Introduction

Nowadays, the lack of energy resources and environmental pollution has forced researchers to explore the next-generation cathode materials to replace the traditional commercial LiCoO2 cathode. Ni-rich layered cathode with a high specific capacity, low cost, and environmentally friendly is one of the most promising cathodes [1,2,3,4,5,6], whereas the lack of stable structure and excellent cycle life cathode materials has also hindered its commercialization of LIBs. To pursuit of high capacity with more nickel content of cathode, the unsatisfied cycle life and poor thermal stability appear, especially at elevated high cutoff potential [7,8,9]. During the high-temperature calcination process, the nickel (II) in the transition metal sites easily migrates to lithium sites on account of a similar radius of Li and Ni (0.76 Å vs. 0.69 Å) which induce severe cations mixing region on the surface with inactive NiO-like phase [10,11,12,13]. Besides, the detrimental phase transformation of the spinel phase blocks the channels of lithium ions diffusion and induces the degradation of electrochemical performance. Moreover, the oxygen release aggravates the side reaction of cathode material and electrolyte at the in-depth delithiation process, even the fatal thermal runaway of LIBs [14, 15].

To address above the tough problems, numerous strategies have been proposed, which devote to decrease cation mixing or modify surface chemistry properties. Such as lattice element doping (Ti4+ [16], Al3+ [17], La3+ [18], Mg2+ [19], Zn2+ [20], Zr4+ [21], PO43− [22]), surface coating (TiO2 [23], ZrO2 [24], CoF2 [25], SnO2 [26]), and so on. Although surface coating can suppress surface side reactions by constructing an inactive coating layer at the beginning of the cycle, severe cation mixing resulting from TM ions migration in bulk at extending cycles limits any improvement [27]. Therefore, to fundamentally address these problems of layered Ni-rich cathodes, it is necessary to consider the precise regulation of the crystalline structure of the material. Lattice doping can alter the electronic structure at atomic level, such as cation ordering, charge redistribution, and the change of metal–oxygen covalency [28, 29]. From a commercial perspective, material cost and electrochemical properties should be the two main considerations. Therefore, cheap and abundant elements (such as Al, Fe, Mg, Ti, Ca, and Mn) might be preferred for use in layered Ni-rich cathode, while expensive elements (such as Rh, Ru, Ga, and Bi) cannot be used in commercial applications [9, 30]. Due to same valence state, similar ionic radius Al3+ to Co3+ (0.535 Å vs. 0.545 Å), and the much stronger covalent bond than Co. So, Al doping can reduce the Li+/Ni2+ mixing to maintain the structural integrity and mitigate the formation of surface residual Li compounds [14, 31]. However, the single modification strategy has little effect on the electrochemical performance at elevated cutoff potential. As far as we know, Ti element can enhance the electronic and ionic conductivity for the charge compensation and the larger ionic radius (0.605 Å) than Co [16, 32]. Consequently, the synergistic effect of Al and Ti cations co-doping strategy may improve the integrity of structure and inhibit the unfavorable phase transformation at different cutoff potential. As far as we know, there are few reports on the co-doping of Al and Ti in LiNi0.8Co0.1Mn0.1O2 cathode.

Herein, we proposed a novel strategy involving uniform Al and Ti co-doping in LiNi0.8Co0.1Mn0.1O2 cathode. The modified cathode was obtained by mixing the precursor Ni0.8Co0.1Mn0.1(OH)2, LiOH·H2O, Al2O3, TiO2, and subsequent high temperature calcination. The schematic diagram of material preparation is shown in Fig. 1a. Compared with the pristine sample, the modified sample with the incorporation of Al and Ti cations exhibits excellent rate capability and satisfying discharge capacity as well as at elevated cutoff potential.

Figure 1
figure 1

a Schematic illustration of the preparation process. b XRD patterns of NCM and NCM-AT samples. c Crystal structure diagram-based LiMO2 model. The fitting spectra of Rietveld refinement of NCM (d), and NCM-AT (e)

