.

Figure 2

XPS spectra of CNMN (a) survey; (b) Co 2p; (c) Mn 2p; (d) Sn 4d; (e) N 1s; (f) O 1s.

The electrochemical performances of as-prepared CNMN were investigated, as well as those of CoSnO3 and CoSnO3@NC nanoboxes, for comparison. Figure a depicts the first three cycles of cyclic voltammetry (CV) curves of the CNMN anode at a scan rate of 0.1 mV/s. In the first anodic scan, there are three reduction peaks located at ∼0.9, ∼0.65, and ∼0.1 V, respectively, which can be assigned to the conversion reaction of Li to Li2O and Sn4+ to Sn and formation of solid electrolyte interphase (SEI) film, while disappearing in the following cycling processes. Meanwhile, in the first cathodic scan, a weak but broad cathodic band and a broad oxidation peak appear at about 0 to 0.5 and ∼1.3 V, which can be ascribed to the dealloying process of LiSn to Sn and the oxidation of Mn to Mn2+, respectively. Furthermore, during the following anodic scan, the reduction peak appearing at about ∼0.5 V, which is shifted from ∼0.3 V, probably originates from the optimized kinetics and the microstructural change of the anode active materials caused by the generation of Mn and Li2O after the initial lithiation reaction. The two weak oxidations peaks located at ∼1.8 and ∼2.3 V are related to the oxidation of Sn to SnO2 and Co to CoO, respectively, which agrees with CV curves and discharge/charge profiles in Figure S5 in the SI. These above results are consistent with the discharge/charge profiles of the CNMN anode at a current density of 0.1C (Figure b). During the initial three cycles at 0.1C, the as-prepared CNMN delivers the first cycling discharge/charge capacity of about 1571.2/1104.7 mA h/g, related to the initial coulombic efficiency (ICE) of 70.31%. The irreversible loss of 1st cycle capacity could impute to the decomposition of the electrolyte and the formation of SEI film on the electrode. During the first cycle discharge, there are clearly gradual plateaus beginning at ∼0.3 V and a short weak plateau that could correspond to the conversion of Mn2+ to Mn and dealloying of Sn4+ to Sn, as well as the formation of an SEI layer. In the next two cycles, the plateaus related to Mn2+ to Mn shift to ∼0.5 V, which is consistent with the change of the voltage platform in Figure b. It is worth noticing that two curves of the 1st and 2nd cycle almost overlap, displaying that CNMN owns outstanding cyclic stability. Furthermore, after 180 cycles, the specific capacity slowly increases to 1220 mA h/g; this phenomenon is observed also in the cycling performance test (Figure c).

Figure 3

Electrochemical performance of CNMN (a) CV of CNMN at 0.1 mV/s; (b) charge/discharge profiles of CNMN at the current density of 0.1C; (c) cycling performance of CNMN, CoSnO3@NC, and CoSnO3 at the current density of 0.1C; (d) rate performance of CNMN and CoSnO3@NC at current densities ranging from 0.1C to 0.2C, 0.5C, 1C, 2C, and 5C and then back to 0.1C; (e) cycling performance of CNMN at the charging/discharging current density of 0.5C; (f) cycling performance of CNMN at the charging/discharging current density of 5C.

