Effect of lithium borate coating on the electrochemical properties of LiCoO2 electrode for lithium-ion batteries

The effect of a protective coating of fused lithium borate, Li3BO3, on the physicochemical and electrochemical characteristics of LiCoO2 has been studied. A cathode material produced by the SCS method using binary organic fuel, glycine and citric acid. The influence of the experiment conditions on the morphology, crystal structure and specific surface of lithium cobaltite was studied. Electrochemical testing of LiCoO2∙nLi3BO3 samples, n = 5 and 7 mass %, has been performed in the cathode Li|Li+-electrolyte|LiCoO2∙nLi3BO3 half-cell using 1M LiPF6 in EC/DMC mixture (1:1) as electrolyte in the 2.7-4.3 V range at normalized discharge current С/10, С/5, С/2. The maximal initial discharge capacity of 185 mAh/g was detected for the samples with 5 mass % Li3BO3. The coulomb efficiency of optimal materials in the 40 th cycle was 99.1%.


Introduction
Lithium cobaltite LiCoO 2 (LCO) is used as a cathode material since 1990, and in spite of the appearance of such promising cathode materials as LiNi 1/3 Mn 1/3 Co 1/3 O 2 (NMC), LiNi 0.8 Co 0.15 Al 0.05 O 2 (NCA), LiMn 2 O 4 (LMO) etc. it is still employed as a component of lithiumion batteries (LIB) with low discharge rates in portable gadgets [1]. Chemical interaction between the electrolyte and the cathode material leads to non-recoverable losses of lithium cations reduced service lifetime and accelerated capacity failure of LIB. The manufacturers of LIB materials try to eliminate this effect by decreasing the specific surface of dispersed materials, by using microgranulation processes or applying modifying coatings on cathode material particles, protecting 3 them from the action of acid fluorine-containing components of electrolyte. In particular, Al 2 O 3 , ZrO 2 , ZnO, SiO 2 , TiO 2 and other oxides are used as protective coatings [2][3][4][5]. Recently appeared publications reporting the application of glasses and boron and lithium based oxides as coatings [5][6][7]. A. Nagasubramanian et al. [6] studied the effect of LiBO 2 coating on the electrochemical performance of orthorhombic LiMnO 2 cathode. ShuangYuan Tan et al. [7] showed that using glasscoated NMC/Li 2 O·2B 2 O 3 can increase the discharge capacity retention of the cathode from 22.5% to 57.8% at −40 °C. Among compounds in the Li 2 O-B 2 O 3 system, lithium borate Li 3 BO 3 should be noted [8][9][10]. It has the lowest melting temperature, 715±15 °C [11], which allows applying it as a flux for more refractory compounds, creating dense protective coatings. In addition, Li 3 BO 3 is a lithium ion conductor [10], and its molecular mass is smaller than that of LiCoO 2 and other cathode materials. Li 3 BO 3 coating also increases the concentration of Li + in the contact layer with electrolyte.
Most coating strategies are based on the sol-gel method or impregnation of cathode materials powders with salt solutions with subsequent drying and annealing [1][2][3][5][6][7]. However, the application of lithium borates via solutions guarantees neither synthesis of the nominal composition of lithium borate nor the density of the coating after annealing at 500 °С. It can be assumed that addition and distribution of LiBO 2 , Li 2 B 4 O 7 or Li 3 BO 3 compounds in the cathode bulk with subsequent annealing at melting temperatures of the corresponding eutectics may cause positive effect.
cathode material can be produced almost without sewage; 2. energy consumption for the decomposition of precursors is reduced compared to conventional technologies, since the method employs internal exothermal processes requiring only relatively moderate energy consumption for evaporation of reaction solutions and preliminary heating of xerogel before the beginning of redox reaction; 3. dispersion and chemical activity of produced precursor may reduce the time of hightemperature annealing to attain the monophasic; 4. although the initial solutions contain nitrates, combustion proceeds with almost complete transformation of nitrogen dioxides into molecular nitrogen;

5.
this method allows to control the dispersion of the material while reducing the costs for milling and lessening the risk of pollution of the material during milling. 4 However, in the LiNO 3 -Co(NO 3 ) 2 -glycine (urea) systems, a redox reaction of LCO formation proceeds intensely with outflow of a considerable part of precursor with effluent gases outside the reactor. Under stoichiometric combustion conditions, the combustion rate may have an explosive character. It is reasonable to expect that the described above effects detected under conditions of laboratory experiments will be multiply strengthened if the mass of the material is increased. This problem can be solved by using of less energetic fuel for controlled reduction of SCS rate, for example, sucrose, ammonium acetate starch, citric acid and oxalic acid [24][25][26][27][28].
In this paper, we report structural, morphological, dimensional and electrochemical characteristics of LCO powders produced in SCS reactions with glycine and citric acid with subsequent coating with fused lithium borate Li 3 BO 3 .

