1 Introduction

Lithium-ion batteries have been widely used because of their advantages such as high-working voltage, high energy density, excellent rate performance, safety, long cycle life, low self-discharge, and no memory effect [1]. However, drawbacks [2] such as large first charge–discharge irreversible capacity loss, instability at high-voltage cycle, and high cost of LiCoO2 limited their applications. Preparation of cathode materials with low price, high capacity, and good stability becomes the current research direction.

Mn-based Li-rich material as cathode material for lithium-ion batteries was the research focus referring to layered Li2MnO3, and layered LiMO2 (M=Ni, Co, Mn) formed a solid solution [3]. First proposed by Numata [4] in 1997, Li-rich material attracted widespread concern for its high capacity and stable cycle characteristics. Compared with conventional layered cathode material, Mn-based Li-rich material, xLi2MnO3·(1−x)LiMO2 (M=Ni, Co, Mn) [5], contains a certain amount of lithium in transition metal layer. Thinking from the perspective of defect chemistry, incorporation with the excess amount of lithium prevents from forming vacancies due to Li volatilization and from making the excess lithium into interstitial lattice to form a solid solution. Although it showed excellent electrochemical performance, its poor circulation performance and low first coulombic efficiency cannot be ignored, which can be improved by doping with metal ions such as Mg, K, Fe, Sn, Zn, Zr, etc. [6–11] or coated with C, Al2O3, CeO2, and fluoride [12–16]. The doped metal ions may generally maintain the structure stability of the material to promote circulation performance, and surface coating could prevent the direct contact of the active substance with the electrolytic solution to avoid the decomposition of the electrolytic solution. Recently, hollow microspheric active material that can improve the rate capability and cycle performance was synthesized, as the hollow sphere increases the surface area and shortens Li+ migration path, stabilizing material microstructure [17].

Many methods have been used to synthesize Mn-based Li-rich materials, such as solid phase sintering method [18], sol–gel method [19], combustion method [20], hydrothermal method [21], and co-precipitation method [22]. The sol–gel method is to use a highly active compound as a precursor, together with raw materials to form a liquid phase and then be subjected to hydrolysis and condensation to generate a stable transparent sol system. The resulting gel network is filled with the solvent without fluidity, which was then calcinated to produce nanomaterials. The sol–gel method has many advantages; for instance, it can achieve the molecular-level mixing of the raw materials and precursors, it requires relatively low-temperature treatment, small and narrow diameter distribution of products can be produced, simple equipment and easy operation are needed and so on [23]. Chelating agents have effects on the performance of the synthesized materials. In this work, Li-rich material, Li1.2Mn0.54Ni0.13Co0.13O2, was prepared by a sol–gel method using different chelating agents. The effects of the chelating agents on the structure and electrochemical performance of the materials were investigated.

2 Experiment

2.1 Preparation of materials

All the chemicals used in this work were of analytical grade and with no further purification. The cathode material Li1.2Mn0.54Ni0.13Co0.13O2 was synthesized by the sol–gel method using different chemical compounds as a chelating agent. A stoichiometric amount of Ni(CH3COO)2·4H2O, Co(CH3COO)2·4H2O, Mn(CH3COO)2·4H2O, and LiCH3COO·H2O was dissolved in distilled water with 5% excess of LiCH3COO; the solution was labeled as A in which total cation concentration was 2 mol L−1. An amount of chelating agent was dissolved in distilled water, and the solution was labeled as B. Solution B was added dropwise into stirring solution A. Thereafter, the pH value of the solution was adjusted using ammonia water, and the solution was evaporated under continuous stirring at 90 °C until the viscidity maroon aquogel was formed. After drying at 120 °C in a drying oven overnight, the xerogel was crushed, subsequently heated at 450 °C for 5 h in air to decompose the organic constituents and acetate components. The sample was then grounded, pelletized, and calcined at 900 °C for 12 h followed by grounding for later use as the temperature cooled down to room temperature. In this study, citric acid, glycolic acid, and polyvinyl pyrrolidone (PVP, K30) were used as chelating agents, and the corresponding samples were labeled as CS, GS and PS, respectively.

2.2 Performance measurements

All the samples were characterized using X-ray diffraction (XRD, Rigaku Ultima D/max-RB 12 KW) with a Cu Kα radiation source. Data were collected in the range 10–90° at a scanning rate of 5° min−1. Morphological studies on the samples were performed using a Hitachi SU8010 scanning electron microscope (SEM).

