Enhanced Cycling Stability through Erbium Doping of LiMn₂O₄ Cathode Material Synthesized by Sol-Gel Technique.

In this work, LiMn2-xErxO₄ (x ≤ 0.05) samples were obtained by sol-gel processing with erbium nitrate as the erbium source. XRD measurements showed that the Er-doping had no substantial impact on the crystalline structure of the sample. The optimal LiMn1.97Er0.03O₄ sample exhibited an intrinsic spinel structure and a narrow particle size distribution. The introduction of Er3+ ions reduced the content of Mn3+ ions, which seemed to efficiently suppress the Jahn⁻Teller distortion. Moreover, the decreased lattice parameters suggested that a more stable spinel structure was obtained, because the Er3+ ions in a ErO₆ octahedra have stronger bonding energy (615 kJ/mol) than that of the Mn3+ ions in a MnO₆ octahedra (402 kJ/mol). The present results suggest that the excellent cycling life of the optimal LiMn1.97Er0.03O₄ sample is because of the inhibition of the Jahn-Teller distortion and the improvement of the structural stability. When cycled at 0.5 C, the optimal LiMn1.97Er0.03O₄ sample exhibited a high initial capacity of 130.2 mAh g-1 with an excellent retention of 95.2% after 100 cycles. More significantly, this sample showed 83.1 mAh g-1 at 10 C, while the undoped sample showed a much lower capacity. Additionally, when cycled at 55 °C, a satisfactory retention of 91.4% could be achieved at 0.5 C after 100 cycles with a first reversible capacity of 130.1 mAh g-1.

1. Introduction

With increasing environmental awareness, many people have realized the importance of green travel, which is very useful for reducing environmental pollution and protecting human health. As an optimal choice for green travel, electric vehicles with rechargeable batteries have become very popular all over the world. Meanwhile, lithium-ion batteries, as the power source, have been developed quickly in recent years [1,2,3,4,5,6,7,8]. It is generally known that there are four major classes of mature cathode materials, namely LiCoO2 [9,10], LiFePO4 [11,12], LiNi1−x−yCoxMyO2 (M = Mn, Al) [13,14], and LiMn2O4 [15,16], for batteries. Among these materials, LiMn2O4 shows many virtues such as mature production technology, cheap production costs, non-pollution characteristics, and so forth [17,18,19,20]. However, the large-scale commercial applications of this material have been seriously restricted because of its poor cycling life and high-temperature performance, which are mostly a consequence of Jahn–Teller distortion, manganese dissolution, and non-uniform particle-size distribution [7,21,22,23,24]. Therefore, there is a tremendous need to optimize this material for better performance.

Among the existing, numerous solutions, many researchers generally prefer surface modification and cation doping [17,25,26,27,28,29,30]. The surface modification can make a positive contribution to the improvement of the cycling life by sealing off the active material from the electrolyte corrosion to suppress the manganese dissolution. However, this strategy cannot fundamentally inhibit the Jahn–Teller distortion [25,26]. Thus, the optimization effect of surface modification is limited. Therefore, many researchers have chosen to use the cation doping strategy to optimize the cycling life of LiMn2O4 [31,32,33]. Yu et al. [34] prepared the Li1+xMn2−xO4 samples by a solid-state sintering method. The obtained Li1.06Mn1.94O4 sample with extra Li showed better cycling performance, because the introduction of lithium ions can weaken the ordering of lithium ions and enhance the structural stability of the sample. Huang et al. [31] reported the synthesis of LiCuxMn2−xO4 by a low-temperature molten-salt combustion method, and their results showed that the Cu-doping can optimize the average particle size and size distribution. More importantly, the Cu substitution significantly improved the cycling performance. Furthermore, the LiAlxMn2−xO4 samples synthesized by a solution combustion technique showed better cycling life as a result of the reduced Jahn-Teller distortion by Al-doping [35]. These results indicated that the cycling life of LiMn2O4 could be improved by doping with other cations. Such optimization effects have been corroborated well by others [32,36,37]. In addition, the synthetic method could have a significant influence on the electrochemical properties. So far, the LiMn2O4 cathode materials have been obtained by a solid-state method [38,39], hydrothermal method [40,41], combustion method [31,42,43], co-precipitation method [44], sol-gel method [45,46,47], and so on. Among these methods, the sol-gel method is highly suitable to prepare the high-performance, doped LiMn2O4 due to the following advantages: (1) the reactants are evenly mixed at the molecular level in the process of forming a gel; (2) uniform cation doping can be achieved due to the intimate mixing of chemical contents in a solution; and (3) the chemical reaction is carried out easily, because the diffusion of the components is on a nanometer scale, which requires relatively low synthetic temperature.