Starting materials
For the combustion synthesis of LCO powders, cobalt(II) nitrate hexahydrate (99%) and

Synthesis
Lithium nitrate combined with cobalt nitrate imparts excessive combustion intensity to SCS reactions. In the methods where glycine used as a fuel / reductant, the combustion rate during synthesis of cathode materials of LIB can be considerably decreased by replacing lithium nitrate by lithium carbonate or lithium citrate [24]. Besides, the reduction of the fraction of cobalt nitrate (oxidizer) due to its replacement by cobalt citrate also lowers the SCS intensity allowing the yield increasing of the resulting material [24].
In this work, we used one-step mode of LCO production. For this purpose, a 150 ml solution of cobalt nitrate (66.67 g/dm 3 Co) and citric acid (237.5 g/dm 3 ) (solution 1) was placed into a 2 dm 3 reactor, to which 120 cm 3 lithium citrate solution (286.68 g/dm 3 Li 2 HC 6 H 5 O 7 ) was added. Cobalt (II) carbonate hydroxide hydrate and glycine successively added to the resulting solution ( Table 1).
The reaction solution was heated on an electrical heater with a capacity of 1 kW (the temperature of the heater was 550-600 °С) for dehydration and initiation of SCS reaction. The produced LCO precursor was ground in a ball mill with grinding bodies made of stabilized 5 zirconium oxide prior and between the annealing at 650, 800 and 850 °С. The annealing duration at each stage was 10 h.
After certification, LCO powders were coated with fused Li 3 BO 3 (LBO) produced preliminarily in a solid-phase reaction using lithium carbonate and boric acid at 560-600 °С for 35 h with intermediate grinding. 5 and 7 mass % of LBO were added to the initial LCO powder and mixed in a ball mill for 1 h, then they were annealed at 750 °С for 5 h. After the first annealing, the LCO/LBO samples were repeatedly ground and annealed for the second time at 750 °С for 5 h to produce better uniform coating.

Characterization of powder samples
The diffraction patterns of the powders were taken at room temperature with a Shumadzu XRD-700 (Cu Kα1 radiation, 2θ = 10-80°) diffractometer, equipped with PDF2 database. The refinement of the crystal structure according to the Rietveld method carried out using the software package FullProf [29]. SEM images obtained with a JEOL JSM 6390 LA microscope. Specific surface area (S) of powders was determined by BET nitrogen desorption during heating in a SORBI N4.1 (Meta, Russia). The particle size distribution of the obtained powders was determined using a Horiba LA-950V2 laser particle meter.

Electrochemical measurements
The electrochemical properties of the LCO powder samples were studied using two-electrode

Results and discussion
SCS as the chosen LCO obtaining method allowed to increase the mass of the obtained product without growing the reaction temperature and rate, i.e. without discharge of the produced powder. Evaporation of the solution led to the formation of a semi-sphered dried gel covered with a violet film. The dry outer layer gradually closed the whole semi sphere, inside which a wet gel remained (Fig. 1а) also transforming gradually into dry precursor material. The redox reaction proceeded in the form of a heating wave (Fig. 1b) with low violet flame. There was no discharge of the material outside the reactor, and no visible traces of nitrogen dioxide were present. When the combustion was completed, a bulky flake black powder of LCO precursor was obtained.

