Effect of nanostructured carbon coatings on the electrochemical performance of Li1.4Ni0.5Mn0.5O2+x-based cathode materials

Nanocomposites of Li1.4Ni0.5Mn0.5O2+x and amorphous carbon were obtained by the pyrolysis of linear and cross-linked poly(vinyl alcohol) (PVA) in presence of Li1.4Ni0.5Mn0.5O2+x. In the case of linear PVA, the formation of nanostructured carbon coatings on Li1.4Ni0.5Mn0.5O2+x particles is observed, while for cross-linked PVA islands of mesoporous carbon are located on the boundaries of Li1.4Ni0.5Mn0.5O2+x particles. The presence of the carbon framework leads to a decrease of the polarization upon cycling and of the charge transfer resistance and to an increase in the apparent Li+ diffusion coefficient from 10−16 cm2·s−1 (pure Li1.4Ni0.5Mn0.5O2+x) to 10−13 cm2·s−1. The nanosized carbon coatings also reduce the deep electrochemical degradation of Li1.4Ni0.5Mn0.5O2+x during electrochemical cycling. The nanocomposite obtained by the pyrolysis of linear PVA demonstrates higher values of the apparent lithium diffusion coefficient, a higher specific capacity and lower values of charge transfer resistance, which can be related to the more uniform carbon coatings and to the significant content of sp2-hybridized carbon detected by XPS and by Raman spectroscopy.


Introduction
LiNi 0.5 Mn 0.5 O 2 -based electrode materials [1] were proposed as a less expensive alternative to LiCoO 2 for high energy density Li-ion batteries containing less toxic elements than cobalt. The reasonable combination of their electrochemical properties (C = 160-180 mAh·g −1 at C/10; U = 2.5-4.6 V) remains attractive until now [2][3][4]. Most of the studies deal with Li(Ni,Mn)O 2 with equimolar amounts of nickel and manganese. The influence of the Ni/Mn ratio on the properties of these materials is discussed in [5]. One of the obstacles to the practical application of LiNi 0.5 Mn 0.5 O 2 is related to the internal ion exchange (ion mixing) of Li + and Ni 2+ due to the very similar ionic radii. This process could be partially suppressed by the introduction of extra lithium into LiNi 0.5 Mn 0.5 O 2 . Extra lithium ions are usually located in the (Ni,Mn) sublattice with the formation of Li [Li x Ni 0.5−(x/2) Mn 0.5−(x/2) ]O 2−δ [6,7]. The practical application of these materials is limited by the insufficient electronic conductivity of Li 1+x (Ni,Mn)O 2 materials [8] and their ability to catalyze the organic electrolyte decomposition at high potentials and currents [9,10].
The most common way of overcoming this problem is the modification of cathode materials by introducing additives and by depositing coatings that would suppress the interaction of electrolyte and the surface of particles. Various kinds of materials have been tested for surface modification, namely other cathode materials (LiMnPO 4 [11], LiMn 2 O 4 [12], LiCoO 2 [13], LiNiO 2 [14]) or simple binary compounds such as CaF 2 [15], TiO 2 [16], ZnO [17] and Al 2 O 3 [18].
During the assembly of the lithium-ion cells, the cathode materials are always mechanically mixed with carbon black in order to enhance the electronic conductivity. However, this method of carbon introduction provides the contact only between the external surfaces of cathode materials aggregates, but it does not improve the coherence and hence the electrical conductivity inside the aggregates.
Both kinds of problems can be solved by the fabrication of conducting carbon coatings on the surface of Li(Ni,Mn,Co)O 2+x particles [19][20][21][22]. Such coatings can be obtained by the impregnation of oxide powders with the solution or the melt of organic compounds accompanied by subsequent heat treatment in the absence of oxygen or at limited oxygen access [23][24][25]. These carbonaceous coatings improve the electrokinetic properties of cathode materials, which results in enhanced capacities at high discharge rates [26]. The variation in the electrochemical properties is governed by the increase in the electronic conductivity of the obtained coatings, which depends on both the organic precursor composition and the preparation conditions. Thus, the investigation of the influence of physicochemical properties of carbon coatings on the electrochemical properties of Li(Ni,Mn,Co)O 2+x /C composites proves to be an important task. The choice of an organic precursor is an important challenge. In the case of Li(Ni,Mn)O 2 , it is rather complicated due to the high oxidizing ability of this material, which leads to its intensive interaction with both the products of the pyrolysis of organic compounds [27] and carbon [28]. As it has been shown before, the rate of this interaction could be significantly reduced by using some polymer precursors [28]. They allow for a solid conducting film consisting of pyrolysis products to be obtained at moderate temperatures (300-350 °С). Poly(vinyl alcohol) (PVA) could be one of such precursors. It eliminates water at temperatures higher than 230 °С, forming carbon residues with a system of conjugated double bonds [29,30].
In the present paper, the electrochemical properties of Li 1.4 Ni 0.5 Mn 0.5 O 2+x /C composites are discussed, the preparation and characterization of which have been described before in detail [28,31]. The melts of linear or cross-linked PVA served as carbon source. We investigated the influence of composition and micro/nanomorphology of the carbonaceous coatings obtained by the pyrolysis of various kinds of PVA on the kinetics of lithium intercalation-deintercalation and the electrochemical properties of Li 1.4 Ni 0.5 Mn 0.5 O 2+x /C composites.

