Titanium-Doped Nickel-Rich Layered LiNi_0.8Co_0.2O₂ as High-Performance Cathode Material for Lithium-Ion Batteries

Abstract

LiNi_0.8Co_0.2−xTi_xO₂ cathode materials with different titanium (Ti) contents have been prepared from Ni_0.8Co_0.2−xTi_x(OH)₂ (0 ≤ x ≤ 0.05) precursors. Phase and morphology analyses indicated that the products were microspheres in space group R$\overline{\textit{3}}$m assembled from smaller secondary particles. Phase and elemental analyses indicated more serious cation mixing in the layered materials obtained with Ti doping. The LiNi_0.8Co_0.2−xTi_xO₂ material with x = 0.0125 showed optimal electrochemical performance with improved cycling performance (capacity retention ratio of 95.8% after 100 cycles) and high rate capability (162 mAh g⁻¹ at 2C). Therefore, Ti doping can be considered an efficient approach to improve the electrochemical properties of such cathode materials.

Introduction

Lithium-ion batteries (LIBs) have been widely used as electrochemical energy storage devices in many fields including portable electronics, electric vehicles, and energy storage systems because of their outstanding advantages such as light weight, long cycle life, and low cost.^1,2 However, the relatively low discharge capacity of traditional commercialized LIB cathode materials such as olivine LiFePO₄, layered LiCoO₂, and spinel LiMn₂O₄ has greatly limited their industrial application and relevant device performance.³ Compared with traditional LiCoO₂ material, the layered solid solution LiNi_xCo_1-xO₂ (0 < x < 1) has been reported to be one of the potential alternative candidates due to its higher capacity and lower cost.^4,5 The utilization of lithium ions in the intercalation and deintercalation processes is higher in LiNi_xCo_1−xO₂ materials, especially LiNi_0.8Co_0.2O₂, than in LiCoO₂ material.^6,7

Many different methods have been reported for synthesis of LiNi_0.8Co_0.2O₂ microstructures, including room-temperature growth,⁸ the sol–gel method,⁹ emulsion drying,¹⁰ coprecipitation,¹¹ and the hydrothermal approach.¹² Particle size control, precursor selection, and the thin-film technique have significant effects on the electrochemistry of LiNi_0.8Co_0.2O₂ cathodes.^3,13,14,15 Nevertheless, the electrochemical stability of these Ni-rich layered oxides remains poor due to surface degradation, gas release, and cation mixing of LiNi_0.8Co_0.2O₂ structures during electrochemical cycling.^16,17,18 To solve these problems, surface modification with MgO and coating with Li₂ZrO₃ on LiNi_0.8Co_0.2O₂ have been reported to improve the structural stability.^19,20 Metal doping has also been studied for LiNi_0.8Co_0.2O₂ modification; For example, Nb⁵⁺ doping at Li⁺ sites in the LiNi_0.8Co_0.2O₂ crystal lattice resulted in an acceleration of lithium-ion migration and an enhancement of the high-rate performance.²¹ Doping of size-invariant Zn²⁺ into LiNi_0.8Co_0.2O₂ structures decreased the polarization and improved the reversibility of the redox processes during cycling.²²

Titanium (Ti) doping has been considered as an efficient approach to improve the electrochemical properties of cathode materials including Li₃V₂(PO₄)₃,²³ Li₂FeSiO₄,²⁴ Li₃Fe₂(PO₄)₃,²⁵ and layered metal oxide. Ti doping creates cation vacancies in the crystal structure that can improve the electronic conductivity and ionic diffusion of cathode materials.²⁶ The electrochemical inactivity of Ti⁴⁺ increased the structural stability of the cathode and inhibited crack formation in the crystal structure during electrochemical cycling.^24,27 For LiNi_0.80Co_0.15Al_0.05O₂ cathodes, Ti substitution at Co sites can suppress the phase transformation during charging.^28,29 For LiNi_0.8Co_0.1Mn_0.1O₂ cathodes, Ti doping reduced the cation mixing between Li⁺ and Ni²⁺ through the pillar effect, which favors capacity retention.^30,31 For LiNi_0.8Co_0.2O₂ cathodes, Liu’s group³² and Zhang’s group³³ reported Ti doping of LiNi_0.8Co_0.2O₂ by using Ni_xCo_y(OH)_z as the precursor. The Ti doping process was carried out through calcination of Ni_xCo_y(OH)_z in presence of LiOH and Ti precursor. However, Ti species mainly appeared on the particle surfaces. Uniform distribution of Ti element in LiNi_0.8Co_0.2O₂ particles cannot be ensured but is important to improve the structural stability and ion diffusion in the LiNi_0.8Co_0.2O₂ particles.