A comparison of the cycling performance at a low current of about 0.1C of CNMN, CoSnO3@NC, and CoSnO3 is carrried out and displayed in Figure c, between the voltage range of 0.005–3 V. During the first cycle, the cycling performance of CNMN delivers discharge and charge capacities of 1572 and 1105 mA h/g, respectively. These results correspond to an initial coulombic efficiency of ∼70.3%, of which the loss probably is caused by the formation of SEI film. In contrast, CoSnO3@NC and CoSnO3 deliver initial discharge/charge capacities of 1209/888 and 1408/988 mA h/g, slightly lower than that of CNMN. However, the capacities of the two samples, CoSnO3@NC, and CoSnO3, go through the same significant decay trend in the following cycles from 2nd to 65th and then CoSnO3@NC shows a slight increasing capacity while the CoSnO3 reversible capacity stabilizes at 350 mA h/g. Although as-prepared CNMN and CoSnO3@NC show a similar tendency in the cycled capacities, CNMN owns much higher capacity with discharge capacities of 924 mA h/g at the 50th cycle, 1014 mA h/g at the 100th cycle, and 1195 mA h/g at the 200th cycle. The significant improvement in the electrical performance of CNMN is due to the extra layer of MnO on the surface of CoSnO3@NC. Besides, benefiting from the unique features of hybrid structures, e.g., the carbon layers (inner and outer), the carbon interlayer, and the hollow structure, can not only maintain the nanobox structure from collapsing completely, as well as accommodate the volume expansion of active materials during lithium insertion/extraction, but also improve electric conductivity of the anode. Moreover, it can be clearly observed that there is a fading in the first 20 cycles, which could be owing to the nanobox with metal oxides suffering complete oxidation into ultrafine metallic alloys.[55] In contrast, there is a gradual increase in the capacity during the cycling process with ∼2 mA h/g per cycle from the 20th to the 200th cycle. The reason why the capacity of CNMN gradually increases can be attributed to the following aspects: (i) CNMN anodes undergo an activation process, where more active sites are provided in electrode materials.[56] (ii) As depicted in the discharge/charge profiles of CNMN electrode at the 100th and 180th cycle (Figure b), it is obvious that the discharge profiles gently shift to higher voltage and the charge profiles gradually change to lower voltages during the cycling number increase from 100 to 180. This suggests the gradually reduced electrode polarization, hinting at the enhancement of reversible capacity.[27] (iii) The electrolyte gradually decomposes during cycling, while Mn2+ oxidizes to a higher oxidation state of Mn4+.[39]

Figure d displays the rate performances of CNMN at rates of 0.1C, 0.2C, 0.5C, 1C, 2C, and 5C, then a return to 0.1C, wherein the as-prepared electrodes deliver average specific capacities of 1048, 942, 880.4, 762.4, and 706.2, 601 mA h/g, respectively, then return to 934 mA h/g, along with a little fading compared with the first five cycles, which could impute to incomplete activation in the first five cycles at 0.1C. Even at 5C, the CNMN electrodes still exhibit a higher specific capacity (601 mA h/g) than the theoretical value of commercial anode electrodes (graphite about 372 mA h/g).[1] As illustrated in Figure e, for evaluating the long cycle life performance, the as-prepared CNMN was tested under 0.1C in the first five cycles in forming stable SEI film and active electrodes, then under 0.5C in the following 295 cycles, showing an excellent cycling ability. The initial discharge capacity of CNMN at 0.5C was 896 mA h/g. Subsequently, the discharge capacity decreased to 806 m Ah/g at the 15th cycle and rose to 878 mA h/g at the 100th cycle, as well as 956 mA h/g at the 200th cycle and 876 mA h/g at the 300th cycle with only 0.07 mA h/g fading per cycle after 300 cycles. Meanwhile, the tendency of specific capacity at 0.5C is similar to that at 0.1C but the values are lower, which may be caused by CNMN at 0.1C exhibiting more diffusion-controlled behavior in contrast to that at 0.5C dominated by capacitive behavior. To further investigate the cycling performance at a high rate, the charging/discharging at a high current density of 5C was performed (Figure f). In the first cycle, the coulombic efficiency loss is owing to the formation of the SEI film on decomposition of the electrolyte. Thus, in the following charging/discharging cycles, the coulombic efficiency increases, indicating an improvement of reversible capacity. Moreover, the slight capacity increase from 374.3 (1st cycle) to 529 (60th cycle) is ascribed to the activation of materials as well as the oxidation of Mn2+ to the higher state of Mn4+. After a 250 times faster charging/discharging cycle at 5C, the CNMN electrodes retain a reversible capacity of 422.6 mA h/g. The cycling performance of CNMN at 0.5C and 5C is better than that of CoSn alloy/carbon, Sn-MnO/carbon, and CoSnO3 hollow cubes-MnOx /carbon materials reported in the literature (Table S2 in the SI).