Electrochemical measurements were performed using galvanostatic cycling with two-electrode coin-cells (type CR2025). The sample: acetylene black: PVDF = 8:1:1 with an appropriate amount of N-Methyl-2-pyrrolidone (NMP) was ground to a uniform syrupy substance. After the slurry was uniformly coated on aluminum foil, it was moved into the vacuum drying oven and then formed into electrode sheets. In a glove box filled with argon, metallic lithium was used as the negative electrode sheets with the prepared sheets as cathode to assemble CR2025 button cell. The electrolyte used 1 M LiPF6 in EC:DEC (1:1). Constant current charge–discharge cycle test was carried out on battery test system at room temperature; the voltage range was 2.0–4.8 V.

A CHI electrochemical workstation (CHI660e, Shanghai, China) was used to perform cyclic voltammograms (CV) of the prepared materials between 2.0 and 4.8 V at a sweep rate of 0.1 mV s−1 at room temperature. A two-electrode system will be used to conduct the Electrochemical Impedance Spectra (EIS) of the cells in which the working electrode is cathode and the reference and the auxiliary electrode is lithium anode by the CHI660e electrochemical workstation before charge and discharge at frequencies from 10 MHz to 0.01 Hz with an AC perturbation signal of 5 mV. The EIS was analyzed using the Zsimpwin software. The potentials throughout the paper are referenced to the Li+/Li couple.

3 Results and discussion

3.1 Characterization of Li1.2Mn0.54Ni0.13Co0.13O2

Figure 1 shows the XRD pattern of the samples synthesized with different chelating agents. The lattice parameters of the samples are calculated and presented in Table 1. It can be seen from Fig. 1 that the diffraction peaks are sharp, indicating the well crystallinity of the samples. All peaks can be indexed to PDF-#20-1115 showing a α-NaFeO2 structure with \(R - \bar 3m\) space group. The well-separated diffraction peaks of (006)/(012) and (108)/(110) demonstrate the layered structure of the synthesized lithium-rich material [24]. The values of c/a and I (003)/I (104) generally are indicative of the ordering and the cation mixing degree of the layered structure, respectively [25, 26]. From Table 1, the c/a values of the composites were all larger than 4.899, indicating that the materials have a better layered structure. The PS sample has a largest c/a value among the three samples, which illustrates that this sample has a good rate performance. The I (003)/I (104) values of the synthesized materials were all greater than 1.2, meaning less cationic mixing, which means that the cationic mixing of CS sample is the smallest. The c value represents the distance between Li2O layer and MO layer. It can be seen from Table 1 that the c-axis of the material changes slightly with different chelating agents. The PS sample with a larger distance between Li2O layer and MO layer leads to easier intercalation of Li-ions, which could result in high discharge capacity. Dahn [27] found that R-factor (R = (I (006) + I (012))/I (101)) could be used to measure the structural order of the layered material—the lower the value, the higher the order of the material that possesses better electrochemical performance. Considering the R-factor in Table 1, the sample prepared with citric acid has the highest R value, which means that its electrochemical performance may be the worst. Meanwhile, the values of the sample prepared with glycolic acid and PVP are almost the same.

Fig. 1
figure 1

XRD patterns of Li1.2Mn0.54Ni0.13Co0.13O2

Full size image
Table 1 Lattice parameters of samples prepared with different chelating agents
Full size table

The morphology and size of the samples have certain influence on the electrochemical performance. Small particle size will lead to a better rate performance. When the material is in nanoscale, the diffusion path for Li-ions’ migration is greatly reduced, resulting in the increased carrier de-embedding and improved high-current discharge performance. Figure 2 shows SEM images of materials synthesized with different chelating agents. It can be seen that different chelating agents have little effects on the dispersibility of the particles. The particle size of the material with citric acid is the largest, whereas the particle size of the material with PVP is the smallest, indicating that macromolecular chelating agent can reduce the particle sizes. As a result, the large-rate performance of the material with PVP could be better than that of the other two materials.

Fig. 2
figure 2

SEM images of a, b CS sample; c, d GS sample; and e, f PS sample

Full size image

3.2 Electrochemical performance

Studies on the de-/insertion Li-ions process of the lithium-rich materials reveal that the potential vs. Li+/Li value of lower than 4.5 V indicates that the reaction accords with the traditional lithium-ion deintercalation theory; that is, Li+ deintercalating from Li1.2Mn0.54Ni0.13Co0.13O2 lattice follows Ni2+/Ni4+ and Co3+/Co4+ redox processes. Once the potential is higher than 4.5 V, Li+ escapes from the material lattice and reacts with O2− to form Li2O, corresponding to the obvious plateau at 4.5 V in the initial charging process. It should be noted that only part of Li+ can be embedded in the material in the initial discharge process; Li+ and O2− vacancies caused by Li2O emerged from the electrode material were occupied by metal ions partly. Owing to the disappearance of lattice vacancies, a part of Li+ could not be embedded into the lattice, which causes the large irreversible capacity loss in the initial charge/discharge process [28–30].