In the work described herein, Ni_0.8Co_0.2−xTi_x(OH)₂ was prepared by the coprecipitation method and served as the precursor for LiNi_0.8Co_0.2−xTi_xO₂ (0 < x ≤ 0.05) cathode materials. The crystallinity and surface morphology of Ni-enriched materials can be significantly affected by Ti doping, as indicated by x-ray diffraction (XRD) analysis and scanning electron microscopy (SEM) analysis, respectively. Electrochemical testing of LiNi_0.8Co_0.2−xTi_xO₂ (0 < x ≤ 0.05) was carried out to study the influence of the Ti doping on the cycling stability and rate performance of the LiNi_0.8Co_0.2O₂ cathode. The electrochemical reversibility and ionic conductivity of the LiNi_0.8Co_0.2−xTi_xO₂ cathodes were studied by using cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) tests.

Experimental Procedures

Synthesis

Ni_0.8Co_0.2−xTi_x(OH)₂ precursors were obtained by the coprecipitation method. Firstly, NiSO₄·6H₂O, CoSO₄·7H₂O, and TiOSO₄·8H₂O were dissolved in water together in argon with the total concentration of Ni, Co, and Ti in the polymetallic solution fixed at 2 mol L⁻¹. The molar ratio of Ni and Co to Ti was 0.8:(0.2−x):x (x = 0, 0.0125, 0.025, and 0.05). NaOH solution (4 mol L⁻¹) and ammonia (28 wt%) were mixed at volume ratio of 10:1 as the alkaline solution. Then, the pH of the polymetallic solution was adjusted to 11.3 ± 0.05 using a water bath at 55°C by adding the alkaline solution dropwise in argon. After aging for 12 h, the obtained Ni_0.8Co_0.2−xTi_x(OH)₂ precipitate was filtered, washed with deionized water repeatedly, and dried at 100°C for 24 h. Thereafter, the Ni_0.8Co_0.2−xTi_x(OH)₂ precipitate was sintered together with LiOH·H₂O at molar ratio of 1:1.1 at 550°C in air for 5 h then calcined at 750°C in oxygen for 12 h. Finally, LiNi_0.8Co_0.2–xTi_xO₂ (x = 0, 0.0125, 0.025, and 0.05) cathode materials were obtained after grinding and screening. The samples with x values of 0, 0.0125, 0.025, and 0.05 are denoted as LNCTO-0, LNCTO-1.25, LNCTO-2.5, and LNCTO-5, respectively.

Characterization

The crystalline structure of the prepared materials was characterized by powder x-ray diffraction (XRD) analysis (PANalytical X’Pert Powder) measurement using Cu K_ɑ radiation in the 2θ range from 10° to 90° with step size of 0.02°. Rietveld refinement was carried out by using the TOPAS 5 software package to obtain the lattice parameters. Morphological studies of the prepared samples were carried out by scanning electron microscopy (SEM, JEOL, JSM-7800F). Energy-dispersive spectroscopy (EDS) was employed in SEM to investigate the distribution of the elements. x-Ray photoelectron spectroscopy (XPS, ESCALab250) was employed to determine the elemental distribution of the materials. The actual content of Ti, Ni, and Co elements in each product was determined by inductively coupled plasma optical emission spectroscopy (AGILENT ICP-OES 730).