Full size image

Experiment section

Materials synthesis

The precursor material was obtained via the traditional hydroxide co-precipitation method. Specifically, the uniformity mixing solution of NiSO4·6H2O, CoSO4·7H2O, and MnSO4·5H2O (in a chemical stoichiometric ratio of Ni/Co/Mn is 8:1:1) was continuously fed into a 500-mL reactor with an appropriate feeding rate. Meanwhile, NH4OH solution and NaOH solution were fed into the reactor simultaneously at a proper feeding rate and amount. For removing O2, the atmosphere of the reactor was filled with N2. The reactor temperature remained at 50 ˚C, and the pH value of the reactor was maintained accurately at 11.0 via pH meter. After the 12 h reaction, the final precipitate was filtered and washed with de-ionized water for several times and dried in a vacuum oven at 120 for 10 h. Finally, the precursor material was mixed with LiOH·H2O (0.5% excess) for 30 min. The well-mixed material was preheated for 5 h at 480 °C and then heated for 15 h at 820 °C under an oxygen flow, and then naturally cools it to room temperature. The resulting material was labeled NCM. The preparation and calcination process of the modified material was the same as that of the raw material, only 1 mol % Al2O3 and 1 mol % TiO2 were added in the mixing process and the modified material was labeled NCM-AT.

Materials characterizations

The crystal structure of all samples was detected by powder X-ray diffraction (XRD) using Cu Kα radiation under the angle of 10–80°. The full-pattern refine treatment was obtained via GSAS software. The surface morphology was obtained by scanning electron microscopy (SEM). The surface element distribution was evaluated using energy-dispersive spectroscopy (EDS) mapping. The transmission electron microscope (TEM, FEI) was used to investigate the microstructure or micromorphology. The element's valence state and composition were obtained by the X-ray photoelectron spectroscopy (XPS). The elements binding energies in the spectra of all chemical elements were calibrated by standard C 1 s spectra (284.8 eV), and XPS PeakFit software was performed the spectra fitting of all samples.

Electrochemical measurements

All prepared materials using coin type half-cells (CR2025) to evaluate the electrochemical performance. To obtain the final electrodes, the as-prepared cathode powders, conductive acetylene black, and material binder (weight ratio of 80:13:7) were mixed, and N-methyl pyrrolidone (NMP) is the mixture solvent. After mixing in a ball mill for 20 min at 520 rpm, the mixing slurry was coated on the aluminum foil evenly and dried for 12 h at 120 in a vacuum oven. Then, the coated aluminum foil was punched into a wafer with a diameter of 14 mm, and the loading mass of electrode was approximately 2 mg. Finally, the half-cells were assembled with the Li metal as the counter electrode in a glove box under argon atmosphere. The electrolyte was 1 M LiPF6 soluble in a mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC) (volume ratio of 1:1). NEWARE software was used for evaluating electrochemical performance. The half-cells were tested in the voltage range of 2.7–4.3 V and 2.7–4.5 V (vs. Li/Li+) at room temperature, and 1 C refers to 180 mA g−1. The electrochemical impedance spectroscopy (EIS) was obtained by electrochemical workstation with the frequency (100 kHz – 0.01 Hz).

Results and discussion

Figure 1b shows the XRD patterns of NCM and NCM-AT samples under the angle of 10–80°. All samples exhibit the α-NaFeO2 layered structure with the space group of R-3 m [33]. Then, the well-ordered layered crystal structure of the two samples can be indicated by the apparent splitting of (006)/(102) and (108)/(110) pairs. Moreover, there are no other impurity peaks in the two samples, implying that the successful incorporation of Al3+ and Ti4+ into the crystal structure and co-doping has little effect on the layered structure. Based on a layered LiMO2 model (Fig. 1c), the full-pattern Rietveld refinement was carried out to GSAS software, and the well-fitting spectra are shown in Fig. 1d, e. Typically, the values of Rwp and Rp (< 10%) indicate that refinement results are credible (Table 1) [34]. As can be seen the lattice parameters in Table 1, it is easily known that the lattice parameters (i.e., a, c, c/a) of the modified sample are larger slightly than these of the pristine sample. Specifically, the lattice parameters a is 2.872722 and 2.875890 Å, respectively, and the lattice parameters c is 14.204408 and 14.233287 Å, respectively. As is known to all, the parameter of c is related to the lithium ions diffusion channel [35, 36]. Therefore, the larger value of c is beneficial to lithium ions diffusion when repeated cycling process. The increase in lattice parameter c on account of the large radius Ti4+ (0.605 Å) doping into the bulk structure. Moreover, the c/a value of all materials larger than 4.9 also provides further evidence of a well-ordered layered structure. Typically, the Li/Ni mixing deteriorates the electrochemical performance due to the presence of an inactive NiO-like phase. In layered materials (such as NCM811), the ratio of I(003)/I(104) can represent the degree of cations mixing (Li/Ni). And the higher the ratio is, the smaller the mixing degree is [37]. According to the results of Table 1, the value of pristine and modified samples is 1.321 and 1.384, respectively, suggesting that a modified sample possesses low-content Li/Ni mixing. After the refinement of material, the Li/Ni mixing ratio of NCM and NCM-AT is 2.19% and 1.42%, respectively. It can be concluded that the mixing degree of the modified material decreases and the structural stability of the material is improved.