In this work, the LiMn2−xErxO4 (x ≤ 0.05) samples were successfully synthesized by the sol-gel technique with erbium nitrate as the erbium source. The influence of the erbium-doping content on the structures, morphologies, and electrochemical properties of the LiMn2−xErxO4 samples obtained by the sol-gel process is discussed. The optimal Er-doped LiMn2O4 sample showed the intrinsic spinel structure and narrow particle size distribution. More importantly, this sample exhibited excellent cycling stability, superior high-rate capability, and outstanding high-temperature performance.

2. Materials and Methods

The LiMn2−xErxO4 samples were obtained by a sol-gel process with erbium nitrate (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) as the erbium source. Figure 1 shows the schematic illustration of the synthesis of these Er-doped LiMn2O4 samples. Firstly, stoichiometric lithium hydroxide (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) and citric acid (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) were weighed to prepare the mixed solution. Under vigorous stirring, a mixed solution of erbium nitrate and manganese acetate (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) was added dropwise into the abovementioned solution at 50 °C. After continuous stirring for 30 min, NH3·H2O (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) was added dropwise into the mixed solution to adjust the pH to 8.0, and the temperature was subsequently adjusted to 80 °C. After further stirring for a few hours, a reddish-brown sol was formed, which was then dried at 110 °C. The obtained dried gel was calcined at 450 °C for 4 h and then further heated at 780 °C for 12 h.

Figure 1

Schematic illustration of the synthesis of the LiMn2−xErxO4 (x = 0, 0.01, 0.03, 0.05) samples.

The crystal structures of the obtained erbium-doped spinels were studied by X-ray diffraction (XRD, Bruker DX-1000, Karlsruhe, Germany) with Cu Kα radiation (λ = 0.15406 nm). The lattice parameters of these erbium-doped samples were obtained by using MDI Jade 5.0 software. The surface morphologies and microstructures were determined by using scanning electron microscopy (SEM, JEOL JSM-6360LV, Tokyo, Japan) with an energy dispersive X-ray spectrometer (EDX, EDAX Inc., Mahwah, NJ, USA). X-ray photoelectron spectroscopy (XPS) was obtained by using a Thermo ESCALAB 250XI instrument (Thermo Fisher Scientific, Waltham, MA, USA) with a monochromatic Al Ka (1486.6 eV) X-ray source.

The active electrode consisted of the obtained erbium-doped spinels, conductive acetylene black, and polyvinylidene fluoride (weight ratio = 85:10:5). The anode material and diaphragm were lithium foil and Celgard 2400 polymer (Celgard, Charlotte, NA, USA), respectively. A mixture of 1 M of LiPF6, ethyl methyl carbonate (EMC), dimethyl carbonate (DMC), and ethylene carbonate (EC) was used as the electrolyte (VEMC:VDMC:VEC = 1:1:1, (Guangzhou Tinci Materials Technology Co., Ltd., Guagnzhou, China)). The electrochemical measurements were carried out on a NEWARE battery testing system (NEWARE, Shenzhen, China). The electrochemical impedance spectroscopy (EIS) was carried out by using a CS-350 electrochemical workstation (Wuhan Corrtest Instruments Crop., Ltd., Wuhan, China). The impedance plots were recorded by applying an AC (alternating current) voltage of 5 mV amplitude in the frequency range of 0.1–100 kHz.

