Improved Electrochemical Performance of Rb_xLi_1.27-xCr_0.2Mn_0.53O₂ Cathode Materials via Incorporation of Rubidium Cations into the Original Li Sites

The Rb_xLi_1.27-xCr_0.2Mn_0.53O₂ (x = 0, 0.01, 0.03 and 0.05) cathode materials were prepared by solid reaction method. In a typical process, according to the chemical composition, a certain amount of lithium hydroxide, kept in excess of 5 wt% in the stoichiometric ratio, was dissolved in distilled water. Cr₂O₃ (99%), Rb₂CO₃ (99%) and MnO₂ (99%) powders were added to form a suspension with continuous stirring. The resulting suspension was poured into a ball mill and treated for 2 hours to achieve a particle size of 100∼300 nm using Zirconia beads as the grinding media. Then the ball-milled slurry was spray dried and the precursor powders were obtained. Finally the powders were calcined at 500°C for 3 hours and then fired at 900°C for 12 hours in nitrogen to form the target Rb_xLi_1.27-xCr_0.2Mn_0.53O₂ (x = 0, 0.01, 0.03 and 0.05) cathode materials.

The phase purity and crystal structure determination of the prepared powders were identified by TD3200 X-ray diffraction method with Cu Kα radiation (λ = 1.54056Ǻ) carried out at 40 KV and 30 mA. The data were collected in the 2θ range of 10°–80°. The XRD data with Rietveld refinement were carried out using Jade 9 software. The morphologies of the samples were examined by scanning electron microscopy (SEM, Hitachi, SU8200).

The electrochemical performances of the samples were investigated using electrodes CR2016 coin-type cells, which were assembled inside a glove box filled with Ar. For electrochemical characterization, the working electrodes were prepared by mixing 80 wt% cathode material, 10 wt% conductive carbon, and 10 wt% polyvinylidene difluoride (PVDF) in N-methyl- 2pyrrolidone (NMP) solvent. After thorough mixing, the homogenous slurry was cast onto an aluminum foil current collector and then dried under vacuum at 120°C for 12 h. The electrochemical performance of the prepared electrodes were determined on a Land CT2001 battery tester at the voltage of 2.0–4.8 V. Electrochemical Impedance Spectroscopy (EIS) was conducted by an electrochemical workstation (Autolab Pgstat302n) and the data were collected in the frequency range 0.05–500 KHz.

Phase identification was carried out with XRD measurements. Fig. 1 showed the XRD patterns of the cathode Rb_xLi_1.27-xCr_0.2Mn_0.53O₂(x = 0, 0.01, 0.03 and 0.05) materials prepared at 900°C. It is clear from the patterns that all the peaks are sharp and well-defined. All the peaks can be indexed into a hexagonal α-NaFeO₂ structure, except several ordering peaks between 20° and 30° which was regarded as the superlattice of Li₂MnO₃ type structure by several groups.^24–26 As presented in Fig. 1, the Rb modified materials exhibit no new peaks or impurities compared to the pristine Li_1.27Cr_0.2Mn_0.53O₂, implying that Rb ions were incorporated into the mother lattice and did not form new compounds in the prepared powder. The slightly shifted peaks shown in the inset imply that the crystal parameters were slightly increased by Rb doping since the ionic radius of Rb ions is much bigger than that of Li ions. Furthermore, it can be observed from the patterns the splitting of the (006)/(012) and (018)/(110) doublets, which indicated a highly ordered layered structure for the prepared cathode materials. The above XRD results strongly indicated the successful synthesis of a pure phase for the layered hexagonal product in the experiment, and the obtained Li_1.27-xRb_xCr_0.2Mn_0.53O₂ compound can be regarded as a normal layered Li[Li_0.27-xRb_xCr_0.2 Mn_0.53]O₂ compound. For samples with x less than 0.03, no Rb-related peak is observed, implying that Rb ions have been successfully introduced into the host layered structure. When the content of Rb ion increased to x = 0.05, some other diffraction peaks were observed between 12° and 30°, showing the formation of impurity phases and therefore a suitable doping concentration of 3% was determined.

