One-Spot Facile Synthesis of Single-Crystal LiNi0.5Co0.2Mn0.3O2 Cathode Materials for Li-ion Batteries.

The layered lithium-metal oxides are promising cathode materials for Li-ion batteries. Nevertheless, their widespread applications have been limited by the high cost, complex process, and poor stability resulting from the Ni2+/Li+ mixing. Hence, we have developed a facile one-spot method combining glucose and urea to form a deep eutectic solvent, which could lead to the homogeneous distribution and uniform mixing of transition-metal ions at the atomic level. LiNi0.5Co0.2Mn0.3O2 (NCM523) polyhedron with high homogeneity could be obtained through in situ chelating Ni2+, Co3+, and Mn4+ by the amid groups. The prepared material exhibits a relatively high initial electrochemical property, which is due to the unique single-crystal hierarchical porous nano/microstructure, the polyhedron with exposed active surfaces, and the negligible Ni2+/Li+ mixing level. This one-spot approach could be expanded to manufacture other hybrid transition-metal-based cathode materials for batteries.

Introduction

Li-ion batteries have become one of the most promising power supplies for hybrid electric vehicles (HEVs) and electric vehicles (EVs) in the last few years due to the growing demand for energy.[1−3] LiCoO2 was first synthesized by Goodenough in 1980,[4] and widely used for commercial batteries by Sony Corporation in 1991 due to the high working voltage and excellent rate capacity.[5] However, the disadvantages such as high cost, limited specific capacity, low thermal stability, and unreliable safety have led to shifting the research focus on layered ternary materials such as LiNi1–MnCoO2. The electrochemical properties of LiNi1–MnCoO2 vary with the proportion of nickel, cobalt, and manganese in the structure.[6] High Ni content of the NCM materials contributes to a greater capacity at the cost of a complicated preparation process and security, high Co content improves the processing ability and rate performance, and high Mn content enhances the structural stability but reduces its capacity.[7] Because of the relatively high energy density, low cost, and the balance in the property such as cycling performance, rate capacity, and stability compared with LiCoO2, LiNi0.5Co0.2Mn0.3O2 (NCM523) is regarded as a hopeful candidate for the commercial lithium-ion batteries.[8,9] To acquire higher-energy-density lithium-ion batteries, one simple way is to enhance upper cutoff voltage of the NCM to obtain higher specific capacity of the positive electrode.[10] However, enhancing the upper cutoff voltage generally reduces the battery lifetime by accelerating the negative parasitic reactions between the positive electrode surface and the electrolyte at high voltage. The methods for modifying the positive electrode surface with a core–shell structure and adding electrolyte additives were verified to improve the lifetime of high-potential Li-ion batteries.[11−14] Hence, the combination of surface modification techniques and the electrolyte additive is of great significance to increase the performance of Li-ion batteries with the high upper cutoff voltages. However, the thick and incomplete covered coatings with inactive oxides on NCM materials could observably reduce the rate capacity and energy density of batteries, and it is not easy to search for new and effective high-voltage electrolytes. Therefore, the focus has been drawn back on the material. Many studies have been done to enhance the properties of cathode materials, especially the micron-sized single primary NCM particle with less gas emission and higher capacity at higher voltage and temperature, as reported by Li.[15−17] But the reported synthesis methods were complex and difficult for industrial production. The coprecipitation method usually results in nonuniform distribution of elements and inferior electrochemical performance.[18−20] The main shortcomings of the hydroxide coprecipitation method is the oxidation of Mn2+, which could lead to manganese oxyhydroxide impurity phases.[21] The advantage of the carbonate coprecipitation method is that the oxidation state of the cation is maintained at 2+ but the experimental conditions such as temperature, pH, and mixing speed require to be charily controlled to acquire uniform precipitates, which makes this method complex.[22] The products produced by the sol–gel method usually expressed lower tap and volumetric density. What is worse, it requires time-consuming drying steps.[23] Now, the industrial production methods involve spray drying, granulation, grinding, washing, and other complex processes, which may pollute the environment. Therefore, it is still a great challenge to prepare layered lithium transition-metal oxides with a favorable single-crystal structure in a simple and time-saving way.

