Introduction

In recent years, lithium-ion batteries have received intense attention as the power source in hybrid electric vehicles (HEVs), plug-in hybrid electric vehicles (PHEVs), and full electric vehicles (EVs) among the currently available energy storage technologies due to their high energy density, long cycle life, and environment friendly [1, 2]. However, the traditional lithium-ion batteries commonly based on layered LiCoO2, LiMn2O4, and LiFePO4 can hardly fulfill the requirement of high-power applications. Thus, design and develop alternative positive-electrode materials with higher capacity, longer cycle life, and lower cost are one of the key challenges for lithium-ion batteries.

Due to its higher energy density than traditional cathode materials, more and more researchers focus their attention on Ni-rich cathode materials, such as LiNi0.8Co0.15Al0.05O2 (NCA), LiNi0.6Mn0.2Co0.2O2 (622), LiNi0.7Mn0.15Co0.15O2, LiNi0.8Mn0.1Co0.1O2 (811), and LiNixMn1−xO2 (x ≥ 0.6) [2]. Some of them have been applied in lithium-ion battery (LIB) systems as novel and promising positive-electrode materials. Unfortunately, their cycling and thermal stability are not satisfactory, further research is very urgent [3,4,5, 10, 17]. According to researchers’previous work, doping with small amounts of elements, such as Mo6+ [6], Al3+ [7], Cr3+ [8, 9], Mg2+ [8, 15, 16, 18], Fe3+ [11, 13], K+ [12], Zr4+ [14], Ti4+ [19, 20], is considered as an effective way to improve the electrochemical performance of Ni-rich cathode materials. Recent research progresses proved that Nb has become one of the most attractive candidates as doping elements since it can enhance the structural stability and electrical conductivity of cathode materials [21, 30]. Yi et al. [22, 23] reported that Nb-doped LiMn1.5Ni0.5O4 can decrease the charge transfer resistance of LiMn1.5Ni0.5O4 and improve the rate performance. Wu et al. [24] synthesized LiNi1/3Co1/3Mn1/3O2 and the high cutoff voltage performance was improved by Nb doping. Also, the electrochemical properties are significantly enhanced in lithium-rich layered cathode materials through niobium (Nb) doping [25], and corresponding theoretical calculations verify the results [26, 27]. However, the influence of Nb doping on structure and electrochemical properties of LiNi0.7Mn0.3O2 cathode materials have rarely been investigated.

In this paper, Nb-doped LiNi0.7Mn0.3O2 was synthesized by calcining the mixtures of LiOH·H2O, Nb2O5, and Ni0.7Mn0.3(OH)2 precursor from a simple continuous co-precipitation method and the impact of Nb substitution on structure, charge transfer resistance, valence state of Ni and Mn, and electrochemical properties were mainly discussed.

Experimental

Material preparation

Nb-doped Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) cathode materials have been prepared successfully by calcining the mixtures of LiOH·H2O, Nb2O5, and Ni0.7Mn0.3(OH)2 precursor from a simple continuous co-precipitation method. Stoichiometric amounts of NiSO4·6H2O and MnSO4·H2O were dissolved in deionized water to form 2.0 mol/L mixed solutions of Ni and Mn. In the same way, 4 mol/L NaOH solution and proper amount of NH3·H2O solution were added to form the mixed alkaline solutions. Metal hydroxide precursors Ni0.7Mn0.3(OH)2 were prepared via co-precipitation in a continuously stirred tank reactor under 55 °C solutions, N2 atmosphere, and the pH of 11.2 ± 0.1. After the completion of the reaction, the resulting suspension was washed with distilled water several times and dried in an oven at 100 °C for about 24 h. Finally, the Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) cathode materials were obtained by calcining the mixtures of the hydroxide precursors, proper amount of Nb2O5, and 10% excess amount of LiOH·H2O (to compensate for lithium loss during the high-temperature calcination process [28]) at 550 °C for 6 h, subsequently at 820 °C for 15 h under flowing air condition.