Fig. 1. Formation of the product in SCS reactions
The LCO powders obtained after combustion possessed high dispersion and revealed chemically non-equilibrium state due to incomplete crystal lattice formation processes. This is connected with a short combustion time (the reaction mass was in the high-temperature region for less than 3-5 min); low density of the material (as a result of large amount of gaseous products) hampers the completion of diffusion processes during LCO formation. Usually, the SCS process is supplemented with high-temperature annealing, in this case in the range of 650-850 °С, to remove carbon-containing impurities and form the LCO crystal structure (Table 1, Fig. 2). Upon annealing, the crystal lattice parameters of the produced LCO corresponded to the literature data [30].
Reflections of the fused LBO coating were not recorded, probably due to its glassy character.  The performed sedimentation analysis revealed that LCO represents finely dispersed powders forming agglomerates with the maximal diameter, D max , of <30 µm (Table 2) and the average particle diameter, D av , of 8.4-11.5 µm. The fraction of particles less than 5 µm is rather large, from 8 6 to 30%. The application of borate coating considerably changes the particle size distribution in the material; D av increases to 11-45 µm and depends explicitly on the milling conditions and load. The fraction of particles with a diameter less than 5 µm decreases to 1.1-1.3% (Table 2), which should have a positive effect on the cathode stability during the interaction with electrolyte. However, the presence of aggregate fractions larger than 30 µm required the classification of powders before applying the electrode mass.
Coating of LCO with a layer of fused lithium borate practically does not change the specific surface of the material (Table 2), remaining equal to 0.8-0.98 m 2 /g. This value is larger than the traditional values for commercial cathode materials, 0.4-0.6 m 2 /g, but possibly has positive impact on the electrochemical characteristics during cycling. The advantage of coating by fusion is that fusion and spreading of LBO on the surface of LCO particles and agglomerates decreases the coating thickness and increases the probability of connection of small particles, reducing the influence of electrolyte.
The sizes of LCO-LBO powders are slightly larger than the base (LCO) particles, it contain a smaller percent of fine fractions due to enhanced sintering into agglomerates and additional annealing time (growth of primary crystallites) (Fig. 3).  The charge-discharge characteristics of cathode materials based on LCO were studied in the potential range of 2.75-4.3V vs Li/Li + . The charge and discharge rate of the first 10 cycles were 0.1C. Then, at the same charge rate, 10 cycles performed with a discharge rate of 0.2C, 0.5C and again 0.1C. Fig. 4 shows the capacity dependences on the cycle number based on a series of powders: LCO, LCO + 5% LBO and LCO + 7% LBO as samples with the most stable characteristics.
The discharge capacity of the sample without coating at the first cycle was 166 mAh/g. A significant increase of the discharge capacity (up to 185 mAh/g) was achieved by using a 5% LBO coating. When the LBO content of the sample increases to 7%, the discharge capacity drops to 164 mAh/g. The negative effect probably related to the increasing thickness of the coating layer and/or the formation of less conducting glasses during the interaction of fused LBO with LCO. Fig. 4. Cyclic performance of the LCO samples: 1 -LiCoO 2 , 2 -LiCoO 2 +5% LBO, 3 -LiCoO 2 +7% LBO All samples retain 94, 93 and 93% of the original discharge capacity after 40 cycles when the cycle rate returned to the original rate 0.1C. In addition, the coulomb efficiency of these samples is more than 99% throughout all 40 charge-discharge cycles (Table 3).  The charge-discharge curves have a plateau at a potential of 3.9 V and two small quasiplateaus at 4.1 and 4.2 V. These plateaus correspond to the peaks on the cyclic voltammograms ( Fig. 6). According to the literature data [32][33][34][35][36], the main peak at 3.9 V is related to the first order transition from LCO to Li 0.8 CoO 2 ; the two less pronounced peaks at ~4.06 and ~4.17 V are associated with phase transitions to the monoclinic structure and back to the hexagonal structure.
Based on the obtained data, it is possible to conclude that borate coating does not affect the electrode polarization, as the discharge plateau does not change. Thus, borate coating increases the stability of cobaltite cycling at elevated rates. When the rate increases to 0.5C, the capacity falls by 4% for coated samples and by 8% for uncoated samples.

Conclusions
The best electrochemical characteristics, discharge capacity of 185 mAh/g at 0.1C and coulomb efficiency of 99.1% after 40 cycles, were demonstrated by the sample obtained in the SCS reaction of cobalt nitrate with lithium citrate and glycine and coated with 5 mass % of LiBO 3 . Fused LiBO 3 coating increases the cobaltate cycling stability at elevated rates.