Enabling 4.6 V LiNi_0.6Co_0.2Mn_0.2O₂ cathodes with excellent structural stability: combining surface LiLaO₂ self-assembly and subsurface La-pillar engineering

INTRODUCTION

Despite the rapid development of lithium-ion batteries (LIBs) in recent decades, their finite energy and power densities are regarded as the main barriers to their further application^[1-4]. The cathode is the most critical component and is crucial in determining the working voltage, energy density and cost of a LIB^[5-9]. Various promising cathode candidates with high capacity, including LiCoO₂ and Ni-rich materials, have been extensively designed for LIBs^[10,11]. Among the available candidates, nickel-rich layered cathode materials, namely, LiNi_xCo_yMn_1-x-yO₂ (NCM, x ≥ 0.6), have demonstrated excellent market prospects owing to their advantages of high specific capacities, large discharge plateaus, low price and hypotoxicity^[12-14]. Although the specific capacity of NCM materials increases significantly with the nickel content, the resulting cation mixing and interfacial side reactions lead to a rapid performance decay upon cycling. In addition, the NCM materials with more than a 70% nickel content manifest apparent safety concerns, thereby hindering their large-scale application^[15-19]. The increasing cutoff voltage of NCM can also increase its specific capacity. For instance, the discharge capacity of LiNi_0.6Co_0.2Mn_0.2O₂ (NCM622) can be enhanced from 180 to 200 mAh g^-1 with an increase in the cutoff voltage from 4.4 to 4.6 V. However, at 4.6 V, the surface side reactions and unamiable phase transitions increase in NCM significantly, resulting in capacity fading^[20-24]. Furthermore, the surface lattice structure of NCM622 can suffer from the uncontrollable cation mixing, leading to an irreversible structural transformation and degradation.

Many approaches, including elemental doping and surface coating, have been adopted to overcome the aforementioned issues and enhance the structural stability of Ni-rich cathodes. Lattice doping using a variety of extrinsic ions, such as Ti^[25], Na^[26], Mg^[27,28], Zr^[29], Nb^[30], Al^[31] and La^[32], has been already used to plausibly improve the electrochemical performance of NCM materials. Inorganic oxides, such as MgO^[33], Al₂O₃^[34], AlF₃^[35], Li₃PO₄^[36], SiO₂^[37], In₂O₃^[38], TiO₂^[39], ZrO₂^[40] and Li₂MnO₃^[41], have been coated on the surface of NCM materials to alleviate the layer-to-rock salt phase transformation and improve the cycling stability. Furthermore, strategies of ion doping combined with surface modification have also been used to improve the electrochemical performance of NCM. Yang et al. used Ti doping and a La₄NiLiO₈ coating to significantly restrain the cation mixing of NCM811 and improve its capacity retention after 200 cycles^[42]. Ming et al. employed Zr and Al doping of LiNi_0.83Co_0.12Mn_0.05O₂, which resulted in a remarkable enhancement in capacity and voltage retention^[43]. However, the above studies focus mostly on the electrochemical performance of NCM materials in a lower potential range of 3.0-4.3 V. At an elevated voltage of 4.6 V, few reliable experimental results are available, and hence the underlying reaction mechanism concerning the capacity retention rate of NCM materials in this broader potential range is poorly understood^[44-47].

La is a relatively inexpensive rare-earth material that is widely used in industry. The La-O bond has an energy of 798 kJ mol^-1, much higher than that for Ni-O (392 kJ mol^-1), Co-O (368 kJ mol^-1) and Mn-O (402 kJ mol^-1). Moreover, La-based oxides have excellent conductivity and thermal stability. As surface coating materials, it is reported that La-based oxides can improve the conductivity of NCM, inhibit the side reactions and alleviate the layer-to-rock salt transformation at high voltages^[48-50]. On this basis, in this work, we design a dual modification strategy to synthesize a NCM622 cathode with surface LiLaO₂ self-assembly in conjunction with subsurface La pillaring, as shown in Scheme 1. The LiLaO₂ coating on the surface of NCM622 is beneficial for inhibiting the side reactions between the cathode and electrolyte, thereby relieving the phase transformation from the layered to rock salt phase and improving the capacity retention at high voltages. The La dopant, as a pillar in the NCM622 lattice, increases the c lattice parameter to improve the Li-ion diffusion capability and suppress Ni antisite defect formation. Density functional theory (DFT) calculations confirm that the Li-ion diffusion and electronic conductivity are greatly enhanced in the dual-modified NCM622 compared with the pristine material. As a result, the dual-modified NCM622 exhibits excellent cycling stability at 4.6 V (over 200 mAh g^-1 at 0.1 C, 188 mAh g^-1 at 1 C and a 76% capacity retention during 200 cycles), which is much better than the pristine material (45% capacity retention after 200 cycles).