Herein, we propose a novel method to synthesize LiNi0.5Co0.2Mn0.3O2 with uniform size distribution and a polyhedron single-crystal structure using glucose and urea in one spot. In brief, glucose is used as the carbon source and urea as the chelating center; nitrate is then added in the glucose and urea to form a deep eutectic solvent; this homogeneous anhydrous solution could lead to the uniform distribution and homogeneous blending of the transition-metal ions at the molecular level. Dehydration of glucose initiates sol–gel reactions of the metal-oxide precursors in a controlled manner and results in homogeneous nucleation of metal-oxide species and uniform distribution of nanocrystals of the same size, phase, and composition, and therefore the formation of well-distributed NCM523 single crystals with uniform size and structures after high-temperature treatment. Compared with the conventional methods, the one-spot approach can produce layered NCM523 with homogeneous size and uniformly dispersed transition-metal ions without further washing, drying, and distillation repeatedly.

Experimental Section

Synthesis of NCM523

All reagents used in this experiment were purchased from commercial manufacturers and used directly with no further purification. Three grams of glucose was mixed with 5 g of urea (mole ratio is 1:5), which was then heated to 120 °C for 0.5 h until a clear, transparent viscous liquid was obtained. Co(NO3)2·6H2O (Aldrich, 99%, 0.3316 g), Ni(NO3)2·6H2O (Sinopharm, 98%, 0.8288 g), and Mn(NO3)2·4H2O (Sinopharm, 97.5%, 0.4290 g) with a stoichiometric amount were dissolved in the glucose and urea solution; the precursor was mixed with Li2CO3 (Aldrich, 99%, 0.2213 g). Here, 5% of Li was excess for compensating for the evaporation of Li at high temperatures. After stirring at 120 °C for about 2 h, the above solution turned into a yellow molten phase, as shown in Figure b; then, the molten system was dried in an oven at 180 °C for more than 3 h until it completely dried. The sponge-like dried materials were obtained, as shown in Figure b, and calcined at 500 °C for 5 h and 900 °C for 10 h at a heating rate of 3 °C/min in air to remove the carbon template and increase the crystallinity. The whole synthesis process is illustrated in Figure a.

Figure 1

(a) Schematic diagram of the synthesis process, (b) thermogravimetric (TG) and (c) Fourier-transform infrared (FT-IR) analyses of the precursor.

Material Characterization

The thermal analysis of the precursor was done using the thermogravimetric method (STA449 F5). The crystalline structure of the powder sample was confirmed by X-ray diffraction (XRD) (Bruker D8 Advance) equipped with Cu Kα at the scanning rate of 2°/min in the 2θ range of 5–90°. The Fourier-transformed infrared spectrum was obtained using Bruker Tensor II in a transmission mode. The composition of the as-prepared NCM523 powder was confirmed by the inductively coupled plasma mass spectrometer (ICP-MS, Agilent 7700). The elements’ electronic state was measured by X-ray photoelectron spectroscopy (XPS) (Thermo ESCALAB 250XI). The microstructure, composition, and morphology of the powder sample were observed by scanning electron microscopy (SEM) (JSM-5510L), energy-dispersive X-ray spectroscopy (EDS), and transmission electron microscopy (TEM, Titan G260-300).