Electrochemical Measurements

LiNi_0.8Co_0.2–xTi_xO₂ powders were mixed with polyvinylidene fluoride as binder and super P as conductor at mass ratio of 90%:5%:5% in N-methyl-2-pyrrolidone (NMP). The mixed slurry was coated uniformly on aluminum foil and dried at 120°C for 10 h in vacuum. The electrochemical performance of LiNi_0.8Co_0.2–xTi_xO₂ was evaluated by using CR2032 coin cells with lithium metal as reference electrode and Celgard 2300 as separator. LiPF₆ (1 M) in ethylene carbonate (EC)/ethylmethyl carbonate (EMC)/dimethyl carbonate (DMC) (1:1:1 v/v/v) was used as electrolyte. Electrochemical measurements were performed between 2.5 V and 4.3 V (versus Li/Li⁺) using a Neware test instrument. Cyclic voltammetry (CV) measurements were carried on an electrochemical workstation (CHI660C, Shanghai Chenhua) at scan rate of 0.1 mV s⁻¹ between 3.0 V and 4.6 V (versus Li/Li⁺). Electrochemical impedance spectroscopy (EIS) tests were performed at full charged state after cycling in the frequency range of 0.001 Hz to 100 kHz by applying an alternating-current (AC) amplitude of 5 mV on a CHI660C electrochemical workstation.

Results and Discussion

The XRD patterns of four LiNi_0.8Co_0.2–xTi_xO₂ samples are shown in Fig. 1a. All the diffraction peaks of the four samples can be indexed to the layered α-NaFeO₂ structure in space group R$\overline{\textit{3}}$m (Fig. 1a).^34,35 No peaks corresponding to TiO₂ or other impurity phases were observed, indicating the high purity of the obtained α-NaFeO₂ phase in the four samples. Two portions of the XRD patterns marked by dotted boxes in Fig. 1a are shown enlarged in Fig. 1b, c. Clear splitting of both (006)/(012) and (108)/(110) doublets can be observed in Fig. 1b, c for all four samples, indicating the well-developed structural ordering of the obtained Ni-rich layered materials.^36,37

Fig. 1

(a) XRD patterns of LiNi_0.8Co_0.2−xTi_xO₂ samples and partial enlargements of (b) (006)/(102) and (c) (108)/(110) doublets.

To investigate the influence of Ti doping on the crystal structure of LiNi_0.8Co_0.2O₂, the lattice constants and other structural parameters were calculated and are summarized in Fig. 2 and Table I. The high c/a value of LiNi_0.8Co_0.2O₂ (> 4.94) indicates its layered structure. Moreover, compared with LiNi_0.8Co_0.2O₂, the increase of the c/a value found for the three LiNi_0.8Co_0.2−xTi_xO₂ samples indicates an improvement of the layered crystal structure upon Ti doping of LiNi_0.8Co_0.2O₂.^38,39 The I(003)/I(104) values of the samples decreased with increasing Ti doping amount, demonstrating the more serious Li⁺/ Ni²⁺ cation mixing in layered materials after Ti doping.^40,41 In addition, the cell volumes were slightly enlarged upon Ti doping of LiNi_0.8Co_0.2O₂, leading to enlarged Li-ion transport channels and improved lithium diffusion in Ti-doped crystal structures.³¹ However, excessive Ti in the LiNi_0.8Co_0.2−xTi_xO₂ materials may result in a reduction of the reversible capacity because of the electrochemical inertness of Ti⁴⁺ ions.^42,43

Fig. 2

Rietveld refinement results of XRD data for (a) LNCTO-0, (b) LNCTO-1.25, (c) LNCTO-2.5, and (d) LNCTO-5.