Material characterization

The structures of the samples were characterized by X-ray diffraction (Bruker D8 Advance, Cu-Kα radiation with a step of 0.02° at 40 kV and 40 mA). X-ray Rietveld refinement was performed with Reflex programs. The morphology and particle size of Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) were investigated with scanning electron microscopy (ZEISS EVO/MA10 with an accelerating voltage of 20 kV). The TEM images were obtained by a transmission electron microscope (FEI Tecnai G2 F20). In addition, the corresponding sample was dispersed into ethanol and transferred onto a Cu grid with holey carbon foil, then dried for TEM observation. Cyclic voltammetry (CV) of the cells were collected at room temperature on Ivium-n-Stat electrochemical workstation between 2.5 and 4.6 V with the scan rate of 0.1 mV s −1. EIS measurements were carried out in the same electrochemical station with 10 mV AC signal and a frequency range from 105 to 0.001 Hz. Charge-discharge performance was characterized galvanostatically in the voltage range of 2.75–4.35 V using a NEWARE cell test instrument at room temperature.

Battery preparation

The electrochemical performances of all samples were measured in the CR2032 coin-type half-cells. The cathode active materials slurry was prepared by mixing cathode material powders with Super P carbon black and polyvinylidene fluoride (PVDF) with a weight ratio of 90:4:6 in N-methyl-2-pyrrolidine (NMP) at room temperature. The slurry was stirred overnight and coated onto aluminum foil, then dried thoroughly at 120 °C. The resulting electrode film was subsequently pressed and punched into pellets and dried in a vacuum chamber at 60 °C for 12 h. Coin cells were assembled in an argon-filled glove box with 1 M LiPF6 in a mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) as the electrolyte solution, and Celgard 2300 polypropylene was employed as the separator.

Results and discussion

XRD phase

The results of X-ray diffraction patterns of the prepared Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) materials are respectively shown in Fig. 1. All samples could be indexed by a hexagonal α-NaFeO2 structure [29]. The slight split of (006)/(102) and (108)/(110) indicates that the products possess typical layered structure characteristics [30]. However, there is some impurity phase of Li3NbO4 in the patterns of Li[Ni0.7Mn0.3]0.96Nb0.04O2, meaning that Nb5+ do not completely substitute Ni(Mn) 3a site. Moreover, with the x increase (x from 0 to 0.02), the (003) and (104) diffraction peaks slightly shift to lower 2θ, which means the interlayer spacing is extended after doping with Nb5+, so the diffusion coefficient of Li+ may increase [31].

Fig. 1
figure 1

XRD patterns of the prepared Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) materials

Full size image

Lattice constants shown in Table 1 were obtained by Rietveld refinement method and the results of X-ray diffraction patterns are illustrated in Fig. 1. The little difference between the experimental and calculated patterns and the low values of Rwp demonstrated that these were successful refinements. The lattice constant a and c and cell volume V expand regularly with the Nb-doping content increase, which is close to the results reported by Wu et al. [24]. This is attributed to the larger Nb5+ (0.64 Å) substitutes Ni3+ (0.56 Å) and Mn4+ (0.53 Å) sites. However, the change of cell parameters begins irregularly with the further increase of Nb content, which mainly because of the Nb5+ react with Li+ to form a new Li3NbO4 phase. According to the reports before [32], the integrated intensity ratio of (003) and (104) peaks (I(003)/I(104)) and the value of c/a, are regarded as an important index of the cation mixing. The larger I(003)/I(104) and c/a value, the lower Li+-Ni2+ ion mixing in the doped samples. Nb substitution has improved the crystallinity of LiNi0.7Mn0.3O2.

Table 1 Lattice parameters of Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) samples obtained from Rietveld refinement of the XRD patterns
Full size table

Microtopography

The SEM images of Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) samples are shown in Fig. 2. Compared with the undoped sample, the particles of all the Nb substitution samples are made up of abundant small primary grains. This may increase the electrode-electrolyte contact area and facilitate the lithium-ion transportation, and then improve the electrochemical performance. But the average particle size of Nb substitution samples is about 10 μm, which was roughly same as the LiNi0.7Mn0.3O2 sample. The spherical morphologies are similar with Li’s [25] report.