Electrochemical Properties

The electrodes were prepared by the following steps. The synthesized NCM523 was mixed with poly(vinylidene fluoride) (PVDF) and carbon black (super P) in 1-methyl-2-pyrrolidinone (NMP, Macklin, 99.5%) at a mass ratio of 8:1:1. Then, the slurry was coated on the aluminum foils using a glass rod and dried in a vacuum oven at 100 °C for 12 h. Button cells (CR2032) were assembled with the NCM523 electrode, Li counter electrode, a separator (2325; Celgard, Tokyo, Japan), and a 1 M solution of LiPF6 in a mixture of ethyl carbonate/dimethyl carbonate (EC/DEC = 1:1 vol %) in a glovebox filled with argon gas. The mass loading and the thickness of active cathode materials for all cells were controlled to 2.4 mg/cm2 and 0.054 mm, respectively. The Land CT2001 battery test system was used for the galvanostatic charge/discharge tests in the voltage range of 2–4.2 and 2–4.5 V under the condition of 0.1C at the temperature of 25 °C. The rate capability tests were obtained at different current ranges from 0.05C to 1C between 2.0 and 4.5 V. Cyclic voltammetry (CV) measurements were tested on the electrochemical workstation (CHI760E Electrochemical Workstation) at a scanning rate of 0.1 mV/s and a potential range of 2.0–4.2 and 2.0–4.5 V. At the end of 50 cycles, electrochemical impedance spectroscopy (EIS) measurements were performed on the measured cells at room temperature using an impedance analyzer (CHI760E Electrochemical Workstation) in a frequency range of 1 MHz to 1 mHz and the signal amplitude of 5 mV.

Results and Discussion

Figure b shows the TG profiles of the precursor; the first reduction of mass from 100 to 300 °C may result from the water loss under condensation of aldohexose and urea to N-substituted glucosylamine; a brown viscous liquid could be obtained in this stage, as shown in Figure b. With the increase of temperature, further condensation and polymerization occurred between the glucose and urea and the highly entangled polymer chains were formed; amide groups exposed in the chains could chelate Ni2+, Co3+, and Mn4+ at the atomic level that guaranteed the intimate mixing of transition-metal ions.[24] The second mass reduction may correspond to the diffusion of gas, nucleation, growth, and development of the NCM523 crystal structure.

The surface functional groups of the NCM523 precursor were investigated by Fourier-transform infrared spectroscopy, and the result is shown in Figure c. The peaks at 1406, 1577, and 3436 cm–1 correspond to −CN, −C=O, and −NH, respectively, which indicates the existence of amide groups in the entangled polymer chains. These amide groups uniformly distributed along the glucose chains could in situ chelate the transition-metal ions (Ni2+, Co3+, and Mn4+) and obtain the NCM523 polyhedron assemblies.

The phase purity of the prepared sample was first measured by X-ray diffraction. The XRD data reflect a representative pattern for the structure of α-NaFeO2 with an R3̅m space group. No impurity phases were detected. The XRD diagram also shows the clear peak separation of (006)/(102) and (108)/(110), which indicates good organized layer structures with good ordering for the material. To obtain more detailed crystal lattice parameters of the sample, Rietveld refinement was used. As displayed in Figure and Table , respectively, the refined unit-cell parameters are a = b = 2.877245 Å, c = 14.255477 Å, and c/a = 4.96, which are in conformity with those reported previously;[25] the high c/a values show good regulated metal ions in the transition-metal layer and minimal degree of cation mixing. The ratio of I003/I104 is another important sensitive parameter, and the calculated value of the XRD data is 1.38 by Rietveld refinement, which is greater than 1.2, indicating that the lithium ions are in low cation mixing at the 3a site, transition-metal ions (Ni, Mn, and Co) at the 3b site, and oxygen ions at 6c site.[26,27] The results confirm that this one-spot method could be successfully adopted to fabricate the NCM523 cathode material.

Figure 2

XRD patterns of the as-prepared LiNi0.5Co0.2Mn0.3O2.