Zoom In Zoom Out Reset image size

The experimental XRD patterns were further analyzed by Rietveld refinement to study the structure in detail. Fig. 2 shows the Rietveld refinements of the XRD patterns for Rb_xLi_1.27-xCr_0.2Mn_0.53O₂ (x = 0, 0.01, 0.03 and 0.05). A structural model with R-3m space group was employed during the Rietveld refinement. To better understand the effect of Rb doping on structure of the prepared samples, lattice parameters and atomic occupancies of all samples were provided on the basis of high symmetry R-3m space group, as listed in Table I and Table II. Generally the weight profiled factor Rwp and expected R factor (Rexp) in Table I are two important factors to evaluate the refinement results, and it is reliable and acceptable when the Rexp is below 10%.²⁷ Therefore, in our case, as can be seen in Table I, the small weight profiled factor Rwp and expected R factor (Rexp) demonstrated the proposed structural model R-3m is correct, and the refinement results are acceptable. It is also found by the refinement that Rb ions were placed in the original Li sites (3b site) and the related site occupancies are very close to the doping concentration. Therefore, Rb ions were introduced into the layered structure in our case. Furthermore, it can be seen from Table I that the unit volumes of Rb replaced samples were enlarged in comparison with the sample without Rb (x = 0). The variations of cell volumes explicitly suggest that Rb ions, with larger ionic radius have been successfully introduced into the bulk crystalline structure of Li layers by replacing the original Li ions which caused the expansion of the lattice. The same tendency was observed for cell parameter of a axis value. However, difference was observed for c axis value. When Rb doping concentration reaches x = 0.03, sample Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ (x = 0.03) demonstrates the largest value of parameter c which stands for the interspacing of transition metal layers. The increased c value can reduce the energetic barrier impending Li-ions migration and facilitate the Li insertion and extraction upon electrochemical cycling. Further increase in x value (x = 0.05) did not keep the tendency, but decreased c axis value. Therefore, we can judge from the above results that with small amount of Rb replacing, the crystal lattice was uniformly expanded, but large distortion in crystal parameters will be formed when large amount of Rb ions were introduced. The reason for such distortion in crystal lattice is not clear yet and will be studied elsewhere.

Zoom In Zoom Out Reset image size

SEM observation of the prepared powder.—Surface morphologies, as well as particle sizes, were investigated by SEM. Fig. 3 shows the SEM images of the pure and Rb-incorporated Li_1.27Cr_0.2Mn_0.53O₂ materials prepared at 900°C under the N₂ atmosphere. It can be seen from the micrographs that the four samples display very similar morphology of round primary grains with smooth surface and the average grain size for the primary particles was determined as 100–300 nm. It can also be observed from the micrographs that with the increase of doping concentration, the particle size is slightly reduced, indicating that the doping of Rb ions is beneficial in achieving smaller crystalline size, although the reason is not clear yet. Microspores among the nano sized grains were also observed for all sample powders. Such nano-micro structures of the spherical powders provide enough spaces for electrolyte and thus good electrochemical performance was expected.

Zoom In Zoom Out Reset image size

The initial charge/discharge curves of Rb_xLi_1.27-xCr_0.2Mn_0.53O₂ (x = 0, 0.01, 0.03 and 0.05) electrodes at a constant current density of 20 mA·g⁻¹ (0.1C) between 2.0 and 4.8 V at room temperature were compared in Fig. 4. It is clear from Fig. 4a that, Rb substituted samples exhibit super electrochemical performance over the pristine one. The undoped Li_1.27Cr_0.2Mn_0.53O₂ cathode delivered a discharge capacity of 182.7 mAh·g⁻¹ at first run, while the Rb substituted electrodes show higher capacity, with 213.5 mAh·g⁻¹ for x = 0.01, 229.1 mAh·g⁻¹ for x = 0.03 and 195.6 mAh·g⁻¹ for x = 0.05. Meanwhile, the Li_1.27-xRb_xCr_0.2Mn_0.53O₂ electrodes offered an initial coulombic efficiency of 73.8%, 77.3%, 78.4% and 77.2% for x = 0, x = 0.01, x = 0.03 and x = 0.05. Compared with the undoped Li_1.27Cr_0.2Mn_0.53O₂ material, sample Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ shows a much higher discharge capacity, which can be ascribed the fact that slight lattice doping expands the interplanar spacing and thereafter reduces the activation barrier for lithium migration and thus allows more Li⁺ intercalation and deintercalation inside the layered structure. As reported by Wu et al., Li_1.27Cr_0.2Mn_0.53O₂ shows the first round of coulomb efficiency of 70.7%⁸. Therefore, it can be concluded that, our approach, rubidium doping is effective in improving the coulombic efficiency at first run.