Fig. 2
figure 2

SEM images of LiNi0.7Mn0.3O2 (a, b), Li[Ni0.7Mn0.3]0.99Nb0.01O2 (c, d), Li[Ni0.7Mn0.3]0.98Nb0.02O2 (e, f), Li[Ni0.7Mn0.3]0.97Nb0.03O2 (g, h) and Li[Ni0.7Mn0.3]0.96Nb0.04O2 (i, j)

Full size image

To further investigate the changes in microstructure and morphology for the LiNi0.7Mn0.3O2 electrodes before and after Nb doping, TEM and HRTEM images were recorded as shown in Fig. 3. HRTEM images with different lattice spacing were detected and assigned to the {003} plane. Compared with the undoped samples, the lattice spacing of the samples doped by 2% Nb was 0.476 nm, which was slightly larger than the value of the undoped sample (0.470 nm). Nb doping made the lattice slightly expand in the c axis direction and increased the lattice spacing, which would be more conducive to the transmission of lithium ions. This conclusion was in agreement with the results discussed in the section on the XRD investigation.

Fig. 3
figure 3

Images of LiNi0.7Mn0.3O2 materials. a Undoped. b Doped with 2% Nb

Full size image

Electrochemical performance

Cyclic voltammetry (CV) is a useful electrochemical tool to study the redox behavior in an electrochemical reaction. Figure 4 shows cyclic voltammograms of Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) electrodes between 2.75 and 4.6 V with a scanning rate of 0.1 mV s−1, and values of the third cycle CV peak are listed in Table 2. The redox peaks mainly located near 3.9 V are ascribed to the oxidation/reduction of Ni2+/Ni4+or Ni3+/Ni4+, and small redox peaks appear near 4.3 V can be attributed to the phase transition of hexagonal to hexagonal (H2 to H3) for Ni-rich layered compounds [32, 34]. The first cycle is different from the subsequent ones due to the mechanism of initial activation and stabilization which forming a solid electrolyte inter-phase [33, 34] or the cation mixing [35]. Compared to the undoped sample, Nb-substituted samples exhibit less potential difference between anodic peak and cathodic peak (△ψV) showed in Table 2. All these results manifest that Nb substitution could reduce electrode polarization.

Fig. 4
figure 4

Cyclic voltammograms of LiNi[0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) materials at a scan rate of 0.1 mV/s

Full size image
Table 2 Values of the CV peaks for LiNi[0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) materials
Full size table

The initial charge/discharge curves and the capacity data of all the samples are shown in Fig. 5a and Table 3, respectively. It can be seen that the initial charge capacity of LiNi0.7Mn0.3O2 is 259.12 mAh/g, approaching their theoretical values of 274 mAh/g. However, with the increasing Nb doping, the initial discharge capacity decreased. As illustrated in Table 3, the electrodes with x = 0, 0.01, 0.02, 0.03, and 0.04 deliver discharge capacities of 179.78, 184.32, 170.85, and 164.03 mAh/g, severally.

Fig. 5
figure 5

a Initial charge-discharge capacity curves of all samples. b Cycling performance of all samples. c Rate performance of all samples

Full size image
Table 3 Initial charge-discharge capacity data of all the samples proceed at 0.1 C
Full size table

Figure 5b shows the cycling performance of Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) samples operated between 2.75 and 4.35 V at 0.2 C. The discharged capacity of pristine sample and Nb-doped sample (1, 2, 3, 4%) at the 50th cycle are 133.2, 150.6, 156.3, 144.2, and 141.7 mAh/g, respectively, which corresponds to capacity retention of 75.8, 90.0, 91.8, 89.9, and 88.3%. Therefore, Nb substitution could improve the cycle performance effectively.

The rate capability test results for the Li[Ni0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) cells are shown in Fig. 5c. Due to the internal polarization of lithium intercalation/deintercalation reaction, the discharge capacity of all electrode materials gradually decreased with the current density increased [36]. It is easy to find that the x = 0.02 sample exhibits better performance than x = 0 sample, especially at high rates. While the x = 0.04 sample manifests worse rate property due to the impurity Li3NbO4 phase in material. The expanded Li+ channels and the facilitated migration of Li+ ions in Nb5+ substitution samples (x = 0.01, 0.02, 0.03) contribute to the good rate performance. These results are consistent with the expectation from previous lattice analysis.

