Introduction

LiCoO2 is the traditional positive-electrode material for lithium ion batteries. However, its low energy density and high cost are currently driving a search for more efficient, economic, and stable materials for use in lithium ion batteries [1–4]. Doping with transition metals such as Mn and Ni has been the main focus of research in this field over the last few decades; an example of a useful doped material of this kind is Li[Ni1/3Co1/3Mn1/3]O2 (NCM 111). However, NCM 111 can only deliver a capacity of ∼155 mAh g−1, which is rather low if it is to be used in batteries for next-generation electric vehicles (EVs). Also, cycling this material at an upper cutoff potential >4.5 V causes the capacity to rapidly fade due to phase transitions from a rhombohedral structure to H1-3 and O1 phases [5–10].

Sakaebe et al. tested LiNi1/3Co1/3Mn1/3O2 at various upper cutoff potentials ranging from 4.5 to 4.9 V and found that the discharge capacity increased with increasing upper cutoff potential, but increasing the upper cutoff potential also led to a fast decline in capacity. They attributed this rapid drop in capacity to polarization caused by increased charge-transfer resistance (R ct), indicating degradation of the interface between the electrode and electrolyte. The increase in polarization and R ct can be effectively suppressed by applying an Al oxide coating, as this inhibits the formation of a rock-salt-structured phase in the surface region of bare NCM 111 during cycling [11].

Thackeray et al. proposed a new strategy to improve the stability of LiMO2 materials by using Li2MnO3. They reported that preparing lithium-rich layered oxide materials (LLOs) xLi2MnO3–(1−x)LiMO2 can yield high electrochemical capacities of ∼250 mAh g−1 when charged to high potentials of >4.6 V, and they also reported that Li2MnO3 plays a major role in structural stability during the charge–discharge process. However, cycling this material at a high upper cutoff potential of >4.6 V still resulted in an irreversible loss of capacity during the initial charge/discharge cycle and damaged the surface of the electrode, leading to high cell impedance [12, 13]. To overcome this issue, Kim et al. created a fluorine-substituted variant of 0.3Li2MnO3–0.7Li[Mn0.60Ni0.25Co0.15]O2, treated the resulting material with acid, and then analyzed the electrochemical performance of this material. Their results showed that the pristine material delivered a discharge capacity of ∼200 mAh g−1 at 0.5 C and retained 85 % of its capacity after 30 cycles in the potential range 2∼4.6 V, while the surface-modified and fluorine-doped material retained 92 % of its capacity when cycled under the same conditions [14]. Thus, the surface treatment appeared to improve the stability of the material. Nevertheless, it retained only 93 % of its capacity after 30 cycles in the potential range 2∼4.6 V. Therefore, LLOs with improved cycling stability and energy density (achieved by cycling them at higher upper cutoff potentials) are still required.

We recently reported a series of materials of formula [Li1.2−x Na x Ni0.62Co0.14Mn0.248O2], where x = 0, 0.05, 0.08, 0.10, and 0.12. Our results showed that coating an LLO with manganese oxide greatly enhanced its cycling stability and suppressed side reactions with the electrolytes. Sodium doping increases the lattice parameter c due to the fact that Na has a larger ionic radius than Li. This facilitates the diffusion of Li ions, reduces charge-transfer resistance, and improves the electrochemical performance of the materials [15].

Herein, we describe a new approach that allowed us to successfully synthesize an LLO material with high energy density and high stability, xLi2MnO3–(1−x)LiMO2, a core–shell material in which the core is LiMO2 and the shell is Li2MnO3. Thus, the aims of this study were (1) to increase the discharge capacity of NCM 111 by preparing NCM 111 with the formula xLi2MnO3–(1−x)LiMO2, i.e., a lithium-rich material with a high capacity [12], and (2) to improve the thermal stability of this material by making it a core–shell structured material with a core of LiMO2 and a shell of Li2MnO3. In this structure, Li2MnO3 has two advantages: it can participate in the electrochemical reactions at potentials >4.5 V, resulting in higher capacities, and it can increase the structural and thermal stability of the material during the charge–discharge process [14, 16, 17].