Zoom In Zoom Out Reset image size

The typical charge and discharge profiles for the different cathode are plotted in Fig. 5. The cells are cycled between 2.0 and 4.8 V at a current density of 20 mA·g⁻¹ (0.1C) at room temperature. For comparison, the pure Li_1.27Cr_0.2Mn_0.53O₂ shows a discharge capacity of 189.3 mAh·g⁻¹, 202.3 mAh·g⁻¹, 196.2 mAh·g⁻¹, 188.1 mAh·g⁻¹, 186.7 mAh·g⁻¹ and 178.6 mAh·g⁻¹ on 1st, 10th, 20th, 30th, 40th and 50th cycle, respectively, whereas the discharge capacity of Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ under the same testing condition is 224.9 mAh·g⁻¹, 247.6 mAh·g⁻¹, 234.7 mAh·g⁻¹, 226.4 mAh·g⁻¹, 225.3 mAh·g⁻¹ and 222.1 mAh·g⁻¹. By comparison, an obvious increase of 18–25% was achieved by Rb replacing. Furthermore, it is recently reported that the Al-doped cathode delivered an initial discharge capacity of 198 mAh·g⁻¹ with a current density of 12 mA·g⁻¹.²⁸ In our case, the initial discharge capacity for sample x = 0.03 is 224.9 mAh·g⁻¹. It can be seen that the Rb-doped electrodes is also an effective way to improve the initial discharge capacity, and the sample Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ (x = 0.03) electrode displays the best specific capacity among the tested four electrodes.

Zoom In Zoom Out Reset image size

The rate performance was investigated from 0.1C to 5C as presented in Fig. 6a. Although all the electrodes showed a decrease in capacity with the increased discharge rates, the Rb-doped electrodes showed higher capacity than the Li_1.27Cr_0.2Mn_0.53O₂ electrode at each rate, and among them the Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ electrode displays the best rate performance. For comparison, the pure Li_1.27Cr_0.2Mn_0.53O₂ shows an average discharge capacity of 182.7 mAh·g⁻¹, 172.8 mAh·g⁻¹, 167.9 mAh·g⁻¹, 157.1 mAh·g⁻¹, 146.6 mAh·g⁻¹ and 65.5 mAh·g⁻¹ at 0.1C, 0.2C, 0.5C, 1C, 2C and 5C, respectively, whereas the discharge capacity of Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ under the same testing condition is 229.1 mAh·g⁻¹, 215.6 mAh·g⁻¹, 190.7 mAh·g⁻¹, 173.1 mAh·g⁻¹, 157.7 mAh·g⁻¹ and 86.1 mAh·g⁻¹, giving nearly 20∼30% increase at each rate. It is also found from the figure that the capacity can be easierly recovered after 5C charge/discharge, showing the potential of the cathode material in practical application. It also can be seen from the figure that sample Li_1.26Rb_0.01Cr_0.2Mn_0.53O₂ (x = 0.01) and Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ (x = 0.03) both have higher discharge capacity than sample Li_1.27Cr_0.2Mn_0.53O₂(x = 0) at each rate, especially at high rate of 5C. The average discharge capacities for samples with x = 0.01, x = 0.03 and x = 0.05 at 5C are 81.0 mAh·g⁻¹, 92.0 mAh·g⁻¹ and 71.1 mAh·g⁻¹, respectively, while that for samples with x = 0 only maintains a value of 65.3 mAh·g⁻¹. Fig. 6b illustrates the cycling performance at room temperature. It is clear that the cycling stability of the samples improve significantly when the Rb is incorporated into the lattice. After 50 cycles at 0.1C, the sample with x = 0.03 can still deliver a capacity of 222.1 mAh·g⁻¹ with capacity retention of 98.8%, while the sample with x = 0 only presented a specific capacity of 178.6 mAh·g⁻¹ with capacity retention of 94.3%. We also noticed the earlier work to improve the rate performance of such materials by the Al-doping method, where capacity of 198 mAh·g⁻¹ below 0.05C rate²⁸ for Li_1.27Cr_0.2Mn_0.53O₂ was reported. What's more, there is an increase of capacities during the first 10 cycles. It was demonstrated by the other researchers²⁹ that the capacity increase at the first several cycles was caused by the continuous activation of the initial Li-rich layered phase. Therefore, it seems that our approach, the Rb replacing method, is also an effective way to improve the rate capability and therefore improves the electrochemical performance.