3. Results and Discussion

Figure 2 presents the XRD results of the LiMn2−xErxO4 (x = 0, 0.01, 0.03, 0.05) samples obtained by the sol-gel method. As shown in Figure 2a, all the Er-doped LiMn2O4 samples showed the obvious characteristic diffraction peaks of spinel-type lithium manganese oxide (JCPDS No. 35-0782), suggesting that the introduction of a small amount of erbium ions did not have detectable influence on the material’s structure [35,42]. All the LiMn2−xErxO4 (x = 0.01, 0.03, 0.05) samples maintained the inherent spinel structure of LiMn2O4. According to the previously reported results [21,45] and according to the reported references [21,48], the (220) peak of LiMn2O4 is particularly sensitive to the other cations at tetrahedral sites (8a). If the doped ions inhabit the tetrahedral sites, the (220) peak should appear in the corresponding XRD pattern. However, the (220) peak cannot be observed in the XRD patterns in all the LiMn2−xErxO4 (x = 0, 0.01, 0.03, 0.05) samples. This indicates that the erbium ions replaced the manganese ions at the octahedral sites in the Er-doped LiMn2O4 samples.

Figure 2

XRD results of the LiMn2−xErxO4 (x = 0, 0.01, 0.03, 0.05) samples: (a) the consecutive XRD patterns and (b) the representative magnified XRD patterns.

Table 1 lists the corresponding crystal parameters of these samples. The lattice parameters of these erbium-doped samples were obtained by using MDI Jade 5.0 software. It is obvious from these data that all the Er-doped LiMn2O4 samples possessed a Fd-3m space group. As the Er-doping content increased, the LiMn2−xErxO4 (x = 0.01, 0.03, 0.05) samples showed smaller lattice parameters and cell volumes. Figure 2b shows the magnified map of the (111), (311), and (400) peaks. It can be clearly seen that the introduction of erbium ions caused a shift toward the higher angle, which further indicated the decrease of the crystal parameters. These results suggest the formation of a more stable spinel structure [20,49]. This is principally because the Er3+ ions in the ErO6 octahedra showed stronger bonding energy (615 kJ/mol) than that of the Mn3+ ions in the MnO6 octahedra (402 kJ/mol) [50]. In addition, it should be noted that the Er-doped LiMn2O4 samples showed higher (311)/(400) peak intensity ratios, which have much to do with the cycling life of LiMn2O4 [21,51]. An analysis of the previously published results indicated that the introduction of erbium ions may play a constructive role in enhancing the electrochemical properties.

Table 1

Crystal parameters calculated from the XRD patterns of the LiMn2−xErxO4 samples.

Sample	Space	a (nm)	Volume (nm3)	I311/I400
LiMn2O4	Fd-3m	0.82334	0.55813	0.9054
LiMn1.99Er0.01O4	Fd-3m	0.82291	0.55725	0.9513
LiMn1.97Er0.03O4	Fd-3m	0.82177	0.55494	0.9917
LiMn1.95Er0.05O4	Fd-3m	0.82053	0.55244	0.9983

Figure 3 presents the SEM images of the LiMn2−xErxO4 (x = 0, 0.01, 0.03, 0.05) samples obtained by the sol-gel technology. It can be clearly observed that the introduction of the erbium ions had a certain influence on the surface morphology of the powders. For the undoped LiMn2O4 particles shown in Figure 3a, the particle size distribution was unsatisfactory because of severe particle agglomeration. By contrast, all the Er-doped LiMn2O4 particles (Figure 3b–d) showed relatively good surface morphology with relatively little particle agglomeration. When the Er-doping content increased, the mean diameter of the LiMn2−xErxO4 (x = 0.01, 0.03, 0.05) samples showed a decreasing tendency. In particular, the LiMn1.97Er0.03O4 particles shown in Figure 3c presented the most uniform size distribution, which is conducive to the enhancement of cycling life [21,28,32]. These results indicate that the introduction of erbium ions can effectively optimize the size distribution, which contributes to the improvement of the cycling stability. Figure 4 shows the SEM-EDX pattern and SEM-mapping results of the LiMn1.97Er0.03O4 sample as a representative sample of the Er-doped LiMn2O4 samples. The SEM-EDX pattern shown in Figure 4a can confirm the successful incorporation of erbium ions in the doped LiMn2O4 samples. As shown in Figure 4b–d, the SEM-mapping results show the homogeneous distribution of the manganese, erbium, and oxygen elements in the Er-doped LiMn2O4 samples.