Experimental section

Materials and methods

Ni0.5Co0.5CO3 was synthesized by a co-precipitation method in which NiSO4⋅6H2O and CoSO4⋅7H2O at a molar ratio of 1:1 were dissolved in deionized water to prepare a 1 M solution, which was then added to a 2 M Na2CO3 solution to form a suspension. This was aged for 12 h under a flow of N2, with the temperature and pH of the solution maintained at 60 °C and 11, respectively. The suspension was filtered, washed, and dried in vacuum at 60 °C.

The core–shell material was prepared as follows. 0.528 g of MnSO4⋅H2O were dissolved in 50 ml of deionized water and added to 100 ml of an aqueous suspension containing 0.5 g of the precursor Ni0.5Co0.5CO3 and 0.4 g of Na2CO3. Aging was performed for 12 h at 60 °C. The resultant suspension was filtered, washed three times with deionized water, and dried in a vacuum at 60 °C.

Li1.15Na0.5(Ni1/3Co1/3)core(Mn1/3)shellO2 was obtained by mixing a stoichiometric amount of the precursor with appropriate amounts of Li2CO3 and Na2CO3. The mixed powder was first heated at 450 °C for 8 h, followed by sintering at 900 °C for 12 h in air. The heating rate was maintained at 2 °C min−1.

The pristine LiNi1/3Co1/3Mn1/3O2 material was synthesized by a co-precipitation method. Stoichiometric amounts of NiSO4⋅6H2O, CoSO4⋅7H2O, and MnSO4⋅H2O were dissolved in deionized water in which the total concentration of metals was 1 M; this was denoted solution A. Solution A was added to a 2 M Na2CO3 solution under vigorous stirring and then aged for 12 h under an N2 atmosphere. The temperature and pH of the solution were maintained at 60 °C and 11, respectively. The resultant suspension was filtered, washed, and dried in a vacuum at 60 °C. LiNi1/3Co1/3Mn1/3O2 was obtained by mixing stoichiometric powder of the precursor with an appropriate amount of Li2CO3. The mixture was first heated at 450 °C for 8 h and then baked at 900 °C for 12 h in air. The heating rate was maintained at 2 °C min−1.

Physical characterization

Powder X-ray diffraction was accomplished on a Shimadzu XD-3A diffractometer using Cu Kα radiation (λ = 0.15406 nm), at a scan rate of 0.15° 2θ min−1 with a step size of 0.02° in the 2θ range 10–80°.

Transmission electron microscopy (TEM) was performed on a JEM2000EX (JEOL, Tokyo, Japan) transmission electron microscope. The morphology was observed using a scanning electron microscope (SEM, FESEM, SU8010, Hitachi, Tokyo, Japan). The electrode was prepared as follows. The synthesized material was mixed with acetylene carbon black (AB) and poly(vinylidene fluoride) (PVDF), weight ratio 80:10:10, and then dispersed in 1-methyl-2-pyrrolidinone (NMP). The resulting slurry was spread onto a disk of aluminum foil with a loading mass of 2∼3 mg cm−2 and then dried in a vacuum oven at 65 °C for 24 h. CR 2032 coin cells were assembled in an Ar-filled glove box with a Li counter electrode and 1 M LiPF6 dissolved in ethylene carbonate (EC) and dimethyl carbonate (DMC) used as the electrolyte. The cells were galvanostatically charged/discharged within the range 2.5∼4.7 V at room temperature, while the pristine material was charged to an upper cutoff potential of 4.5 V. The cells were galvanostatically charged at different current densities to 4.7 V. This was followed by an additional constant-voltage (CV) charging step at 4.7 V until the current reached 70 % of the charge current, and then the cells were discharged galvanostatically to 2.5 V using a battery cycler (battery testing system, Land, Wuhan, China) at room temperature. The current densities used were 20, 100, 200, 400, 1000, and 2000 mA g−1.