The merit of the as-prepared CNMN anode on its reaction kinetics was further studied by cyclic voltammograms (CVs). As shown in Figure a, the locations of the two peaks, the reduction peak (below) and oxidation peak (above), shift toward opposite directions as the scan rate ranges from 0.2 to 1.0 mV/s, showing the increase in the reaction overpotential, which further confirmed the reversibility of CNMN electrode. The indicated relationship is as shown in the following eq (v: scan rate; i: measured current): the b-value of 0.5 relates to the diffusion-controlled dynamics, while the b-value of 1 corresponds the ideal capacitive-controlled dynamics. The diffusion-controlled process, such as intercalation, conversion, and alloying, endows the anode with high capacity, while the capacitive behavior promises fast charge transfer.[57]For the CNMN sample, the b values for the reduction and oxidation cycling process, after the linear-fitting, can be determined to be 0.744 and 0.734, respectively (Figure b). The fitting results suggest that CNMN anode exhibits both a capacitive behavior and diffusion-controlled process during the lithiation/delithiation process but the capacitive-controlled dynamics dominate the cycling process. Besides, the total capacitive contribution can be calculated on the basis of the following eq (57)in which i(v) represents the measured current at a fixed potential, v represents the scan rate, and k1V, k2V1/2 stands for the capacitive and diffusion contributions, respectively. Meanwhile, k1 and k2 serving as constants can be obtained by the following eq evolved from the above equation[57]On the basis of the above formula, the contribution of the capacitive capacity for the as-prepared CNMN anode at different sweep rate scans, ranging from 0.2 to 1.0 mV/s, is determined and shown in Figure c,d. Figure c depicts the capacitive-controlled current contribution (pink region) to the total current obtained at the scan rate of 0.8 mV/s, which intuitively shows that the capacitive-controlled contribution is generated mainly at the peak voltage in both reduction and oxidation processes. The capacitive-controlled contribution is about ∼74.3 and ∼25.7% for the cation intercalation-controlled charge of the total current, demonstrating that the as-prepared CNMN owns excellent cycle stability and high specific capacity. In addition, as depicted in Figure d, the gradually increased proportion of capacitive-controlled contribution is from 46.4 to 75.9% along with the increase of scan rates from 0.2 to 1.0 mV/s, which can be ascribed to the fact that the lithium diffusion process declines at high rates.

Figure 4

Reaction kinetics of the electrochemical behaviors of CNMN: (a) CV profiles at various scan rates ranging from 0.2 to 1.0 mV/s; (b) the determination of the b value representing the relationship between the peak current and scan rate; (c) scheme of the contribution of capacitive and diffusion currents at 0.8 mV/s (the pink region represents the capacitive contribution); (d) percentages of capacitive and diffusion contribution at different scan rates, respectively.