Table 1 Refined structure parameters of NCM and NCM-AT
Full size table

To obtain N-rich with excellent electrochemical performance, it is necessary to design an appropriate precursor for the latter lithiation reaction. The morphology of precursor is shown in Fig. 2a–c; a spherical secondary particle is composed of primary flake particles via hydroxide co-precipitation method, using to prepare cathode materials. After evenly mixed with lithiation sources and calcined at high temperatures, all obtained materials retain the spherical shape and the secondary particle size was ~ 10 μm (Fig. 2d–i). The surface morphology of primary particles was different when carefully observed. As can be seen from the yellow ellipse dotted line box from Fig. 2f, i, the secondary particle of NCM-AT is denser, and the porosity of the surface is smaller, implying that Al and Ti co-doping strategy has some effect on the surface morphology where the side reaction between electrolyte and cathode material could be inhibited to some extent. Moreover, to further understand the existence of Al and Ti cations on the NCM-AT material, EDS mapping adopted and the images of the elements of NCM-AT are shown in Fig. 2j-o. It is easy to see the elements of Al and Ti of the NCM-AT surface show a uniform distribution. Therefore, combined with XRD analysis, uniform Al and Ti cations are successfully incorporated into the crystal lattice.

Figure 2
figure 2

SEM image of precursor (a–c), NCM (d–f) and NCM-AT (g–i). EDS mapping images of NCM-AT sample (j–o)

Full size image

To further investigate the microstructures of materials, HRTEM was performed, and Fig. 3 shows the corresponding images of all materials. Figure 3a and b exhibits the images of primary particles of NCM and NCM-AT, respectively. Taking from the yellow square zone in Fig. 3a, it is easy to see the interplanar spacings of 0.236 nm indicate the well-layered structure of NCM. Moreover, the fast Fourier transformation (FFT) of the white region in the bulk of NCM provides further evidence for the layered structure. Similarly, the (003) crystal plane of the layered structure was detected in the white square zone of NCM-AT material [30, 38]. It can be concluded that the two materials exhibit a high ordered layered phase. In conclusion, this phenomenon reveals that the incorporation of Al and Ti cations is successfully doped into the crystal lattice and is no visible effect on the structure of the material.

Figure 3
figure 3

HRTEM image and corresponding FFT pattern of NCM (a) and NCM-AT (b)

Full size image

To further figure out the valence state's variation of surface elements after modification, XPS analysis was obtained due to its high surface sensitivity. The elements of Ni, Co, Mn, O, Ti, Al, and C were detected in the NCM-AT (Fig. 4a). The binding energy peaks of 73.6 and 457.6 eV are attributed to the Al and Ti 2p (Fig. 4b, c) [39]. It is concluded that Al and Ti cations are successful incorporation into the lattice structure. Normally, the binding energy peaks of Ni2+ and Ni3+ are approximately 855.20 eV and 856.40 eV, respectively [40]. As shown in Fig. 4d, e, all samples are mixed valence of + 2 and + 3 for Ni oxidation state. Moreover, the ratio of Ni2+/Ni3+, via the semiquantitative analysis, for NCM-AT sample, is larger than the NCM sample. As the incorporation of high valence Al3+ and Ti4+, for the charge compensation, a fraction of high valence Ni3+ reduces to low valence Ni2+. To maintain the structure stability, the Mn oxidation state of all samples remains + 4 in Ni-rich layered materials (Fig. 4f) [41].

Figure 4
figure 4

The survey spectrum of all samples (a). XPS spectra of Al 2p (b) for NCM-AT. Ti 2p (c) for all samples. Ni 2p for NCM (d) and NCM-AT (e), and Mn 2p (f) for all samples