Results and discussion

The materials were synthesized by coprecipitating stoichiometric amounts of Ni2+ and Co2+ sulfates with Na2CO3 and aging for 12 h, during which the pH and temperature were maintained at 11 and 60 °C, respectively. The material was synthesized at a high pH to precipitate a higher percentage of cobalt from the solution. Furthermore, the surface of the synthesized material is partially negatively charged at high pH, aiding the creation of a homogeneous coating. The precursor was then coated with manganese oxide, as described in the “Experimental section.” The precursor was mixed with appropriate amounts of Li2CO3 and Na2CO3 to produce the cathode material for LIBs. We doped the core–shell material with sodium to further improve its electrochemical performance. Previously, we reported that sodium doping can significantly enhance electrochemical performance, as it results in the expansion of the lattice parameter c [15].

Figure 1 shows SEM images of the core–shell and pristine materials. It can be seen that the particles of the core–shell material are smaller than those of the pristine material, which can be attributed to the longer aging period applied.

Fig. 1a–b
figure 1

SEM images of a the pristine material and b the core–shell material

Full size image

It can be seen that the core–shell material’s primary particles are plate-like in shape and ∼150 nm in size (Fig. 2a). The selected area electron diffraction (SAED) pattern (Fig. 2b) exhibits sets of lattice fringes with interplanar spacings of 0.470 nm, 0.233 nm, 0.171 nm, 0.155 nm, and 0.111 nm, which can be indexed to the (003) R-3m or (001) C2/m, (012) R-3m or (200) C2/m, (221) C2/m, (132) C2/m, and (223)C2/m diffractions of either a trigonal (layered α-NaFeO2 type, space group R-3m) or monoclinic (space group C2/m) structure, respectively (please refer to the XRD data for more information) [12, 18, 19]. This indicates that the core–shell material is heterostructured with R-3m and C2/m phases. The monoclinic Li2MnO3 (C2/m) is compatible with the trigonal LiMO2 (R-3m) structure due to the fact that the R-3m and C2/m phases both possess interlayer spacing of 4.7 Å, which can be attributed to the 003 and 001 planes of the R-3m and C2/m phases, respectively [12] .

Fig. 2
figure 2

a TEM image of the core–shell material, b electron diffraction pattern of the core–shell material, c TEM image of the pristine material, and d electron diffraction pattern of the pristine material

Full size image

To further compare the structures of the pristine and core–shell materials, we characterized the pristine material using a high-resolution transmission electron microscope (HRTEM) (Fig. 2d and e). The pristine material’s primary particles are also plate-like shape and 100∼200 nm in size. The SAED pattern exhibits lattice fringes with spacings of 0.47, 0.235, 0.143, 0.186, and 0.243 nm, which can be indexed to the 003, 006, 018, 015, and 101 diffractions of the α-NaFeO2-type structure (space group R-3m), respectively. These results indicate that the core–shell material is a mixture of layered (R-3m) and monoclinic C2/m phases, while the pristine material is a pure layered (R-3m) structure.

The XRD patterns and the profiles of Rietveld refinements are shown in Fig. 3. As discussed above, the Li2MnO3 (monoclinic C2/m) and LiMO2 (trigonal R-3m) structures are compatible due to the fact that the interlayer spacing of the close-packed layers is ∼4.7 Å for both structures, which can be attributed to the 003 and 001 planes of the R-3m and C2/m phases, respectively [12]. Note that the XRD data are often difficult to interpret uniquely for these two systems because all of the peaks for the R-3m structure can also be indexed to the C2/m structure. However, reflections that appear in the 2θ range ∼20–23° (Cu Kα) can only be indexed to the monoclinic structure and are a signature of the LiMn6 unit in the Li2MnO3 [20]. Therefore, the Rietveld refinements were carried out in this study by assuming that the core–shell material is a mixture of trigonal (R-3m) and monoclinic (C2/m) phases, while the pristine sample is a pure trigonal phase, because no sign of the C2/m phase in the 2θ range ∼20−23° (Cu Kα) has been observed [12–14, 21].