Scheme 1. Synthetic process of LiNi_0.6Co_0.2Mn_0.2O₂ (NCM622) cathode with and without modification.

EXPERIMENTAL

Materials preparation. The Ni_0.6Co_0.2Mn_0.2(OH)₂ precursor was synthesized using a hydroxide coprecipitation method. The details of the coprecipitation process are displayed in the supporting information. NCM622 samples were prepared by mixing Ni_0.6Co_0.2Mn_0.2(OH)₂ precursors and Li₂CO₃ (Li:Ni ratio of 1:0.56). After calcined at 870 °C for 10 h, the samples were cooled to 570 °C and then sintered at 570 °C for 5 h in an oxygen atmosphere. NCM622-La samples were prepared by homogeneously mixing Ni_0.6Co_0.2Mn_0.2(OH)₂ precursors, Li₂CO₃ and La₂O₃ (Li:Ni:La ratio of 1:0.56:0.003) and then sintered according to the above process. The synthesized NCM622-La@LLO was a mixture of NCM622-La, Li₂CO₃ and La₂O₃ with the molar ratio of Ni: Li: La (0.6:0.003:0.003) in an agate mortar and then calcined at 870 °C for 10 h in an oxygen atmosphere.

Materials characterization. The structure and morphology of the samples were evaluated by X-ray diffraction (XRD, Rigaku DMAX 2500, Tokyo, Japan) with Cu Kα radiation and scanning electron microscopy (SEM, Hitachi S-4800, Tokyo, Japan), respectively. The lattice parameters of the samples were refined by Full Prof Suite software. High-resolution transmission electron microscopy (HR-TEM, Tecnai G2 F30) was conducted to further understand the microstructure of the samples. The chemical states of the main elements of the materials were analyzed by X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250Xi spectrometer).

Electrochemical measurements. The cell in a coin configuration (CR2016) assembled with the cathode, separator, lithium plate and electrolyte in an argon-filled glove box was used to investigate the electrochemical properties. The cathode was fabricated by coating the slurry composed of the active material, carbon black and binder at a ratio of 80:10:10 wt.% on Al foil. The electrode mass loadings of NCM622, NCM622-La and NCM622-La@LLO were 1.20, 1.18 and 1.22 mg/cm², respectively. Li metal was used as the counter and reference electrodes. Celgard 2400 was used as the separator. The electrolyte was 1 MLiPF₆ in ethylene carbonate/dimethyl carbonate (1:1 by volume). The as-prepared cells were used for electrochemical measurements with Neware software. Cyclic voltammetry (CV) and AC impedance spectroscopy (0.01-105 Hz) were carried out using a multichannel potentiostat/galvanostat from Princeton Applied Research. The galvanostatic intermittent titration technique was conducted to evaluate the kinetic properties of the electrodes.

RESULTS AND DISCUSSION

The crystal structures of NCM622, La-doped NCM622 (NCM622-La) and La-doped and LiLaO₂-coated NCM622 (NCM622-La@LLO) were studied by XRD, as shown in Figure 1A-C. The diffraction peaks of the (003), (101), (104), (105), (107) and (113) planes are observed [Figure 1A] and indexed to the hexagonal Rm structure (PDF#09-0063), indicating that the doping and coating processes have no obvious influence on the crystal structure of NCM622. The separation of the (006)/(102) and (108)/(110) peaks illustrates a highly crystallized layered structure for the three samples^[51]. The apparent left shift of the (104) peak of NCM622-La indicates that La successfully entered the NCM622 lattice [Figure 1C]. A similar phenomenon occurs on both the (104) and (003) peaks of NCM622-La@LLO [Figure 1B and C]. Additionally, the interplanar spacings of (003) and (104) widen, owing to the introduction of La, which is plausibly conducive to facilitating the ionic diffusion kinetics of Li ions^[52,53]. Moreover, the weak diffraction peaks of (020), (111) and (220) observed in NCM622-La@LLO are attributed to the LiLaO₂ surface coating layer [Supplementary Figure 1].

Figure 1. (A) XRD patterns of NCM622, NCM622-La and NCM622-La@LLO. (B and C) Comparison of (003) and (104) peaks for the three samples. (D-F) XPS comparison of Ni element for the three samples. (G-I) SEM morphology images of NCM622, NCM622-La and NCM622-La@LLO and (J) the corresponding energy-dispersive spectroscopy mapping results of NCM622-La@LLO. TEM images of (K and L) NCM622 and (M and N) NCM622-La@LLO.