Full size image

The initial charge–discharge curves with the voltage range of 2.7–4.3 V at 0.1 C are shown in Fig. 5a. In the 1st cycle, the specific capacity of pristine and modified samples is 177.32 and 189.70 mAh g−1, and the corresponding coulombic efficiency is 83.79 and 85.45%, respectively. This capacity difference of the two samples may be on account of the enlargement of Li ions diffusion channels and low the cation mixing. Figure 5b shows the cycle life of all materials after 200 charge–discharge cycles at 1 C. The specific capacity retention rate of NCM and NCM-AT is 46.67 and 76.75%, respectively. Remarkably, the material exhibits improved cycle life after the modification. To obtain the excellent cathode of higher energy density LIBs, the constant working voltage is crucial. The average discharge voltage of the two materials during cycling is shown in Fig. 5c. From the 1st cycle to the 200th cycle, the average difference voltage of the NCM sample is 0.377 V, with the voltage retention of 90.04%. In sharp contrast, the average difference voltage of the NCM-AT sample is 0.035 V, with voltage retention of 98.78%. The modified sample exhibits an excellent electrochemical performance of average discharge voltage. Therefore, the co-doping strategy of Al and Ti cations can improve the discharge capacity and inhibit the normal discharge voltage fading during repeated cycling. Figure 5d-f presents different discharge capacity under a different current density of all materials. Comparing with the rate capability of NCM, it is easy to see NCM-AT exhibits the excellent rate capability. Moreover, NCM-AT still possesses a high capacity of 128 mAh g−1 at a large current density of 10 C. In comparison, NCM exhibits poor rate capability and attains 105 mAh g−1 under the same condition. The improved capability on account of the inactive Al and Ti ions can act as a pillar role when repeatedly cycling, resulting in reduce structure distortion and stabilize the bulk structure. But usually too many foreign ions will hinder the transfer of lithium ions, so it needs to be controlled in a proper number of foreign ions. Moreover, Fig. 5g shows the cycle life of all materials at 10 C. In sharp contrast, after 200 charge/discharge cycles, NCM and NCM-AT exhibit a discharge capacity retention of 29.34 and 73.36%, respectively. Moreover, Fig. 5h depicts the initial charge and discharge curves of all materials between 2.7 and 4.5 V at 0.1 C. During the initial cycle, the discharge capacity of NCM and NCM-AT is 189.25 and 194.21 mAh g−1, and the corresponding coulombic efficiency is 78.08 and 81.62%, respectively. The modified material possesses a high capacity than pristine material. Also, NCM cathode suffers from severe discharge capacity decay. At the 200 cycles, NCM cathode exhibits a capacity of 89.53 mAh g−1 with a capacity retention of 49.44%, while NCM-AT cathode not only increases the capacity but also shows a better cycle life (Fig. 5i). The NCM-AT cathode delivers an improved capacity of 129.24 mAh g−1 with the capacity retention of 70.93%. In summary, the modified materials possess better electrochemical performance at different cutoff voltages.

Figure 5
figure 5

a Initial charge–discharge curves between 2.7 and 4.3 V at 0.1 C. b Cycle life at 1 C after 200 cycles. c Average discharge voltage over the cycling of all materials. d Rate performance for all materials. e, f Discharge curves at different current density. g Cycle life at 10 C. h Initial charge–discharge curves all materials between 2.7 and 4.5 V at 0.1 C. i Cycle life at 1 C after 200 cycles

Full size image

To explain the enhancement in the electrochemical performance in the discharge capacity retention, the structural changes with electrochemical reactions during cycling were obtained by the dQ/dV profiles at different cutoff voltages (Fig. 6). In general, the multiple peaks represent phase transitions corresponding to the contraction/expansion of lattice parameters. Ni-rich cathode usually underwent several phase transitions, including the hexagonal phase (H1), monoclinic phase (M) at the initial charging stage, and hexagonal phase (H2/H3) at the subsequent charging stage. In this phase transformations, the phase transition of H2/H3 is harmful to the layered phase, causing the bulk structure to collapse and turning the layered phase to the inactive NiO phase [42, 43]. Avoiding the phase transition of H2/H3 is crucial to realize the reversibly of cathodes. The redox peaks for the NCM-AT cathode remain relatively stable and exhibit well reversibly during the extend cycles at the voltage range of 2.7–4.3 V (Fig. 6b). In sharp contrast, the NCM cathode shows macroscopic changes and poor reversibly (Fig. 6a). It can be concluded that NCM-AT cathode mitigates the harmful phase transition to achieve superior capacity and voltage retention during cycling. In addition, the dQ/dV profiles at the voltage window of 2.7 and 4.5 V were obtained (Fig. 6c, d). Regarding the NCM cathode, the position of the peak for anodic shifts to higher voltage range and the peaks position for cathodic shifts lower voltage range during 1st to 100th cycling, indicting the severe polarization after repeated delithiation/lithiation process at elevated cutoff voltage. However, the anodic and cathodic peak positions retain relatively stable in NCM-AT cathode at the same condition. The profile analysis of NCM and NCM-AT shows that the improved samples have good electrochemical reversibility. Because the spinel phase is more stable than the layered structure, the layered phase has irreversible transformation to the thermodynamically stable spinel phase, and the presence of foreign ions inhibits the migration of transition metal ions and is conducive to the maintenance of structural stability and the avoidance of spinel phase transition to some extent [44, 45].