Table I Cell parameters and composition of LiNi_0.8Co_0.2–xTi_xO₂ samples

The morphology and microstructure of the LiNi_0.8Co_0.2−xTi_xO₂ (0 ≤ x ≤ 0.05) samples were observed by SEM; typical images are shown in Fig. 3. All four samples were composed of microspheres with size from several to more than ten micrometers. All the microspheres were assembled from smaller secondary particles, and some spheres were connected together. Many pores could be observed among adjacent secondary particles, providing space for sufficient contact with the electrolyte.^32,40 When the x values was changed from 0 to 0.025, the size of the secondary particles could be controlled in the range from 0.2 μm to 1.5 μm (Fig. 3a–f). However, when x reached 0.05, the secondary particles were larger than 3 μm (Fig. 3g, h). This indicates that the morphology of LiNi_0.8Co_0.2−xTi_xO₂ with low Ti doping amounts (≤ 0.025) may be similar to that of LiNi_0.8Co_0.2O₂, while excess Ti in the crystal (≥ 0.05) had a significant effect on the structure of LiNi_0.8Co_0.2−xTi_xO₂.

Fig. 3

SEM images of LiNi_0.8Co_0.2−xTi_xO₂ cathode materials with (a, b) x = 0, (c, d) x = 0.0125, (e, f) x = 0.025, and (g, h) x = 0.05.

Taking LNCTO-1.25 as an example, EDX elemental analysis was carried out to study the distribution of O, Ni, Co, and Ti elements in the square region of the Ti-doped particle (Fig. 4). The mapping result for Ti element completely overlapped with those of O, Ni, and Co elements, indicating a homogeneous distribution of Ti element in the Ti-doped LiNi_0.8Co_0.2O₂ samples.

Fig. 4

EDX elemental mapping of O, Ni, Co, and Ti for LNCTO-1.25.

To compare the surface elemental composition and valence states of the LiNi_0.8Co_0.2−xTi_xO₂ samples, LNCTO-0 and LNCTO-1.25 were chosen for XPS tests; their corresponding spectra are shown in Fig. 5. Both XPS results show characteristic peaks near 55 eV, indicating presence of Li element in the two samples (Fig. 5a). However, only LNCTO-1.25 showed peaks at 457.3 eV and 463.0 eV, which can be assigned to Ti 2p_3/2 and Ti 2p_1/2, indicating presence of Ti⁴⁺ in LNCTO-1.25 upon Ti doping (Fig. 5b).⁴⁴ The Ni 2p spectrum contained two peaks at 855.0 eV (Ni 2p_3/2) and 872.67 eV (Ni 2p_1/2), and the splitting of the Ni 2p_3/2 peak can be assigned to Ni³⁺ (855.83 eV) and Ni²⁺ (854.75 eV) (Fig. 5c).⁴⁵ Compared with LNCTO-0, the ratio of Ni²⁺ to Ni³⁺ was significantly increased for LNCTO-1.25, indicating more serious Li⁺/Ni²⁺ mixing upon Ti doping.²⁰ This is consistent with the XRD results, and proper cation mixing can enhance the structural stability of layered cathode materials.⁴⁶ The peaks observed near 861 eV and 880 eV in the Ni 2p spectrum can be attributed to satellites of Ni 2p_3/2 and Ni 2p_1/2, respectively.³¹ The binding energies of 780 eV and 794.9 eV correspond to Co 2p_3/2 and Co 2p_1/2, respectively (Fig. 5d). The splitting of the Co 2p _3/2 peak indicates a mixed valence state of Co²⁺ with Co³⁺ in these two samples.⁴⁷ Compared with LNCTO-0, the ratio of Co²⁺ to Co³⁺ was higher for LNCTO-1.25, which may be because of the charge equilibrium in the presence of Ti⁴⁺.⁴⁸

Fig. 5

XPS spectra of (a) Li 1s, (b) Ti 2p, (c) Ni 2p, and (d) Co 2p.