Results and Discussion
TEM micrographs of pure Li 1.4 Ni 0.5 Mn 0.5 O 2+x (LNM) and LNM/C composites obtained by the pyrolysis of both linear and cross-linked PVA are shown in Figure 1. In the case of LNM/C composites obtained from linear PVA, 5-7 nm thick carbon coatings can be clearly observed on the particle surface as well as amorphous carbon bottlenecks between particles forming a consolidated 3D network ( Figure 1B). Concerning the LNM/C composites obtained from cross-linked PVA, the continuous carbonaceous coatings on oxide particles are absent. The carbon is localized as separate islands on the boundaries of Li 1.4 Ni 0.5 Mn 0.5 O 2+x particles ( Figure 1C). The carbon obtained from cross-linked PVA has a sponge-like microstructure with nanometer-sized mesopores ( Figure 1D). Thermogravimetry (TG) curves of the linear and cross-linked PVA demonstrate the intense pyrolysis of the linear PVA at 240-310 °C ( Figure 2A) and the continuous decomposition process of the cross-linked PVA at 240-600 °C ( Figure 2B). The latter shows a significantly higher thermal stability of the cross-linked polymer. The analysis of differential scanning calorimetry (DSC) curves allows one to identify a small "endo"effect at 230 °C before the decomposition of the linear PVA ( Figure 2A) that could be associated with its melting [32]. The corresponding effect is absent in the DSC curve of the crosslinked polymer ( Figure 2B).
The reasons for the different localization of carbon in the samples can be associated with the observed features of the pyrolysis of precursors. Melting of linear PVA causes a wetting of LNM grains with the polymer melt followed by a relatively uniform pyrolysis of the thin polymer films on the surface of oxide crystallites. In the case of cross-linked PVA, the polymer    Another difference between the pyrolysis products of linear and cross-linked PVA concerns their different chemical composition. According to the C 1s XPS spectra of carbon-coated LNM ( Figure 3, Table 1 it takes place over a wider potential range. It means that the carbon nanocoatings enable the polarization decrease upon lithium intercalation [23]. After galvanostatic cycling at 20-100 mA·g −1 , CVA curves were recorded at the 15th cycle ( Figure 4В). In the case of pure Li 1.4 Ni 0.5 Mn 0.5 O 2+x , the potential range of lithium intercalation expands and in the anodic region, the main lithium deintercalation peak shifts towards higher potentials (from 3.78 V at the 1st cycle to 3.85 V at the 15th cycle). This fact confirms a poorer reversibility of lithium insertion/extraction into the pure Li 1.4 Ni 0.5 Mn 0.5 O 2+x .
The electrokinetic properties of LNM/C nanocomposite obtained by the pyrolysis of linear PVA are significantly better than the ones of the corresponding composite obtained by the pyrolysis of cross-linked PVA. In particular, after galvanostatic cycling at 20-100 mA·g −1 , the cathodic peak at ca. 3.6 V is more pronounced and the range of lithium intercalation potentials is narrower ( Figure 4). Apparently, this phenomenon could be ascribed to a more uniform carbon distribution and to the higher content of conductive sp 2 -hybridized carbon in the coatings obtained by pyrolysis of linear PVA compared to the pyrolysis of cross-linked PVA [28]. Another reason for the superior electrokinetic properties of the LNM/C nanocomposite obtained from linear PVA could be associated with the absence of the reversible structural transition in the partially delithiated Electrochemical impedance (EI) measurements were performed to investigate the details of the lithium insertion-extraction processes in LNM/C nanocomposite cathodes ( Figure 5A-D). All the plots are mainly composed of a small intercept at high frequencies, a semicircle at high to medium frequencies and a linear part in the low-frequency region. The real and complex EI parts are normalized to the real surface area of each electrode calculated according to the procedure described earlier [31]. The impedance spectra are fitted using the equivalent circuit model ( Figure 5E) [34,35]. The small intercept is almost the same for the electrodes and corresponds to the solution resistance of the cell (Re). Rf and CPEf stand for the Li + migration resistance and the capacity of the surface layer, respectively. Rct and CPEct stand for the related charge-transfer resistance and the double-layer capacitance, respectively. Zw represents the diffusion-controlled Warburg impedance in the lowfrequency region [34,35]. for the one obtained from cross-linked PVA at 3.75 and 3.72 V, respectively. This difference can be explained by the different morphologies and compositions of carbon-based nanocoatings, in particular, by the higher ratio of sp 2 -to sp 3 -hybridized carbon in the LNM/C composite obtained from linear PVA [28]. This results in a higher electronic conductivity and higher D values. The different amounts of residual intermediates of pyrolysis also play a role.