The detailed mechanism and reaction kinetics of CNMN, CoSnO3@NC, and CoSnO3 are further investigated by electrochemical impedance spectroscopy (EIS) and shown in Figure . The EIS of CNMN is fitted by an equivalent electrical circuit with Z-view software (Figure a). All of the Nyquist plots are composed of semicircles and straight lines (inset of Figure a), where the inner resistance of the cell (Re), the intercept with the X axis, represents the Ohmic resistance and solution resistance, the charge-transfer impedance (Rct) reflects the charge-transfer resistance between the electrolyte interface and electrode, the constant phase-angle element (CPE1) represents the double layer capacitance and the Warburg impedance (Zw), and the linear slope at low frequencies is related to the ion diffusion impedance in the active materials. In fact, the semicircle at the mid-high frequency reflects the charge-transfer resistance, including (Rct and CPE1), wherein CNMN is estimated to be at ∼48 Ω, smaller than that of CoSnO3@NC (∼91 Ω) and CoSnO3 (∼127 Ω), demonstrating smaller interface impedance for CNMN and could offer better Li+-ion diffusion as well as electronic transport. Thus, CNMN results in superior rate capability than that of CoSnO3@NC and CoSnO3. Besides, for evaluating the diffusion process of Li+ ions, we could calculate the apparent ion diffusion coefficient (D) of CNMN, CoSnO3@NC, and CoSnO3 based on the linear slope at low frequencies through the following eqs and 7(58)whereas R is the gas constant, T is the absolute temperature, A represents the surface area of the electrode based on the coin cells, n is the number of the electrons involved in the electrode reaction, F is the Faraday constant, C is the shuttle concentration of Li+ ions, and σω is the Warburg coefficient with ω as the angular frequency related to Z′. According to eq , the slopes of the fitting line are 80.81, 245.77, and 425.38 by calculating the linear fitting for the low-frequency region of the EIS plots (Figure b). According to eq , D is proportional to σω inversely, which implies that CNMN shows the lowest value of σω and owns the highest D. Therefore, the conclusion is drawn that CNMN anode holds a faster charge-transfer rate as well as higher Li+ diffusion coefficient than those of CoSnO3@NC and CoSnO3 anodes, wherein the impedance evolution is consistent with the electrochemical properties. To investigate the integrity of the recycled electrode, we also carried out TEM for the CNMN recovered by washing the PVDF binder away totally. As shown in Figure c, although CNMN occurs with a slight aggregation phenomenon during lithiation/delithiation, it maintains its hollow morphology after 200 cycles with full charge/discharge at 0.1C. Such a depiction demonstrates that CNMN as anode material owns excellent cycling stability.

Figure 5

(a) Nyquist plots of CNMN anode electrodes. The inset shows the equivalent circuit used to fit the EIS; (b) linear relationships between Z′ and ω–1/2. (c) TEM images of CNMN at a 3.0 V potential state after the 200th cycle at 0.1C.

All of the results demonstrate that CNMN exhibits high performance as the anode material of LIBs, especially stable cycling capability, and high specific capacity at 0.5C. The excellent performance of CNMN can be mainly attributed to the following factors. First, the rational combination of multiple active TMO material of CoSnO3 and MnO is applied to achieve enhanced specific capacity. Second, the hollow structure of CoSnO3 and interspace-rich interlayer structure of carbon for MnO cope with the volume expansion by providing more space that is beneficial to stabilize Li+ cycling. Third, the dual N-doped carbon layers not only separate the two TMOs from direct hybridization but also provide lower charge-transfer resistance and increase the utilization of active sites bringing about the promotion of preferred revisable capacity. Thus, the strategy of building a hierarchical structure with two active TMOs and dual-carbon layers could be extended to other metal oxides for electrode materials.

Conclusions

In summary, dual-TMO-based hollow hierarchical CNMN were successfully fabricated through a facile method. CNMN showed excellent performance as anode material for LIBs. Due to the unique hollow structures of the inner core of CoSnO3 and the outer shell of MnO “sandwiched” between dual N-doped carbon layers, the as-prepared CNMN exhibits a reasonable specific capacity (1195 mA h/g at 200th cycle) without fading in 200 cycles at 0.1C. Importantly, CNMN owns an excellent stable cycling performance at 0.5C with only 0.07 mA h/g fading per cycle after 300 cycles. Even after a 250 times fast charging/discharging cycle both at 5C, it still retains a reversible capacity of 422.6 mA h/g. These superior electrochemical performances could be ascribed to the combination of dual TMOs and hierarchical structure that greatly increases the utilization of TMOs, the carbon layers that improve the conductivity, and the hollow and interlayer structures that serve as efficient supports as well as alleviate the volume changing during cycling. The work demonstrated here could provide a novel strategy for designing a hierarchical structure with two active materials and dual carbon layers to achieve enhanced lithium storage performance.