Using the obtained LiNi_0.8Co_0.2−xTi_xO₂ samples as cathode materials, galvanostatic charging and discharging of 2032-type coin cells was carried out between 2.5 V and 4.3 V at different rates. The resultant charge–discharge curves, cycling performance, and rate capabilities are shown in Fig. 6. Moreover, the corresponding electrochemical results are presented in Table II. Similar curves were observed for all four samples (Fig. 6a–d), indicating the limited influence of Ti doping on the lithium insertion/extraction in such batteries.²⁰ The initial discharge capacities of the LiNi_0.8Co_0.2−xTi_xO₂ samples decreased with increasing Ti doping amount (Table II), which may be because of the electrochemical inertness of the doped Ti⁴⁺.^42,43 However, the discharge capacities of both LNCTO-1.25 (189 mAh g⁻¹ and 182 mAh g⁻¹) and LNCTO-2.5 (177 mAh g⁻¹ and 163 mAh g⁻¹) in the 50th and 100th cycles were larger than those of the LNCTO-0 sample (176 mAh g⁻¹ and 154 mAh g⁻¹), respectively, indicating an improvement in the cycling stability upon doping appropriate amounts of Ti in the samples (Table II; Fig. 6e). Furthermore, the rate capability of both LNCTO-1.25 (162 mAh g⁻¹) and LNCTO-2.5 (160 mAh g⁻¹) was higher than that of LNCTO-0 (156 mAh g⁻¹) at 2C (Table II; Fig. 6f). This enhanced cycling stability and rate performance may be because of the improved lithium diffusion in the crystal structures after the cell volumes is slightly enlarged by doping Ti, as discussed regarding the XRD results.³¹ As shown by the SEM results of the samples, the Ti-doped products displayed morphologies (microspheres) and secondary particle sizes (0.2 μm to 1.5 μm) similar to those of LiNi_0.8Co_0.2O₂ when the doping amount was low (x ≤ 0.025), indicating the negligible effect of Ti doping on the product morphology. However, a significant difference in the electrochemical performance (including the cycling stability and rate performance) was found among the obtained LiNi_0.8Co_0.2−xTi_xO₂ (0 ≤ x ≤ 0.025) samples, suggesting a weak correlation between sample morphology and electrochemical performance. In addition, when more Ti (x = 0.05) was doped into the sample (LNCTO-5), poor cycling stability (capacity retention of 76.5% upon 100 cycles) and low rate capability (141 mAh g⁻¹ at 2C) were observed (Table II), which can be attributed to the severe Li⁺/ Ni²⁺ cation mixing and enlarged secondary particles upon excess Ti entering the crystal lattice, as described regarding the XRD and SEM results.^40,41

Fig. 6

Charge–discharge curves of 1st, 50th, and 100th cycles for LiNi_0.8Co_0.2−xTi_xO₂ samples with (a) x = 0, (b) x = 0.0125, (c) x = 0.025, and (d) x = 0.05 at rate of 0.5 C. (e) Cycling performance at 0.5C and (f) rate capabilities obtained from 0.2C to 2C for the LiNi_0.8Co_0.2−xTi_xO₂ samples.

Table II Electrochemical performance of LiNi_0.8Co_0.2−xTi_xO₂ (0 ≤ x ≤ 5%) samples

CV tests were carried out to investigate the effect of Ti doping on the electrochemical behavior of the LiNi_0.8Co_0.2−xTi_xO₂ samples. Figure 7a–d depicts the first three CV cycles of four different LiNi_0.8Co_0.2−xTi_xO₂ samples between 3.0 V and 4.6 V at 1 mV s⁻¹. As shown in Fig. 7a, compared with the second and third CV cycles of the LiNi_0.8Co_0.2O₂ sample, its first cycle showed an obviously different shape, which is because of the initial stabilization and solid-electrolyte interphase (SEI) formation in the first cycle.^49,50 Similar results were observed for the LiNi_0.8Co_0.2–xTi_xO₂ samples with x = 0.0125 and 0.025. For all four LiNi_0.8Co_0.2−xTi_xO₂ samples (x =0, 0.0125, 0.025, and 0.05), nearly overlapping curves are seen for the second and third cycles, indicating improved interfacial stability and electrochemical reversibility after the initial activation in the first cycle.⁵¹ For LNCTO-0, LNCTO-1.25, and LNCTO-2.5, each of the CV curves exhibited three couples of oxidation and reduction peaks, including one sharp and two weak redox pairs (Fig. 7a-c), corresponding to the phase transitions of hexagonal phase (H1)/monoclinic phase (M), monoclinic phase (M)/hexagonal phase (H2), and hexagonal phase (H2)/hexagonal phase (H3) during lithium deintercalation and intercalation.^18,52 However, only one redox pair was observed when x was equal to 0.05 (Fig. 7d), indicating a reduction of the electrochemical activity with the introduction of excessive Ti due its electrochemical inertness in the crystal structure. The peak positions of the H1/M phase transitions (at 3.9 V to 4.1 V) and the corresponding potential differences (ΔV) are presented in Table III for LNCTO-0, LNCTO-1.25, and LNCTO-2.5. Compared with LNCTO-0, a decrease of ΔV for the LNCTO-1.25 and LNCTO-2.5 samples can be found. Therefore, doping with a proper amount of Ti was helpful to weaken the electrode polarization of LiNi_0.8Co_0.2O₂, which can improve the high-rate performance and cycle stability.^53,54