The electrochemical tests are shown in Fig. 3. Figure 3a shows the initial charge and discharge curves of the sample prepared with different chelating agents. It can be seen that the initial discharge capacity of the samples prepared with PVP and citric acid is 231 and 207 mAh g−1, respectively. The initial charge/discharge efficiencies of the samples prepared with citric acid, glycolic acid, and PVP were 70.4, 69.0, and 70.9%, respectively. Figure 3b is the rate performance of different materials. The capacity of CS and GS samples under low discharge rate is similar. When the discharge current reaches 1 C and above, the capacity of CS sample decreases rapidly, i.e., 38 mAh g−1 at 5 C, whereas the PS sample shows excellent rate performance with 99 mAh g−1 discharge capacity at 5 C. Therefore, PVP can improve the rate performance of the material, which is consistent with the previous XRD and SEM analysis.

Fig. 3
figure 3

a Initial charge/discharge curve of CS, GS, and PS samples; b rate performance of CS, GS, and PS samples

Full size image

Figure 4 shows the SEM images of materials synthesized with different chelating agents after 100 charge/discharge cycles. The samples’ surface became unsmooth, and the particle size became bigger after cycles. The surface of the CS sample was the most rough of all and its particle size was the biggest, which corresponded to the worst cycle performance. The unsmooth surface may be caused by the dead lithium during charge and discharge cycles, which caused less available Li+ for intercalation/deintercalation and capacity decay.

Fig. 4
figure 4

SEM images of a, b CS sample; c, d GS sample; and e, f PS sample after 100 charge/discharge cycles

Full size image

Figure 5a shows the cycle performance of the samples at a current of 1 C. It can be seen that the discharge capacity of CS sample under 1 C is the lowest and that of PS sample is the highest. After 100 cycles, the capacity of GS sample decreased from 150 to 138 mAh g−1, with a high capacity retention of 91.7% compared with 70.4% for the PS sample. Therefore, glycolic acid as a chelating agent can improve the cyclic stability of the material, whereas PVP as a chelating agent can increase the discharge capacity of the material.

Fig. 5
figure 5

a Cycle performance of CS, GS, and PS samples; b charge/discharge profiles at 1 C of 2nd, 20th, 40th, 60th, 80th, and 100th cycles for GS sample; c charge/discharge profiles at 1 C of 2nd, 20th, 40th, 60th, 80th, and 100th cycles for PS sample

Full size image

During the charge/discharge processes, besides the electrolytic decomposition, the internal resistance will be changed with the collapsed structure of the cathode material forming Li-dendrite, which results in increased dead lithium and decreased discharge capacity [31]. Charge/discharge profiles at 1 C of 2nd, 20th, 40th, 60th, 80th, and 100th cycles for Li1.2Mn0.54Ni0.13Co0.13O2 samples prepared with glycolic acid and PVP are shown in Fig. 5. With the increase of the cycle number and the decrease of the discharge platform voltage, the charge and discharge capacity of the material decays gradually. The longer the discharge platform is, the greater the available capacity will be. The discharge platform of the PS sample decreased from 3.4–2.9 V to 3.1–2.4 V after100 cycles and that of the GS sample was from 3.4–2.7 V to 3.1–2.4 V. The above-mentioned phenomenon is known as the voltage decay of LMNCOs; it causes a reduced energy density and decreased capacity. This phenomenon may be caused by three reasons—activation of the Li2MnO3 component [32], phase transformation from layered structure to spinel-like structure [31, 33], and the polarization during cycles [34]. The initial discharge platform of the PS sample was the highest, which means higher output power. However, with increasing cycles, the discharge platform voltage of PS sample dropped quickly, and the output power was decreased.

CV was carried out to explore the electrochemical redox responses of different Li1.2Mn0.54Ni0.13Co0.13O2 electrodes. As shown in Fig. 6, all the cathode electrodes have two obvious anodic peaks at 4.0–4.3 V and 4.5–4.7 V in the initial anodic process. The first anodic peak could be attributed to the oxidation reactions of Ni2+ and Co3+, which corresponds to the first voltage plateau during the initial charge/discharge process; the second anodic peak could be attributed to the lithium ion removal and simultaneous oxygen evolution, which is irreversible and corresponds to a large irreversible capacity thus leading to the disappearance of the second anodic peak during the following two scanning laps. In addition, a negative shift of the anodic peak potential was observed, which becomes smaller with the increase of scanning number, illustrating a good reversibility of the cathodes. It should be noted that there is still a little anodic peak at 4.5 V in the second cycle for CS sample, indicating that the CS sample was not fully activated in the first cycle to lead to a low efficiency during the second charge/discharge process. In the reversal scan, a distinct reduction in peak occurs at 3.3 V, corresponding to Mn4+ reduction to balance the charge of oxygen vacancies arising from the loss of oxygen in the initial charge process. The current peaks representing the reduction of Ni4+ and Co4+ were in the potential region of 3.6–4.4 V. The peak intensity of two cathodic peaks is increased as the scan continues, i.e., the cathodic peak is more obvious in the second and third cycle curves, indicating that the cathode material was further activated. These characteristics are consistent with the initial charge/discharge curves of the samples.