XPS was used to evaluate the surface composition and oxidation states of the samples. As illustrated in Figure 1D-F, the peaks of Ni-2p_3/2 at 855.60 and 854.53 eV represent Ni³⁺ and Ni²⁺, respectively. Apparently, compared to that in the pristine sample, the decrease in the Ni²⁺ content in both NCM622-La and NCM622-La@LLO is mainly due to the introduction of La³⁺, whose electrostatic repulsion prevents more Ni²⁺ from migrating to the Li site, thereby inhibiting cation mixing^[54,55]. The XPS analysis comparison of La in the three samples indicates that the La particles exist in both the NCM622-La and NCM622-La@LLO electrodes [Supplementary Figure 2]. The Rietveld refinement results further confirm that the Ni-Li cation mixing decreases after the La doping and LiLaO₂ coating [Supplementary Figure 3 and Supplementary Table 1].

The sizes of the NCM622-La@LLO particles are ~0.5 μm, smaller than that of the pristine NCM622 particles (~1 μm), as shown by the SEM images in Figure 1G-I. The introduction of La in the lithiation of the hydroxide precursor changes the surface energy, which changes the fusion of nanosheets during high-temperature calcination^[56]. Interestingly, compared with the modified NCM622, many small particles appear on the surface of the original sample, which are residual lithium (Li₂CO₃/LiOH)^[38]. The surface of the NCM622-La@LLO particles displays a blurred edge compared with the pristine NCM622 particles, preliminarily proving the LiLaO₂ formed on the surface of NCM622. Moreover, the integral morphologies of the three samples show no significant differences, indicating that the overall morphology of the host material is free from the destruction of the doping and coating processes [Supplementary Figure 4]. The energy-dispersive spectroscopy mapping images of NCM622-La@LLO [Figure 1J and Supplementary Figure 5] show a uniform distribution of La, indicating that LiLaO₂ is evenly coated on the surface of NCM622. Additionally, the HR-TEM images show that an obvious Ni-O phase exists on the pristine NCM622 electrode surface [Figure 1K and L] and the NCM622-La@LLO surface has a clear coating layer [Figure 1M]. The lattice fringes of the NCM622-La@LLO electrode vary significantly from the surface to the interior. The lattice spacing of the coating layer is 0.28 nm [Figure 1N], corresponding to the diffraction peak of the (220) plane of LiLaO₂, further confirming that the surface of NCM622-La@LLO is covered with a thin LiLaO₂ layer.

DFT calculations were used to estimate the surface structure of NCM622 before and after modification with La doping and LiLaO₂ coating, as shown in Figure 2A-I. Using the same structure, the calculated density of states (DOS) of NCM622-La@LLO is much larger than that of NCM622 and NCM622-La at the Fermi level [Figure 2D-F and Supplementary Figure 6], indicating that the electronic conductivity of NCM622-La@LLO is improved^[57]. Figure 2G-I and Supplementary Figures 7 and 8 exhibit the charge density differences between the three samples. An evident electron transfer phenomenon occurs among Ni, Co, La and the surrounding O atoms, and becomes clearer after La doping and LiLaO₂ coating. In addition, the La-O has a high electron aggregation degree compared to the Ni-O, suggesting that La-O has a stronger chemical bond, which will be beneficial to stabilize lattice oxygen during cycling, especially at a high cutoff voltage. Stable lattice oxygen can inhibit the layer-to-spinel/rock salt phase transformation and improve the electrochemical performance.

Figure 2. (A-C) Structural diagrams, (D-F) DOS and (G-I) differential charge densities of the three samples.