Table 1

Structure Parameters Obtained from Rietveld Refinement of the As-Prepared LiNi0.5Co0.2Mn0.3O2

atom	site	X	Y	Z	occupancy
Li	3a	0	0	0.05	0.9693
Ni	3a	0	0	0.05	0.0307
Li	3b	0	0	0	0.0100
Ni	3b	0	0	0	0.4900
Mn	3b	0	0	0	0.3000
Co	3b	0	0	0	0.2000
O	6c	0	0		1

The valences of elements in the NCM523 were determined by X-ray photoelectron spectroscopy (XPS), and the results are summarized in Figure . C 1s with the binding energy of 284.8 eV was used to calibrate. There are two main spin-orbit lines for the Mn 2P spectrum, which are, respectively, 2p3/2 at 642.3 eV and 2p1/2 at 654.0 eV with separation of 11.7 eV, representing the Mn4+ cation.[28] However, a small number of low oxidation state Mn3+ are also detected, as shown in Figure a, with the presence of peaks at 641.8 and 653.2 eV.[29] As shown in Figure b, the Co 2p spectrum reveals two main peaks, which are Co 2P3/2 and Co 2P1/2 with binding energy values of 780.1 and 795.1 eV, respectively. The spin-orbit splitting is 15.0 eV, which refers to the Co3+ cation.[30,31] According to the fitted XPS curves, the Co4+ coexistence is confirmed by peaks at 783.6 and 798.9 eV. As the Ni 2p XPS spectrum shows in Figure c, the intense Ni 2p3/2 peak situated at 854.9 eV is accompanied by a shake-up peak located at 872.4 eV, which clearly indicates Ni2+ cations.[32] However, weaker peaks at 855.7 and 873.3 eV reveal the existence of the higher state of Ni3+. The existence of Mn3+ and Ni3+ cations is attributed to the electron transfer between Mn4+and Ni2+:[33][33] Mn4++ Ni2+ ↔ Mn3++ Ni3+. To supplement the volatilization of Li at high temperature and increase the electrochemical performance, NCM materials are usually prepared with the ratio of Li: transition metal (TM) slightly higher than 1. The excess Li+ existing in the TM layers could cause TM/O to drop below to 1:2; the higher oxidation states of TM elements are produced to balance the oxidation ion. Therefore, the existence of Ni3+ is meaningfully inevitable.[34] In the O 1s spectrum in Figure d, the peak at 529.4 eV corresponds to the Co–O bond, The peaks of Mn–O bonding and Ni–O bonding are, respectively, at 530.1 and 531.3 eV.[35] The XPS results are consistent with the reported NCM layered material.[36]

Figure 3

XPS spectra of (a–d) Ni 2p, Mn 2p, Co 2p, and O 1s in the as-prepared LiNi0.5Co0.2Mn0.3O2.

The structure and the morphology of NCM523 materials were observed by scanning electron microscopy (SEM). As displayed in Figure a,b, no signs of agglomeration were observed, which was favourable for the electrochemical performance. The polyhedrons with the size ranging from 600–1000 nm assembled with a special hierarchical structure, as shown in Figure a. During the heat treatment, the fused polyhedrons are interconnected to form a structure with a particle size ranging from 0.5 to 2 μm; the urea and glucose would decompose and release gases, which will lead to the formation of the observed pore structure. As shown in Figure b, the single-crystal particles with sharp edges fused together. This phenomenon is in good agreement with the XRD results that NCM523 exhibits a hexagonal layered structure. The distribution of the elements is indicated by the energy-dispersive X-ray spectroscopy (EDS) mappings. Figure d–f shows that the Co, Ni, and Mn elements are evenly distributed among the entire region, demonstrating the uniform mixing of transition-metal ions. As shown in the ICP results (Table S1), the stoichiometric ratio of Ni–Co–Mn for the NCM523 material is close to the expected value of 5:2:3; the results indicate that a highly uniform large nickel content could be obtained in the formation process.

Figure 4

(a, c) SEM image and the (b) magnified image of the as-prepared LiNi0.5Co0.2Mn0.3O2 and relevant elemental EDS mapping of (d–f) Co, Ni, and Mn.