For all the investigated materials, the charge-discharge curves for the intermediate cycles obtained at different rates are shown in Figure 7A. In contrast to pure Li 1.4 Ni 0.5 Mn 0.5 O 2+x , both LNM/C nanocomposites exhibit is a well-pronounced plateau at 3.75 V in the discharge curves, especially at a current density of 20 mA·g −1 . Apparently, this fact confirms the polarization decrease for the LNM/C composites discussed earlier during the analysis of CVA results.  The discharge capacity variations upon cycling are represented in Figure 7B. In the case of pure Li 1.4 Ni 0.5 Mn 0.5 O 2+x , the capacity rapidly drops at the 4th cycle. For the LNM/C nanocomposites, the initial capacity values are lower but the stability upon cycling is significantly better. For the LNM/C nanocomposite obtained from linear PVA, the capacity values are usually higher than for the one obtained from cross-linked PVA. They attain a value of 165 mAh·g −1 at a current density of 20 mA·g −1 . This fact correlates quite well with the higher electronic conductivity of this material and the higher values of the Li + diffusion coefficient. Figure 8 displays the XRD patterns of electrode materials after electrochemical cycling at high discharge currents. It can be seen that the splitting of (018)/(110) reflections is much higher in carbon-coated materials compared to the pure Li 1.4 Ni 0.5 Mn 0.5 O 2+x (Figure 8, inset).  Hence, it was demonstrated that even a partial modification of the synthesis technique (additional stage of thermal crosslinking of PVA) leads to LNM/C composites with different properties of the carbon coatings, including a different micro-/ nanomorphology and a different ratio of sp 2 -to sp 3 -hybridized carbon. The last feature is also confirmed by the different integral intensities of D-and G-peaks in the Raman spectra of the pyrolysis products of linear and cross-linked PVA [28,40]. The variation of sp 2 -and sp 3 -hybridized carbon content has to do with the features of the composition and pyrolysis of organic polymer precursors. During the pyrolysis of linear PVA at 200-230 °C, its dehydration is accompanied by the formation of C=C chains with conjugated double-bond systems. In the case of cross-linked PVA, the number of free OH-groups is substantially lower due to the formation of chemical bonds between neighboring linear PVA chains. Hence in the course of the dehydration of this polymer, the formation of C=C double bonds is much less probable.

Conclusion
The various properties of carbonaceous nanocoatings affect both the kinetics of lithium insertion-extraction and the electrochemical properties of LNM/C composites.  [25]. Due to the different character of the pyrolysis processes of linear and cross-linked PVA, the amount of PVA was optimized in order to ensure the comparable carbon content in the final composites (4 and 5.5 wt % for the linear and cross-linked PVA, respectively [28]).
The samples were characterized by XRD using a D/MAX 2500 diffractometer (Rigaku) in the reflection mode with Cu Kα radiation and a curved-graphite [2] monochromator placed in the reflected beam (2θ range 10-90°, step 0.02°, acquisition time 3 s per step).The analysis of diffraction patterns was performed by using WinXPow software and PDF-2 powder diffraction database.
The thermal analysis of the linear and cross-linked PVA was performed in argon by heating to 700 °C at a rate of 10 °C·min −1 (STA 209 PC Luxx thermal analyzer (Netzsch)).
The morphology of the composites and their electron diffraction patterns were analyzed by transmission electron microscopy (Libra 200 MC, Carl Zeiss) at an accelerating voltage of 200 kV and a magnification of 30,000-300,000×.
The X-ray photoelectron spectra were acquired with a Kratos Ultra DLD spectrometer using a monochromatic Al Kα X-ray source that possesses an analysis area of 300 μm × 700 μm. The spectra were recorded in a constant analyzer pass energy mode of 5 eV resulting in a resolution better than 0.3 eV.
For the preparation of the electrodes, the cathode paste (active material (85 wt %), acetylene black (10 wt %, Timcal) and polyvinylidene fluoride (5 wt %, solved in N-methyl-2-pyrrolidone) was coated on a stainless steel net (0.05 mm thick). After pressing, the electrodes were dried in vacuum (0.5-1.0 mbar) at 120 °C for 8 h. The active electrode, counter electrode (Li) and reference electrode (Li) were placed in a hermetic Teflon cell using a porous polypropylene separator. 1 M LiPF 6 solution in ethylene carbonate (EC)-dimethyl carbonate (DMC)-diethyl carbonate (DEC) (1:1:1 by volume) was used as an electrolyte. According to Fischer titration data, the water content in the electrolyte did not exceed 25 ppm. In the course of cell assembly, the water content in the glove box did not exceed 0.5 ppm. All the potential values in this manuscript are referred to the Li/Li + electrode.
The galvanostatic curves and cyclic voltammograms (CVA) were registered using 100N Metrohm Autolab equipment. The cells were cycled in a potential range of 2-4.6 V at a current density of 20-100 mA·g −1 at room temperature. The potential scan rate was 50 µV·s −1 .
Electrochemical impedance (EI) measurements were performed on a "Solartron 1255B" using the cells assembled as described above. The amplitude of the AC signal was 5 mV over a frequency range from 1 MHz to 10 mHz. The Corrware 2 and CorrView 2 software (Scribner Associates) was employed for the electrochemical experiments and the obtained hodographs were treated using "ZView-Impedance Software".