The ionic and electronic conductivities of NCM622, NCM622-La and NCM622-La@LLO measured with a blocking electrode method show that NCM622-La@LLO has the highest values [Figure 3A and B]. The Nyquist plots of the half-cells also reveal the best lithium-ion transport capability of NCM622-La@LLO [Figure 3C]. To understand the dynamic behavior of the three samples, the galvanostatic intermittent titration technique was used at a pulsed current of 20 mA/g [Supplementary Figure 9]. The lithium-ion diffusion coefficient in NCM622-La@LLO is estimated to be ~1.4 × 10^-10 cm² s^-1, higher than that in NCM622 (5.1 × 10^-12 cm² s^-1) and NCM622-La (4.0 × 10^-11 cm² s^-1), suggesting that the La doping and LiLaO₂ coating boost the lithium-ion diffusion kinetics of NCM622. CV was carried out at a scan rate of 0.1 mV s^-1 within the potential range of 3.0-4.6 V [Supplementary Figure 10]. The potential difference (ΔEp) values of the three samples between the oxidation/reduction peaks are 286 mV (NCM622), 187 mV (NCM622-La) and 166 mV (NCM622-La@LLO), respectively. The minimum positional deviation of the oxidation/deoxidization peaks indicates that NCM622-La@LLO exhibits the best electrochemical activity and cycle reversibility. The CV curves of the three samples at different potential scan rates from 0.2 to 1.0 mV s^-1 are shown in Figure 3D-F. Both the oxidation and reduction peak currents of the three samples gradually increase with the scanning rate and shift toward low and high potentials, respectively. The fitting result indicates that the diffusion-controlled electrochemical process dominates the charge/discharge reaction in electrodes^[58-60]. The b values of the cathode peaks of NCM622, NCM622-La and NCM622-La@LLO are 0.76, 0.80 and 0.83, respectively, indicating that the NCM622-La@LLO has more significant surface capacitive effects [Figure 3G]. Figure 3H distinguishes the percentage of the capacitive contribution and diffusion-controlled contribution for the three electrodes at different scan rates. Although the capacitive contribution increases with scan rate, the surface capacitive contribution of NCM622-La@LLO dominates, meaning that the capacitance control process plays an important role in Li-ion storage at higher scan rates^[61].

Figure 3. (A) Ionic and (B) electronic conductivity of NCM622, NCM622-La and NCM622-La@LLO samples. (C) Nyquist plots of the three electrodes. (D-F) CV curves of the three samples at different potential scanning rates. (G) Relationship between logarithmic anode peak current and logarithmic scan rates of the three electrodes. (H) Capacitance contribution calculations of the three electrodes at various scan rates.

The first-cycle discharge/charge capacities of NCM622-La@LLO are 237.2/202.2 mAh g^-1, with a corresponding initial Coulombic efficiency (ICE) of 85.2% [Figure 4A]. There is little difference among the ICEs compared with NCM622 (84.7%) and NCM622-La (86.8%) [Supplementary Figure 11], indicating that La doping and LiLaO₂ coating have little effect on the ICE of NCM622 materials. Figure 4B further compares the electrochemical cycling performances of the three samples after various cycles at 1 C. For NCM622-La@LLO, excellent capacity retention can be achieved at 76.1% after 200 cycles. In contrast, pristine NCM622 undergoes rapid capacity decay, resulting in a lower capacity retention of only 32.7% after 200 cycles. Furthermore, NCM622-La@LLO shows a good charge and discharge plateau above 3.6 V after 200 cycles [Figure 4C-E]. Figure 4F-H provides a detailed analysis of the evolution process of the differential capacitance (dQ/dV) curves of the three samples. The dQ/dV curves display two redox peaks during cycling. However, for the NCM622 sample, the intensity of the redox peak becomes weaker than that of the NCM622-La and NCM622-La@LLO samples when the cycle number increases, indicating that La-pillared and LiLaO₂-coated NCM622 exhibits a more stable structure, a stronger ability to withstand high voltages and better electrochemical performance.

Figure 4. (A) Charge and discharge curves of the first cycles of NCM, NCM-La and NCM622-La@LLO at a current density of 20 mA g^-1 (0.1 C) and (B) corresponding cycling performance at 1 C in the voltage range of 3.0-4.6 V. Charge and discharge curves of different cycles and corresponding dQ/dV curves of (C and F) NCM622, (D and G) NCM622-La and (E and H) NCM622-La@LLO.

The crystal structures and surface morphologies of the samples after 200 cycles were investigated by TEM and SEM, respectively. The TEM images of NCM-La@LLO show that although no rock-salt phase is observed on the NCM-La@LLO surface, it conserves a well-ordered layered structure [Figure 5B]. In contrast, the rock-salt phase exists on the surface region of NCM622 and the spinel phase is detected from the diffraction spots, as shown in Figure 5A. The NCM622 exhibits a salt-spinel-layered structure from the surface to the inner phase, but the NCM-La@LLO exhibits a slight mixture of the layered and spinel phases on the surface. This obvious evidence indicates that La doping and LiLaO₂ coating significantly inhibit the transformation of the layered phase to the rock-salt phase and maintain the structural stability of the NCM material, thereby improving the reversible specific capacity and cycle life. The surface morphologies of NCM622-La@LLO after 200 cycles are shown in Figure 5E. Compared with the NCM622 and NCM622-La particles [Figure 5C and D], no cracks exist on the surface of NCM622-La@LLO, indicating that La doping and LiLaO₂ coating significantly inhibit the phase transition to maintain the structural stability of NCM at high voltages, thus improving the electrochemical performance.