The NCM523 material is further observed by transmission electron microscopy (TEM). Figure a shows a representative single-crystal morphology of particles, which is in agreement with the SEM observation. Figure b displays the high-resolution lattice-resolved TEM image of the sample at the edge of Figure a and shows a clear and uniform interplanar distance of 0.24 nm, which is in agreement with the lattice space of the (010) plane for the layered structure LiNi0.5Co0.2Mn0.3O2 and demonstrated by the electron diffraction of the selected region, as shown in Figure c. These results demonstrate the uniform mixing of transition-metal ions at the atomic level for NCM523.

Figure 5

(a) TEM; (b) high-resolution TEM (HR-TEM), and (c) selected-area electron diffraction (SAED) images for the as-prepared LiNi0.5Co0.2Mn0.3O2.

Figure a shows the cycling performances of NCM523 in the voltage range of 2.0–4.5 V at 0.1C. The first discharge capacity is 244.2 mAh/g. With the increase of the cycling number, the discharge capacity slowly decreased, which is probably caused by the parasitic reactions between the electrolyte and oxidative species found in the electrolyte or the dissolution of transition metal, etc.[37] Furthermore, structural recombination of the positive electrode surface will also lead to the capacity loss, such as the formation of NiMn2O4 and rock salt under the higher voltage operation;[38,39] these new phases could block the Li+ diffusion pathways and lead to the characteristic voltage decay on cycling. The Coulombic efficiency (CE) of NCM523 nearly approaches 100% since the seventh cycle. The low CE and rapid decrease of the capacity for the initial cycles may be caused by the parasitic reactions, which is due to the formation of a solid electrode interface (SEI) at the surface of cathode materials.[40] The cycling performances of NCM523 are comparable with the commercial materials with the morphology of the secondary particles.[41] The cycling performances of NCM523 in the voltage range from 2.0 to 4.2 V at 0.1C are presented in Figure S1. The results are similar to the cyclability results, as shown in Figure a. Figure c displays the relationship between the corresponding cell voltage and specific capacity of charge and discharge. The initial charge plot is composed of a long plateau and a slope area; the long plateau at higher voltage is attributed to the oxidation of other transition metals. The slope area represents the concomitant oxidation of Ni2+ with the extraction of Li+ ions from the LiMO2 component. Figure d displays the cyclic voltammograms of NCM523 in the voltage range of 2.0–4.5 V with a scanning rate of 0.1 mV/s. There is a reduction peak at 3.6 V and an oxidation peak at 3.9 V, which correspond to the deintercalation and intercalation of lithium ions. Figure S1 shows the CV results in the voltage between 2.0 and 4.3 V. The small redox peak separations of NCM523 at both voltage range may benefit from the optimized crystallization condition and minimized cation mixing of Li+/Ni2+. According to Qiao and Kim, the oxidation reaction of Ni2+/Ni4+ corresponds to the plateau lower than 4.2 V; however, the Co-like transition-metal ions with a higher valence redox reaction observed transition-metal redox up to 4.7 V.[42] In recent years, the single-crystal NCM has attracted extensive attention. Compared with their counterparts, the single-crystal structure can increase the electrochemical performance by reducing the diffusion length of lithium and providing a fast lithium-transportation channel. As a result, it is facilitated to the transportation of Li+ along the grain.[43] Hence, NCM523 reveals an excellent initial capacity especially at high voltage. However, to some extent, the smaller particle size compared with the polycrystalline could be a double-edged sword, it may generate more side reactions between the electrolyte and the electrode and produce negligible gas during use.[44] As every coin has two sides, the NCM523 electrode synthesized by this one-spot facial way expressed a unique hierarchical single-crystal structure, low Ni2+/Li+ mixing, competitive initial electrochemical performance at 4.5 V, and showed no negative effect of the cracking but the advantages are offset by the relative lower specific capacity in the long-term though it is competitive with the commercial one. However, further investigation is now undergoing in the way of increasing the grain size and crystal decoration, which could improve the performance of the NCM523 electrode for the long-term application of the Li-ion batteries.