Fig. 7

CV curves of (a) LNCTO-0, (b) LNCTO-1.25, (c) LNCTO-2.5, and (d) LNCTO-5; (e) EIS results of four samples.

Table III Potential differences obtained from CV curves of LNCTO-0, LNCTO-1.25, and LNCTO-2.5

To enable investigation of their resistance parameters and lithium-ion diffusion, Nyquist plots of the EIS results for the pristine and Ti-doped LiNi_0.8Co_0.2O₂ cathode materials are shown in Fig. 7e. The EIS curves of all four samples showed similar shapes, consisting of three semicircles from high- to intermediate-frequency regions and an inclined line in the low-frequency region. The three semicircles can be assigned to the surface-film resistance, electron transfer impedance, and charge-transfer impedance, respectively.^55,56 The inclined line can be associated with lithium-ion diffusion in the electrode, which is known as Warburg diffusion.⁵⁷ Compared with LNCTO-0, the decreased diameter of the semicircle for LNCTO-1.25, LNCTO-2.5, and LNCTO-5 indicates a reduction of the charge-transfer resistance and an improvement of the lithium-ion migration in the resultant electrode upon Ti doping.^58,59,60 As shown in the XRD results, Ti doping led to an increase of the c/a value for three LiNi_0.8Co_0.2−xTi_xO₂ samples, implying an improvement of the layered crystal structure of the Ti-doped samples. Moreover, the cell volume was slightly enlarged upon Ti doping of LiNi_0.8Co_0.2O₂, leading to enlarged Li-ion transport channels and improved lithium diffusion in the Ti-doped crystal structures. Therefore, the EIS results indicate the decreased resistances of the Ti-doped samples compared with pristine LiNi_0.8Co_0.2O₂. However, the I(003)/I(104) value increased with increasing Ti doping amount, indicating more serious cation mixing with more Ti doping. Cation mixing can block lithium diffusion and increase the resistance of cathode materials. As a result, the resistances of the Ti-doped samples were lowest for the LNCTO-1.25 sample.

Conclusions

Layered nickel-rich LiNi_0.8Co_0.2−xTi_xO₂ (x=0, 0.0125, 0.025, and 0.05) materials with size of 5 μm to 10 μm were successfully synthesized by calcination of Ni_0.8Co_0.2−xTi_x(OH)₂ precursors together with LiOH·H₂O. The reversible capacity of LiNi_0.8Co_0.2−xTi_xO₂ (x = 0.0125) was 182 mAh g⁻¹ at 0.5C after 100 cycles, much higher than that of pristine LiNi_0.8Co_0.2O₂ (154 mAh g⁻¹). Moreover, the discharge capacity of LiNi_0.8Co_0.2−xTi_xO₂ (x = 0.0125) was 200 mAh g⁻¹, 193 mAh g⁻¹, 182 mAh g⁻¹, and 162 mAh g⁻¹ at 0.2C, 0.5C, 1C, and 2C, respectively, also being higher than those for pristine LiNi_0.8Co_0.2O₂ samples. To study the electrochemical behavior of LiNi_0.8Co_0.2−xTi_xO₂, CV and EIS tests were carried out. The results indicated a reduced charge-transfer resistance, improved lithium-ion migration, and weakened electrode polarization of the layered structures upon Ti doping.