Figure 5. TEM images of (A) NCM622 and (B) NCM622-La@LLO after 200 cycles, where I^a, II^a, III^a, I^b, II^b and III^b are the fast Fourier transform conversion graphs of the corresponding regions. Surface morphologies of (C) NCM622, (D) NCM622-La and (E) NCM622-La@LLO after 200 cycles. (F) C1s, (G) O1s and (H) F1s XPS spectra of NCM622, NCM622-La and NCM622-La@LLO after 200 cycles.

To further verify the protective effect of the coating layers on the cathode-electrolyte interface, XPS measurements on NCM622, NCM622-La and NCM622-La@LLO after 200 cycles were conducted [Figure 5F-H]. The peaks of the C-H, C-C, C-O and OCO₂ bonds can be observed in the C 1s spectra of the electrodes [Figure 5F]. The binders and conductive substances in the electrode can cause the existence of C-H and C-C bonds, while the C-O and OCO₂ bonds are related to the decomposition of the electrolyte^[62]. Compared to NCM622, weaker peak intensities are observed for the peaks associated with the C-O and OCO₂ of the NCM622-La@LLO cathode, suggesting a lower electrolyte decomposition amount on the surface. In addition, the typical peaks of M-O, Li₂CO₃ and ROCO₂Li can also be observed in the O 1s spectra [Figure 5G]. The ROCO₂Li peak of the NCM622-La@LLO electrode is smaller, meaning that the LiLaO₂ coating layer formed in NCM@LLO can inhibit the decomposition of the electrolyte and the formation of SEI film on the particle surface during electrochemical cycling. Additionally, the LiF peak is significantly weakened in NCM622-La and almost invisible in NCM622-La@LLO, indicating that the La doping and LiLaO₂ coating significantly inhibit the surface side reactions and are helpful for the structural stability of NCM materials.

Cylindrical full batteries with a design capacity of 2500 mAh were assembled to compare the cycling stabilities of NCM622 and NCM622-La@LLO in real applications. Schematic diagrams of the cylindrical full batteries are shown in Figure 6A and B. The NCM622-La@LLO cell undergoes 1000 cycles with an excellent capacity retention of 82.5% [Figure 6C]. The overlap of the dQ/dV curves of the NCM622-La@LLO cathode demonstrates that La doping and LiLaO₂ coating can maintain the structural integrity of NCM materials [Figure 6D and E]. In addition, the charge/discharge curves from the 1st to 1000th cycle also suggest that the NCM622-La@LLO cathode has a higher reversible capacity due to the less electrochemical polarization compared with the NCM622 cathode [Figure 6F and G], which is mainly caused by the enhanced stability of NCM622 after La doping and LiLaO₂ coating. During long cycling [Figure 6H], the average voltage of the NCM622-La@LLO cell is stable at ~3.6 V, further proving the cyclic stability of the NCM622-La@LLO cathode. The excellent electrochemical performance of the NCM622-La@LLO cathode shows its excellent prospects for industrial applications.

Figure 6. Electrochemical characterization of cylindrical full batteries. (A) Schematic and (B) photographs of the configuration of a graphite//NCM622-La@LLO full battery. (C) Cycling performance of full batteries with both pristine NCM622 and NCM622-La@LLO as the cathode and graphite as the anode at 2 C. (D and E) dQ dV^-1 curves at selected cycles. (F and G) Corresponding charge/discharge curves. (H) Comparison of voltage platform attenuation.

CONCLUSIONS

In summary, we present a practical solution to enhance the stability of NCM622 at 4.6 V via the self-assembly of a LiLaO₂ coating on its surface and La pillars in its subsurface. The dual-modified NCM622-La@LLO cathode exhibits a capacity of more than 200 mAh g^-1 and an initial Coulombic efficiency of 85% at a current density of 0.1 C. Even after 200 cycles, NCM622-La@LLO maintains a capacity retention of 1.7 times higher than the pristine material (76% vs. 45%) at 1 C and 4.6 V. Two factors are responsible for the superior electrochemical performance of the NCM622-La@LLO cathode. The first is that the LiLaO₂ coating on the surface of NCM622 is beneficial for inhibiting the side reactions between cathodes and electrolytes and suppressing the phase transformation from the layered phase to the rock-salt phase, thereby improving the capacity retention at high voltage. The other reason is that the La atom dopants, as pillars in the NCM622 lattice, increase the c-axis distance to improve the Li⁺ diffusion rate, and suppress nickel taking the place of lithium. We expect this strategy could provide a direction for managing the internal structure and interfacial stability of NCM, which can be extended to the applications of other cathodes, such as LiMn_2xNi_xO₄ spinel and layered Li-rich oxides.