Figure 3

SEM images of the LiMn2−xErxO4 samples: (a) x = 0; (b) x = 0.01; (c) x = 0.03; and (d) x = 0.05.

Figure 4

(a) SEM-EDX pattern and (b–d) SEM-mapping results of the LiMn1.97Er0.03O4 sample.

Figure 5 shows the XPS spectra of Li1s, Mn2p, Er4d, and O1s in the LiMn1.97Er0.03O4 sample, which was selected as a representative sample of the Er-doped LiMn2O4 samples. The binding energy peaks of the Li1s, Mn2p, and O1s are well shown in Figure 5a,b,d and coincide with the previous reported literature [21]. It is important to note that the Mn2p3/2 binding energy of the manganese element was at 642.4 eV. However, according to the existing literature [36,52], the Mn2p3/2 binding energies of the trivalent and tetravalent manganese ions are at 641.7 eV and 643.1 eV, respectively. Thus, it can be inferred that the manganese element in the LiMn1.97Er0.03O4 sample corresponded to the coexistence state of the trivalent and tetravalent manganese ions. As for the erbium element, the binding energy peak shown in Figure 5c corresponded to the oxidation states for Er4d, which was assigned to Er3+ at 168.8 eV, which agrees with the previous result [53].

Figure 5

XPS spectra of Li1s, Mn2p, Er4d, and O1s in the LiMn1.97Er0.03O4 sample.

Figure 6a presents the first discharge curves of these samples, which were tested at 0.5 C. All the Er-doped LiMn2O4 samples showed similar characteristic discharge curves to that of the undoped spinel. There were two distinct voltage platforms around 4.15 V and 4.00 V, suggesting that the introduction of the erbium ions did not change the electrochemical redox reaction mechanism, as all the LiMn2−xErxO4 samples had two extraction/insertion steps of Li+ ions [29,43]. Figure 6b presents the cycling life of the LiMn2−xErxO4 samples. The cycling life of the LiMn2−xErxO4 (x = 0.01, 0.03, 0.05) samples was significantly improved as the erbium-doping amount increased because of the inhibition of the Jahn-Teller distortion and the improvement of the structural stability. Note, however, that the introduction of more erbium ions had a harmful effect on the reversible capacity of the LiMn1.95Er0.05O4 sample because of the reduction of the trivalent manganese ions. Figure 6c shows the comparison plots of the initial discharge capacities and capacity retentions of these samples. We can clearly observe the positive influence on the capacity retention and the adverse effect on the discharge capacity. These results indicate that introducing an appropriate amount of erbium ions can play an active role in enhancing the cycling life of a sample. Figure 6d presents the long cycling life of the undoped LiMn2O4 and LiMn1.97Er0.03O4 samples. For the optimal LiMn1.97Er0.03O4 sample, the initial reversible capacity could exhibit 130.2 mAh g−1. After 100 cycles, this sample exhibited 123.9 mAh g−1 with an outstanding retention of 95.2%. However, the undoped LiMn2O4 sample showed a poor cycling life with low reversible capacity after the 100th cycle. In particular, the undoped LiMn2O4 sample only delivered 93.7 mAh g−1 with a lower retention of 67.8% after 100 cycles. In addition, we compared the cycling performance of the LiMn1.97Er0.03O4 sample with that of the other doped samples, as shown in Table 2. It can be found that the erbium-doped LiMn2O4 sample show good cycling performance. These analyses further confirm the improvement of the cyclic stability by introducing some appropriate erbium ions into the spinel structure.

Figure 6

(a) Initial discharge curves and (b) cycling performance of the LiMn2−xErxO4 (x = 0, 0.01, 0.03, 0.05) samples; (c) comparison plots of the initial discharge capacities and capacity retentions; and (d) long cycling performance of the LiMn2−xErxO4 (x = 0, 0.03) samples.

Table 2

Comparison of the various doped-LiMn2O4 samples described in the references including this work.