Figure 6

(a) Specific discharge capacity and (b) Coulombic efficiency of the LiNi0.5Co0.2Mn0.3O2 cell as a function of the cycle number. (c) Cycle voltammogram of the first cycle with a scan rate of 0.1 mV/s at the voltage between 2.0 to 4.5 V. (d) Voltage curves of the LiNi0.5Co0.2Mn0.3O2 cell at the voltage between 2.0 and 4.5 V.

To further demonstrate the advantage for the application of the Li-ion batteries, the rate performance was further tested for 30 cycles (from 0.05C to 1C) at the voltage between 2.0 to 4.5 V. As shown in Figure a, the material exhibits comparable high initial capacities at 0.1C but decreases quickly with the cycling number, which indicates it has perfect electrical conductivity and quick lithium transportation due to the smaller grain size; however, at the same time, the serious reaction between the electrolyte and electrode particles leads to the capacity degradation. The more stable performance at higher current density shows that the material is more suitable to use between 0.5C and 1C; the electrode material should be further optimized at lower current density. In addition, the Nyquist plots of the NCM523 electrodes tested for the frequency range of 0.1 Hz to 100 kHz by electrochemical impedance spectroscopy (EIS) are displayed in Figure b with the corresponding equivalent circuit to fit the plots. The spectra for the first and 50 cycles at the voltage from 2.0 to 4.5 V show one Warburg tail at a low-frequency region and one semicircular arc at a high-medium-frequency region.[45−47]Rs is the electrolyte resistance, including the resistance of a solid/electrolyte interface and an electrolyte, while the equivalent internal resistance (RSEI) is relevant to the intercept of the semicircle at high frequency; the sample reveals that the resistance is very low. The charge transfer resistance (Rct) corresponds to the intercept of the medium-frequency region. W represents the Warburg impedance associated with Li+ diffusion; it was found that the sample has a relatively low Rct, which shows faster ionic migration, and as the cycle number increases, Rct is increased slightly from 192 to 218 Ω, which is closely related to the charge/discharge cycling data and rate capacity. The EIS result indicates that the NCM523 sample has relatively high electrical conductivity and extremely low polarization, which is benefited from the high hexagonal arrangement and minimized mixing of the cation of the NCM523 materials.

Figure 7

(a) Rate capabilities and (b) Nyquist plots and the equivalent circuit of LiNi0.5Co0.2Mn0.3O2 before and after 50 cycles.

Conclusions

A layered lithium transition-metal oxide, NCM523 cathode, is successfully synthesized via a facile and low-cost one-spot method, which uses glucose and urea to form a deep eutectic solvent. The hydrogen bonding between the hydroxide group of glucose and the carbonyl group of urea could strengthen the contacts and form relatively stable entangled chains. The amide groups in the glucose–urea chains could in situ chelate Ni2+, Co3+, and Mn4+ in an atomic layer along the chains in the lower temperature, and after the calcination treatment, homogeneous LiNi0.5Co0.2Mn0.3O2 polyhedron assemblies with a desirable single-crystal morphology could be obtained. The NCM523 electrode expressed a competitive initial electrochemical performance. The hierarchical porous nano/microstructure of NCM523 is beneficial to the performance. The hierarchical structure could also act as a shortcut for the quick migration of electrons, electrolytes, and Li+; the exposed active-crystal facet of polyhedrons could supply more pathways for the Li+ diffusion. Moreover, NCM523 exhibits a highly ordered hexagonal arrangement in an a–b plane and minimized antisite mixing of Ni2+/Li+. Furthermore, more comprehensive research is going on in the lab to engineer the phase, composition, and structure of the material to enhance the performance of the NCM523 electrode for the lithium-ion batteries for the long-term application.