In order to investigate the internal resistance change between the 4th cycle and the 50th cycle, electrochemical impedance spectroscopy measurement (EIS) technology was applied; the EIS profiles are demonstrated in Fig. 6a, b, respectively. Rs, Rsf, and Rct calculated from EIS results are tabulated in Table 4. According to the reported references [24, 32], the symbols Rs, Rsf, Rct, and Zw represent the ohmic resistance between the working electrode and the reference electrode, the resistance for Li+ diffusion in the surface layer, the charge transfer resistance, and the Warburg impedance of solid phase diffusion, respectively. Obviously, the resistance for Li+ diffusion in the surface layer (Rsf) and the charge transfer resistance (Rct) of Nb-doped samples are very close to each other in Fig. 6 which is much smaller than LiNi0.7Mn0.3O2(x = 0). It reveals that the ionic conductivity was improved greatly by Nb doping and this is the reason of good electrochemical performance obtained before.

Fig. 6
figure 6

Nyquist plots and equivalent circuit for LiNi[0.7Mn0.3]1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) positive-electrode materials at charge state of 4.0 V after 4 cycles, the scatters represent the experimental data and the continuous lines represent the fitted data

Full size image
Table 4 The fitting results of EIS after the 4th and 50th cycle
Full size table

XPS analysis

As discussed above, LiNi[0.7Mn0.3]1−xNbxO2 materials with the x = 0.02 show the best electrochemical performance. To gain more insights into the oxidation state of transition metals in the layered LiNi0.7Mn0.3O2 material and x = 0.02 sample, X-ray photoelectron spectroscopy (XPS) were recorded as shown in Fig. 7. The Ni 2p spectra were composed of two main peaks, Ni2p 3/2 (855.52 eV) and Ni2p 1/2 (873 eV), which was attributed to Ni3+ ions based on the simple ionic model [37], but another small peak, Ni2p 3/2 (854.39 eV) was corresponding to Ni2+. The Ni was with the mixing valence states. More importantly, the relative peak area ratios of Ni3+/Ni2+ before and after doping were 7.7040 and 4.9575, respectively, meaning that the Ni3+ component gradually decreased, whereas the growth of the Ni2+ component at 854.39 eV became evident. The binding energy of Mn2p 3/2 was about 642.2 eV, in agreement with the value reported for MnO2 [38] in both two samples. The result indicated that the oxidation valence of Mn was kept constant in tetravalent, only Ni changed after doping.

Fig. 7
figure 7

XPS spectra of Ni 2p (a0, a1), Mn 2p (b0, b1) of the pristine LiNi0.7Mn0.3O2, 2% doped-LiNi0.7Mn0.3O2 materials (XPS spectra of a0, b0 refer to pristine sample; a2, b2 refer to 2% doped sample)

Full size image

Conclusion

Li(Ni0.7Mn0.3)1−xNbxO2 (x = 0, 0.01, 0.02, 0.03, 0.04) cathode materials have been prepared successfully by calcining the mixtures of LiOH·H2O, Nb2O5, and Ni0.7Mn0.3(OH)2 precursor formed through a simple continuous co-precipitation method. The effect of Nb substitution on the structure, morphology, and electrochemical properties of LiNi0.7Mn0.3O2 was investigated in detail. The results indicate that Nb substitution has an effective effect on the improvement of cycling performance of LiNi0.7Mn0.3O2. Moreover, Nb substitution could decrease the electrode polarization and the 2% Nb-doped sample has the lowest charge transfer impedance, best cycling performance at 0.2 C under the potential of 2.75–4.35 V among all the doped samples. In the end, X-ray photoelectron spectroscopy (XPS) indicated that the Mn were kept tetravalent but Ni was in + 2 and + 3 mixing valence states and Ni3+ component decreased after doping. This study will be helpful for improving the cycle performance of cobalt-free Ni-rich cathode materials.