Fig. 3a–b
figure 3

Profiles for Rietveld refinement of the a core–shell and b pristine materials. Red dots indicate the calculated profile; black solid line indicates observed data; green dots indicate Bragg positions; blue solid line indicates (Y obs – Y calc)

Full size image

For the trigonal (R-3m) phase, Li, a transition metal (Ni, Co, or Mn), and O occupy the 3a (0 0 1/2), 3b (0 0 0), and 6c (0 0 1/4) sites, respectively. The Ni2+/Li+ disorder between the 3a and 3b sites is considered to be due to the similar ionic radii of the Ni2+ (0.76 Å) and Li+ (0.70 Å) ions [22, 23].

Some selected results of Rietveld refinements obtained using the Fullprof package [24] are listed in Table 1. Table 1 shows that the pristine material exhibits a pure R-3m phase, while the core–shell material is a mixture of R-3m and C2/m phases. Table 1 also shows that the M*–O bonds (M* = Ni, Mn, Co) of the core–shell material are stronger than those of the pristine material. This should result in increased structural stability, decreased oxygen release during charge/discharge processes, and reduced Ni2+ migration to Li slab because it makes vacancy formation in the transition metal layer more difficult [25].

Table 1 Results of Rietveld refinement for the synthesized materials
Full size table

Figure 4 shows SEM-EDS data on the core–shell structure. It indicates that the actual chemical composition of the material is very similar to the proposed composition.

Fig. 4
figure 4

EDS data on the core–shell-structured material

Full size image

XRD data were employed to determine the electron-density distributions in the pristine and core–shell materials using the Fullprof package. The results (Table 2) show that the electronic density at the Li site (0 0 1/2) is lower in the core–shell material, which can be ascribed to the decreased amount of Ni2+ in the Li layer.

Table 2 Electron-density distributions of the synthesized samples
Full size table

The fits of the Ni 2p orbital peaks to the XPS spectra of the pristine and core–shell materials are shown in Fig. 5. Since the binding energy of the Ni(III) 2p orbitals is higher than that of Ni(II), it is possible to calculate the Ni3+ and Ni2+ contents of the synthesized materials. The results show that the Ni2+ content of the core–shell material is more than that of the pristine material. This increase in Ni2+ may arise because some of the Mn4+ from the Li2MnO3 shell interdiffuses into the surface of the core material and Ni2+ from the core material diffuses into the shell, which causes the core and the shell to fuse together. This fusion definitely increases the chemical and thermal stability of the material.

Fig. 5a–b
figure 5

XPS spectra of the a pristine and b core–shell structured materials

Full size image

Electrochemical performance

Figure 6 displays the initial charge–discharge profiles of the samples at 20 mA g−1 and different cutoff potentials. We charged the pristine sample to an upper cutoff potential of 4.5 V because potentials higher than this can lead to a rapid decrease in capacity and structural collapse [11], while the core–shell material could be cycled to an upper cutoff potential of 4.7 V with high stability.

Fig. 6
figure 6

Initial charge–discharge profiles of the pristine and core–shell structured materials at 20 mA g−1

Full size image

The pristine sample delivers an initial discharge capacity of 168 mAh g−1 at 20 mA g−1, while the core–shell material delivers 208 mAh g−1 at the same current density (Fig. 6). This shows that the core–shell structured material delivers a much higher discharge capacity than the pristine material, because LLOs can deliver a high discharge capacity when cycled at an upper cutoff potential of >4.5 V [12–14, 16, 17, 20, 21, 26, 27].

It can be seen that the core–shell material operates at a much higher potential than the pristine material (Fig. 6). The discharge capacity of the core–shell structured material increased to 218 mAh g−1 after cycling for four cycles, due to the electrochemical activation of the Li2MnO3 phase (Fig. 7) [16, 28].

Fig. 7
figure 7

Cycling performances of the pristine and core–shell structured materials at 20 mA g−1 for 10 cycles, followed by 90 cycles at 100 mA g−1. The pristine sample was cycled between 2.5 and 4.5 V, while the core–shell sample was cycled between 2.5 and 4.7 V

Full size image

The pristine and core–shell materials deliver discharge capacities of 203.4 and 155.2 mAh g−1 at a current density of 100 mA g−1, and retain 79 and 90 % of their peak capacities after 90 cycles at room temperature, respectively (Fig. 7). This high stability of the core–shell structure can be attributed to the Li2MnO3-phase shell and stronger M*–O bonds (as mentioned earlier in the discussion of the XRD analysis results), which can provide high structural stability in the highly delithiated state [12, 13, 28].