Sample	Synthesis Method	Initial Capacity and Capacity Retention	Ref.
Li(Li0.06Mn1.94)O4	Solid-state method	123.0 mAh g−1, 86.6% after 200 cycles at 0.5 C (25 °C)	[34]
Li(Zn0.05Mn1.95)O4	Solution combustion method	102.6 mAh g−1, 82.9% after 500 cycles at 1.0 C (25 °C)	[20]
Li(Mg0.08Mn1.92)O4	Solid-state combustion synthesis	101.3 mAh g−1, 98.1% after 40 cycles at 0.2 C (ambient temperature)	[43]
Li(Cu0.05Mn1.95)O4	Molten-salt combustion method	119.0 mAh g−1, 95.0% after 100 cycles at 0.5 C	[31]
Li(Al0.06Mn1.94)O4	Co-precipitation method	117.4 mAh g−1, 97.0% after 100 cycles at 1.0 C (55 °C)	[44]
Li(Cr0.05Mn1.95)O4	Citric acid-assisted combustion method	117.0 mAh g−1, 81.2% after 100 cycles at 0.5 C	[55]
Li(Co0.02Mn1.97)O4	Controlled crystallization method	116.8 mAh g−1, 91.0% after 350 cycles at 296 mA g−1	[56]
Li(Ti0.03Mn1.97)O4	Solid-state method	135.7 mAh g−1, 95.0% after 70 cycles at 0.5 C (room temperature)	[38]
Li(Si0.05Mn1.97)O4	Solid-state method	134.6 mAh g−1, 85.1% after 100 cycles at 0.5 C (room temperature)	[29]
Li(Zr0.02Mn1.97)O4	Solid-state method	113.8 mAh g−1, 95.5% after 50 cycles at 0.2 C (25 °C)	[57]
Li(Er0.03Mn1.97)O4	Sol-gel method	130.2 mAh g−1, 95.2% after 100 cycles at 0.5 C (room temperature)	This work

For the practical application of LiMn2O4, the rate performance is an important factor. The undoped LiMn2O4 and Er-doped LiMn1.97Er0.03O4 samples were tested successively at different rates. Figure 7a shows the corresponding discharge curves of the LiMn1.97Er0.03O4 samples. It can be seen that there were two voltage platforms, which were obvious at 0.2 C (the red color) and 0.5 C, suggesting the diffusion process of the lithium ions [20,33]. When the rate was further increased, these two potential plateaus gradually showed ambiguous boundaries and shifted toward the lower voltage when the cycling rate increased. This result has a lot to do with the polarization effect and ohmic drop [45,54]. Furthermore, when the cycling rate recovered to 0.2 C (the saffron yellow color), it was found that the LiMn1.97Er0.03O4 sample could show similar discharge capacity compared with the initial discharge capacity at 0.2 C (the red color), suggesting the excellent restorative performance of the LiMn1.97Er0.03O4 sample. Figure 7b shows the cycling stability of the undoped LiMn2O4 and the optimal LiMn1.97Er0.03O4 samples at varying rates. When cycled at 0.2 C, the capacities of these two samples reached up to 140.5 and 133.2 mAh g−1, respectively. However, what is important to pay attention to is the reversible capacity of the Er-doped LiMn2O4 sample. With the increasing of the cycling rate, these two samples can show much more different results. In particular, when cycled at 10 C, the LiMn1.97Er0.03O4 showed 80.7 mAh g−1, while the LiMn2O4 samples only showed 20.7 mAh g−1.

Figure 7

(a) Representative discharge curves of the LiMn1.97Er0.03O4 samples and (b) cycling performance of the LiMn2−xErxO4 (x = 0, 0.03) samples at varying rates.

Figure 8 shows the cycling performance of the LiMn2O4 and LiMn1.97Er0.03O4 samples at 10 C. As shown in Figure 8a, the high rate shows a greater negative impact on the characteristic voltage plateaus at around 4.15 and 4.0 V, respectively. For the LiMn1.97Er0.03O4 sample, these two voltage plateaus become blurred to a certain extent. What is worse, the LiMn2O4 sample presented a lower voltage plateau, and the capacity of the LiMn2O4 sample showed severe degradation. Figure 8b presents the cycling life of these two samples at 10 C. It can be found that the initial discharge capacity of the undoped LiMn2O4 sample only reached to 32.5 mAh g−1 with a poor capacity retention of 81.5%. By contrast, the optimal LiMn1.97Er0.03O4 sample displayed a higher discharge capacity of 83.1 mAh g−1. The discharge capacity still showed 78.0 mAh g−1 with an excellent capacity retention of 93.9%. These results suggest that the high-rate performance of LiMn2O4 can be enhance by doping manganese ions with erbium ions in the spinel structure.