Zoom In Zoom Out Reset image size

The whole pattern survey for Li_1.24Rb_0.03Cr_0.2Mn_0.53O₂ powder was presented in Fig. 7a, where peaks representing Rb, Cr, Mn, O, C ions were observed and denoted on the figure. Figs. 7b, 7c showed the XPS spectra of manganese ions (Mn 2P_3/2 and Mn 2P_1/2) after the charging and discharging process respectively. The Mn XPS results for the charging state showed two main peaks in the Mn 2p spectra, which can be assigned to the manganese 2p_3/2 at 641.56 eV and 2p_1/2 at 653.0 eV. The binding energy is well consistent with the data on MnO₂, indicating the existence of Mn⁴⁺ in the Li_1.27Cr_0.2Mn_0.53O₂. The binding energies for Mn 2P_3/2 and Mn 2P_1/2 at the end of discharge were 642.35 and 653.85 eV, respectively, which are very consistent with the binding energies for Mn⁴⁺ (641.8 eV) and (654.0 eV). Such results greatly implied that Mn⁴⁺ cations were not involved in the electrochemical process and remained inactive with the crystalline lattice.

Zoom In Zoom Out Reset image size

Figs. 7d, 7e showed the XPS spectra of chromium ions (Cr 2P_3/2 and Cr 2P_1/2) for the sample after charging and discharging process respectively. The XPS spectrum for sample after charging was dominated by the major peaks with binding energies of 577.85 and 586.97 eV, very close to those of the hexavalent chromium ions in the transition metal layer. Whereas the spectra for sample after discharging were composed of major peaks with binding energies 575.94 and 585.68 eV, showing the existence of chromium in the trivalent state. So, by comparison, it is clear that Cr³⁺/Cr⁶⁺ redox pair, not the Mn redox pair, dominate the electrochemical process in our approach.

XPS investigation on the electrode after discharging was shown in Fig. 7f. Besides the denoted peaks for the compound, the whole survey of the XPS spectra revealed another peak at binding energy around 110 eV. And the detailed information for such a peak was shown as the inset, showing two peaks centering at 110.94 eV and 109.41 eV. Such two binding energies are actually the binding energy for Rb ions, showing the existence of Rb ions in the de-intercalation sample. This means the Rb ions will not be de-intercalated during the electrochemical process, but stay quietly in its lattice sites and d0 not participate in the rocking movement.

In total, these XPS results suggest that, during the charging process, lithium ions were extracted from the lattice, followed by an oxidation of Cr³⁺ to Cr⁶⁺, while Mn ions were not oxidized beyond 4+. However, when the electrodes were discharged, the lithium ions were inserted back into the lithium layer, followed by a reduction of Cr⁶⁺ to Cr³⁺. Only Li cations and the Cr³⁺/Cr⁶⁺ redox pair were involved, Rb and Mn cations remained quietly during the rocking process.