To further investigate the differences between the pristine and core–shell materials, we also tested the half-cells at different current densities (Fig. 8). The core–shell material delivered the highest discharge capacity of 218 mAh g−1 at 20 mA g−1, which was much higher than that of a recently reported material [11]. For example, Xia et al. reported that a solid solution of LiNi0.8Co0.1Mn0.1O2 and Li2MnO3 which was cycled at an upper cutoff potential of 4.5 V delivered a discharge capacity of 207 mAh g−1 and retained 75 % of its capacity after 100 cycles at 20 mA g−1 [29]. Moreover, the core–shell material synthesized in this study exhibited a higher discharge capacity than the pristine material and better stability than it at all current densities (Fig. 8).

Fig. 8
figure 8

Rate capability performances at different current densities of the pristine and core–shell structured materials at room temperature. The pristine sample was cycled between 2.5 and 4.5 V, while the core–shell sample was cycled between 2.5 and 4.7 V

Full size image

The average discharge potential is highly dependent on the discharge profile [30]. In LiMO2 materials, voltage fading is a serious issue because voltage fading results in a dramatic decrease in energy density, even though discharge capacity fading can be prevented [24–29]. Figure 9 compares the average potential profiles of the core–shell and pristine materials, and shows that the core–shell material operates at much higher potential than the pristine material, with less potential fading upon cycling. This indicates that the presence of the Li2MnO3 shell significantly increases the energy density and improves structural stability [16, 29, 31, 32].

Fig. 9
figure 9

Average voltage vs. number of cycles for the core–shell and pristine materials at a current density of 20 mA g−1 for 10 cycles followed by 90 cycles at 100 mA g−1. The pristine sample was cycled between 2.5 and 4.5 V, while the core–shell sample was cycled between 2.5 and 4.7 V

Full size image

Cyclic voltammetry was also carried out to study the differences between the pristine and core–shell materials (Fig. 10). The prepared cells were scanned anodically, followed by a corresponding cathodic scan. A couple of redox peaks were observed for both samples. The anodic peaks located in the potential range 4–4.2 V can be attributed to the redox couple of Ni2+ and Co3+ [33].

Fig. 10
figure 10

Cyclic voltammograms of the pristine and core–shell structured materials

Full size image

The core–shell material displays a broad oxidation peak with a higher current than the pristine material. In addition, one more reduction peak (at ∼3 V) was observed for the core–shell material, which can be attributed to the Mn3+/Mn4+ redox couple; this could explain the increased capacity shown by this material [17].

Conclusions

Li1.15Na0.5(Ni1/3Co1/3)core(Mn1/3)shellO2 was successfully synthesized via coprecipitation and subsequent baking at 900 °C for 12 h in air. Using XRD, TEM, and XPS, we confirmed the structure of this material. The core–shell material is composed of a core of a layered LiMO2 phase (R-3m), while the shell is a monoclinic Li2MnO3 phase (C2/m).

Rietveld refinements show that the core–shell structured material has less Li+/Ni2+ cation disorder than the pristine material. They also reveal that stronger M*–O (M* = Mn, Co, Ni) bonds are present in the core–shell material, which leads to high structural stability during the charge–discharge process.

The core–shell material can be cycled at a higher upper cutoff potential of 4.7 V than that of the pristine material (4.5 V). Therefore, it can be concluded that the presence of the Li2MnO3 shell substantially improves the electrochemical performance of the material in terms of its discharge capacity, energy density, and cycling stability. Li1.15Na0.5(Ni1/3Co1/3)core(Mn1/3)shellO2 delivers a high discharge capacity of 218 mAh g−1 at 20 mA g−1 and retains 90 % of its discharge capacity at 100 mA g−1 after 90 cycles at 100 mA g−1. This discharge capacity is much higher than that of the classic NCM 111 material. The core–shell material displays a higher average voltage, a higher energy density, and better cycling stability than those of the pristine material.