Figure 8

(a) Initial discharge curves and (b) cycling performance of the LiMn2−xErxO4 (x = 0, 0.03) samples at 10 C.

Figure 9a presents the cycling stability of the undoped LiMn2O4 and LiMn1.97Er0.03O4 samples at 55 °C. It can be seen from Figure 7a that the initial capacity of the LiMn1.97Er0.03O4 sample could reach up to 130.1 mAh g−1 at 0.5 C. Moreover, this sample still maintained a high capacity of 118.9 mAh g−1 with an excellent retention of 91.4% after 100 cycles. Unfortunately, the undoped LiMn2O4 sample showed very poor high-temperature cycling performance. After 100 cycles, the undoped sample only showed a lower capacity of 62.5 mAh g−1 with a low-capacity retention of 45.3%. These results suggest that introducing erbium ions can be favorable for enhancing the high-temperature performance of such a sample. Figure 9b shows the rate capability of these two samples at 55 °C. As shown here, the undoped LiMn2O4 and LiMn1.97Er0.03O4 samples showed similar capacities at low rates. However, these two samples presented obvious differences with the increasing of the rates. When cycled at 10 C, the LiMn1.97Er0.03O4 sample could exhibit 78.2 mAh g−1, while the LiMn2O4 sample only showed 18.3 mAh g−1. Based on these results, it can be concluded that the introduction of erbium ions can improve the high-temperature rate performance of LiMn2O4.

Figure 9

(a) Cycling performance and (b) rate capacities of the LiMn2−xErxO4 (x = 0, 0.03) samples at 55 °C.

Figure 10a,b show the EIS results of the undoped LiMn2O4 and LiMn1.97Er0.03O4 samples. As shown here, the high-frequency semicircle represents the charge transfer resistance (R2), which is closely related to the cycling life [21,48]. Thus, the effect of doping manganese ions with erbium ions on the cycling stability was mainly studied. The fitting values of R2 are listed in Table 3. For the LiMn1.97Er0.03O4 sample, the original R2 value only reached 73.4 Ω cm2 but increased to 115.1 Ω cm2 after 100 cycles. The R2 value increase was relatively small with a low growth rate of 56.8%. However, the undoped sample only showed the unsatisfactory R2 value. It can be seen that the undoped spinel showed a higher original R2 value (118.3 Ω cm2). After 100 cycles, the high growth rate reached up to 149.5% with the 100th R2 value of 295.2 Ω cm2. These results indicate that the addition of erbium ions in the spinel structure can have a positive role in decreasing the R2 value and enhancing the diffusion of lithium ions, which is conducive to the improvement of cycling stability [29,32].

Figure 10

Nyquist plots of the LiMn2O4 (a) and LiMn1.97Er0.03O4 (b) samples before cycling and after 100 cycles.

Table 3

Fitting values of the charge transfer resistance (R2) calculated from EIS.

Sample	R2 (Ω cm2) before Cycles	R2 (Ω cm2) after 100 Cycles	Percentage of Increase
LiMn2O4	118.3	295.2	149.5%
LiMn1.97Er0.03O4	73.4	115.1	56.8%

4. Conclusions

In summary, we have successfully used the sol-gel technology to prepare the Er-doped LiMn2O4 samples. All these samples maintained the spinel structure of LiMn2O4 and showed relatively even particle size distribution. The optimal LiMn1.97Er0.03O4 sample showed a better cycling performance. When tested at 0.5 C, this sample delivered a reversible capacity of 130.2 mAh g−1 with an excellent retention of 95.2% after 100 cycles. At higher rate of 10 C, the reversible capacity of the LiMn1.97Er0.03O4 sample peaked at 83.1 mAh g−1, which is far higher than that of the undoped spinel. Moreover, this sample showed outstanding cycling stability at higher temperatures. All of these results indicate that the introduction of erbium ions could enhance the cycling stability of LiMn2O4.