Fig. 6
figure 6

CV curve of: a CS sample, b GS sample, and c PS sample

Full size image

To investigate the kinetic processes of Li-ion intercalation/deintercalation for the Li-rich layered materials, EIS measurements were carried out. Figure 7 is the Nyquist plots of Li1.2Mn0.54Ni0.13Co0.13O2 materials prepared with three different cleating agents. The impedance spectra of the Li-rich layered material generally contain several regions, reflecting the serial nature of Li-ion intercalation/deintercalation processes. Before the electrochemical inactivation of cells, the high-frequency region showing a semicircle is ascribed to Li-ion migration through the surface, the medium-frequency region is ascribed to charge-transfer reaction at the electrolyte-electrode interface, and the low-frequency region is ascribed to solid-state diffusion of lithium ions inside the intercalated compound [35]. The intercept of the high-frequency spectrum and the real part of the impedance diagram represent the ohmic resistance (R s) of the solution. The semicircular diameter corresponds to the surface charge-transfer resistance (R ct), followed by a slash which corresponds to the Warburg impedance of the lithium-ion diffusion. From Table 2, it could be seen that R s of different samples could be ignored compared with R ct, whereas R ct (114.1 Ω) is the smallest for PS. The finding indicates that the ion’s transmission channel of PS is short, which can accelerate the Li+ deintercalation and thus it results in a good rate performance. The surface charge-transfer resistance of the PS sample is lower than that of the other two materials, corresponding to the highest rate of capability among the three materials.

Fig. 7
figure 7

Nyquist plots of PS, CS, and GS samples

Full size image
Table 2 R s and R ct values of PS, CS, and GS samples
Full size table

From the above results, it could be concluded that the material prepared with PVP shows high initial discharge capacity and good rate performance and poor cycle performance. To enhance the cycle performance, graphene was coated on the PS sample. Figure 8 shows the cycle performance at 1 C rate of the pristine PS sample and the one coated with 3% graphene, namely Gr-PS. The capacities and capacity retention rates are shown in Table 3. The discharge capacity of the Gr-PS sample decreased from 162 to 145 mAh g−1, with the capacity retention up to 89.8% after 100 cycles at a current of 1 C. The discharge capacity loss took place within the first 20 cycles, and the capacity retention of later 80 cycles was 95.4% (152–145 mAh g−1). The improvement of the cycle performance could be attributed to graphene, which prevents the direct contact of the decomposition product HF of electrolyte with the active material. The HF could destroy the surface structure of active material to lead to the decrease in discharge capacity [36].

Fig. 8
figure 8

Cycle performance of the PS and Gr-PS samples at 1 C rate

Full size image
Table 3 The capacities and capacity retention rates of the PS, GS, and Gr-PS samples at 1 C rate
Full size table

Figure 9 shows the Nyquist plots of the PS and Gr-PS samples after 100 cycles in discharge state, and the R s and R ct values are shown in Table 4. It can be seen that the value of R s and R ct increased after 100 cycles, and the value of R s also could be ignored compared with that of R ct. After coated with 3% grapheme, the R ct value of PS sample was decreased to 71.57 Ω. The reduction of charge transfer resistance was due to the conductive property of graphite [37]. After 100 cycles, R ct of the PS sample increased to 697.1 Ω, whereas that of the Gr-PS sample was just 364.1 Ω; so Gr-PS sample exhibits a better cycle performance.

Fig. 9
figure 9

Nyquist plots of the PS and Gr-PS samples

Full size image
Table 4 R s and R ct values of the PS and Gr-PS samples
Full size table

4 Conclusion

Citric acid, glycolic acid, and PVP were used to synthesize Li-ion battery cathode material, Li1.2Mn0.54Ni0.13Co0.13O2, by a sol–gel method. It was found that all the samples showed a α-NaFeO2 structure with \(R - \bar 3m\) space group. In the case of PS, the material has the smallest particle size (200–350 nm) and the best rate performance and poor cycle stability. The PS sample has the highest initial discharge capacity (231 mAh g−1) and the highest initial charge/discharge efficiency (70.9%). When glycolic acid was used, the sample displayed excellent cycle performance with the capacity retention rate of 91.7% after 100 cycles under 1 C current. EIS tests show that the CS sample showed the largest charge-transfer resistance (R ct), whereas the PS sample showed the smallest R ct, which could be responsible for its superior discharge capacity and rate performance. Coating with 3% graphene on the PS sample could improve the cycle performance, giving the capacity retention of up to 89.8% after 100 cycles at the current of 1 C.