Figure 6
figure 6

The differential capacity curves of NCM (a) and NCM-AT at different cycles (b) at 2.7–4.3 V, 2.7–4.5 V for NCM (c) and NCM-AT (d)

Full size image

Electrochemical impedance spectroscopy (EIS) was carried to analyze the dynamic behavior. The Nyquist plots of the two cathodes before cycle and after 100 cycles are shown in Fig. 7a. Typically, Nyquist plots are made up of two parts, a semicircle in the medium–high frequency region and a slope line in the low-frequency region [46, 47]. Charge-transfer resistance (Rct) represents charge transfer impedance, and Warburg impedance (Zw) has a relationship to the diffusion coefficient of lithium ions. The solvent electrolyte resistances are expressed by Re. Moreover, based on the equivalent circuit model (Fig. 7b inset), the fitting results of Re, Rf, and Rct showed in Table 2. Before initial cycle, the Re and Rct values of the NCM sample are 3.06 and 31.53 Ω, respectively. However, the NCM-AT shows the smallest values of Re and Rct with 1.04 and 17.07 Ω, respectively. As the charge and discharge progresses, after 100 cycles, the values of Re, Rf, and Rct of NCM are 8.31, 23.67, and 166.76 Ω, respectively. In contrast, the values of Re, Rf, and Rct of NCM-AT are 4.75, 8.48, and 91.49 Ω. It is evident that the modified material exhibits a low impedance, which may be the result of inhibiting the inactive phase transition. The proper amount of pillar ions introduced can stabilize the structure and maintain the complete lithium ion transport channel. The EIS results analysis also agrees with the dQ/dV profiles. Therefore, the incorporation of Al and Ti into the lattice can reduce the lithium diffusion barrier to promote dynamic behavior during cycling. As is known to all, the lithium ions diffusion coefficient (DLi+) was evaluated by the EIS pattern and the calculation equations of DLi+ as follows:

$$D_{{{\text{Li}}}}^{ + } = R^{2} T^{2} /2A^{2} n^{4} {\text{F}}^{4} {\text{C}}^{2} \sigma^{2}$$
(1)
$$Z^{\prime} = R{\text{f}} + R{\text{ct}} + \sigma \omega^{ - 1/2}$$
(2)
Fig. 7
figure 7

Nyquist plots of the electrodes (a) before cycle and (b) after 100 cycles of NCM and NCM-AT electrodes

Full size image
Table 2 The values of Re, Rf, Rct, and DLi+ for electrodes of all samples
Full size table

In the equations, the R refers to the gas constant (8.314 J K−1 mol−1), T represents room temperature (293.15 K), F refers to the Faraday constant (96485 mol−1), A refers to the reaction area of the electrode (1.766 × 10−4 m2), n refers to the number of electrons and losses during cycling, C is the lithium ions concentration, and σ is the Warburg coefficient [47, 48]. According to results from Table 2, the values of diffusion coefficients of NCM sample are 9.17385 × 10–11, 1.16024 × 10–12 at initial cycle and 100 cycles. In contrast, the NCM-AT sample shows good lithium ion diffusivity under all conditions with 1.02465 × 10–10 and 6.57985 × 10–12. It can be seen that the modified samples have a high lithium ion diffusion coefficient, which may be because the mixing degree of Li+/Ni2+ cations is reduced and the blockage of the diffusion channel is avoided. And this result is consistent with the excellent rate performance of the modified sample.

Conclusion

In summary, we successfully synthesize uniform Al3+ and Ti4+ modified LiNi0.8Co0.1Mn0.1O2 cathode by a high-temperature solid-phase reaction. The results show that the co-doping strategy can improve the electrochemical performance. XRD and HRTEM results show that the modified material retains the original intact layered structure. The introduction of foreign ions reduces the degree of material mixing, accelerates the lithium ion transfer, and slows down the harmful phase transition to maintain the structural stability. In addition, the electrochemical performance of the material is still good under high voltage. In general, this simple cation co-doping strategy provides a feasible way to develop high energy density lithium ions layered oxide cathode.