In order to further explore the mechanism for the improved electrochemical performance by Rb⁺ incorporation, electrochemical impedance spectroscopy (EIS) measurements were carried out for the Rb_xLi_1.27-xCr_0.2Mn_0.53O₂ and Li_1.27Cr_0.2Mn_0.53O₂ electrodes as seen in Figs. 8a, 8b. Basically, all Nyquist plots consist of a semicircle in the high-to-medium frequency range and a sloping line in the low frequency region which respectively denotes the charge transfer resistance (Rct) between solid electrode/electrolyte interface and Warburg impedance (Wo) referring to solid-state Li⁺ diffusion in layered oxide.³⁰ Additionally, a very short intercept between semicircle and real axis in the high frequency region represents the solution resistance (Rs). In general, the semicircle can be mainly ascribed to the Li⁺ ion diffusion through the bulk electrode and the charge transfer reaction at the interface of the cathode-electrolyte, while the inclined line in the low frequency region can be assigned to Li⁺ diffusion (Warburg impedance) in the bulk electrode.^31,32 On the basis of equivalent circuit depicted in Fig. 8b; Among them, Rs represents the resistance of the solution, and Rct corresponds to the charge transfer resistance. CPE is related to the double-layer capacitance and passivation film capacitance, while W refers to Warburg impedance .The fitting results are listed in Table III. It is observed that all samples show small Rs values less than 7Ω for the sample cells before and after 50 cycles, indicating good stability of electrolyte composition during the lithiation/delithiation process. By contrast, large changes occurred for Rct values. It can be seen that after 50 cycles, the Rct of Rb_0.03Li_1.24Cr_0.2Mn_0.53O₂ cathode display only a two-fold increase from 84.7Ω to 192.1Ω, whereas the Rct increased dramatically from 102.3Ω to 680.6Ω for the pristine Li_1.27Cr_0.2Mn_0.53O₂ which suggests that the much easier Li⁺ diffusion can be obtained in an expanded lattice cell owing to Rb substitution, which are responsible for its best rate performance presented in Fig. 6.

Zoom In Zoom Out Reset image size

In addition, the diffusion coefficient of lithium ions (D_Li⁺) can be evaluated in terms of the following equation:^33,34

Where R is the gas constant, T the absolute temperature, c the concentration of Li⁺ in the material, F the Faraday constant, A the surface area of the electrode, and σ is the Warburg factor obeying the following relationship:^35,36

Where Z_Re is the real part of impedance, thus ω is the angular frequency at low frequency. The linear relationship of Z_Re and ω^−0.5 is shown in Fig. 9, the slope of the fitted straight line indicating the value of σ. The calculated values of σ and D_Li⁺ are recorded in Table IV. It is clearly seen that sample Rb_0.03Li_1.24Cr_0.2Mn_0.53O₂ exhibits the largest value of D_Li⁺, which is about one order of magnitude higher than that of sample Li_1.27Cr_0.2Mn_0.53O₂. This result is in good agreement with the variation trend of rate capability and cyclic performance, essentially accounting for the performance improvement of Li_1.27Cr_0.2Mn_0.53O₂ cathode material.

Zoom In Zoom Out Reset image size

From the above results, it can be found that the sample Rb_0.03Li_1.24Cr_0.2Mn_0.53O₂ presented the best electrochemical performance among the four samples. The sample showed advantage over the others on the initial capacity, columbic efficiency and rate performance and so on. All the results can be explained by the fact that a fast Li ion transportation was achieved since the largest Li ion diffusion coefficient was observed for the sample, as presented in Table IV. By careful comparison, the trend of the changes in diffusion coefficient is in good accordance with the changes in lattice parameter of c axis. The c values increase with the doping concentration but decrease when x reaches 0.05. Therefore, it seems in our case, the amount of Rb doped in the lattice determined the electrochemical performance: with the increase in Rb concentration, the c axis increases, the interlayer spacing increases, then the diffusion of Li⁺ is enhanced and finally the electrochemical properties is improved. However, when x reaches 0.05, the lattice distortion happened, c axis begins to shrink, the Li⁺ diffusion is restricted, and therefore the electrochemical performance is reduced.