Abstract
In the recent years, olivine LiFePO4 has been considered as a prospective cathode material for lithium-ion batteries. However, low conductivity is an obstacle to the commercialization of LiFePO4; doping the transition metal such as Mn and Ni is one of the solutions for this issue. This work aimed to synthesize the Mn-doped olivines LiMnxFe1−xPO4 at low content of Mn (x = 0.1, 0.2) via the hydrothermal route followed by pyrolyzed carbon coating. The synthesized olivines were well crystallized in olivine structure, with larger lattice parameters compared with LiFePO4. The EXD and TGA results confirmed the coated carbon of 4.14% for LiMn0.1Fe0.9PO4 and 6.86% for LiMn0.2Fe0.8PO4. Both of Mn-doped olivines showed higher diffusion coefficients of Li+ intercalation than those of LiFePO4 that led a good performance in the cycling test. LiMn0.2Fe0.8PO4 exhibited a higher specific capacity (160 mAh/g) than LiMn0.1Fe0.9PO4 (155 mAh/g), and the Mn content is beneficial for the cycling performance as well as ionic conductivity.
1. Introduction
In the last three decades, the world has witnessed urges for significant breakthroughs in the hi-tech industry. Among them, the demand for power storage can provide a large amount of energy with good cycling performance. Lithium-ion batteries (LIBs) have taken those requirements as their advantages. A great deal of research studies have delved deeply into the topic of LIBs in order to develop suitable materials to improve the performance of batteries. The use of cathode materials for LIBs has been an attractive field because they are the determining factor of power density, capacity, lifetime, and cost of the batteries [1, 2]. Besides the popular layered transition metal oxides, olivine materials LiMPO4 (M = Mn, Co, Fe, …), represented by LiFePO4, are competitive candidates to the commercialized LiCoO2. They are of low cost and easy to synthesize and nontoxic. High thermal stability and comparable specific capacity (170 mAh/g) to LiCoO2 (140 mAh/g) are the key to their large-scale application in the future [3, 4].
Olivine materials are categorized as orthorhombic structures with the space group Pnma. The lattice of LiFePO4 contains [FeO6] octahedral sites and [PO4] tetrahedral sites sharing vertices and edges. The octahedrons are connected to form a tunnel structure, which facilitates the migration of Li+ ions [5]. The alternate arrangement of [FeO6] and [PO4] groups, however, leads to low conductivity (∼10−9 S/cm at room temperature), which is the main limitation of LiFePO4 [6]. In order to solve this issue, many research projects have focused on minimizing the particle size to nanoscale with a modern synthetic process. Besides, doping the material with Mn or Ni (LiMnxFe1−xPO4 or LiNixFe1−xPO4) is also a promising solution [7]. Nakamura et al. reported that the Mn2+ substitution in olivine compounds (prepared by the solid-state reaction) expended the unit cell along the c-axis, which facilitated the Li immigration in doped olivines, and Li+ diffusions were found to be in the order of cm2/s [8]. Novikova et al. prepared the olivines LiMnxFe1−xPO4 (x = 0–0.4) by the sol-gel method, and the particles varied in the range of 400 nm to 4 μm. By using Mössbauer spectroscopy, they proved the low-content Mn doping (x = 0.1, 0.2) enhanced the discharge capacity to 142 mAh/g (at the current density of 55 mA/g) due to the solid solution formation under gradual changes in the potential of Mn [9].
Our work aimed to synthesize the Mn-doped olivine LiMnxFe1−xPO4 at low-content of Mn doping (x = 0.1, 0.2) via the hydrothermal route followed by pyrolyzed carbon coating. The structure and morphology of the synthesized olivines were evaluated by X-ray diffraction and FE-SEM coupling with EDX and TGA. We also considered the kinetics of Li+ intercalation (lithium diffusion coefficient, ) determined from cyclic voltammetry results, which are related to the cycling performance.
2. Methods
2.1. Preparation of LiMnxFe1−xPO4
Mn-doped olivines LiMn0.1Fe0.2PO4 and LiMn0.2Fe0.8PO4 were synthesized via the hydrothermal method. 150 mL of aqueous solution of Mn(CH3COO)2.4H2O and Fe(NO3)3.9H2O (at the molar ratio of 1 : 9 or 2 : 8) with a small amount of n,n-dimethylformamide C3H7NO as the reducing agent was stirred at 80°C for 1 hour. The solution was then mixed with LiOH.H2O, H3PO4, C6H8O6, and HNO3 and stirred thoroughly at ambient condition for 6 hours using a magnetic stirrer. The mixture was transferred into a hydrothermal system and annealed at 180°C for 12 hours. The outcome was separated from the suspension by the centrifuge and low-pressure filter, then washed by deionized water, and finally dried at 60°C overnight. In order to coat carbon on the surface of the synthesized olivines, the obtained solid was grounded with glucose at the mass ratio of 1 : 1 and then calcined at 700°C in N2/Ar atmosphere (95/5, v/v) for 3 hours.
2.2. Structural, Morphological, and Electrochemical Characterizations
The crystalline structure of the synthesized olivines was characterized by X-ray diffraction using D8-ADVANCED (Bruker, radiation) in the 2θ range of 10–70° at the scanning rate of 0.020°C/step. The morphology and elemental composition were evaluated by scanning electron microscopy (SEM) and SEM-EDX using a Hitachi FE-SEM S-4800 scanning electron microscope. The proportion of carbon coating on the particles was quantified by thermal analysis from 25°C to 800°C in air at the scanning rate of 10°C/min using a LABSYS evo TG–DSC 1600 system (Setaram, France).
The cathode slurry was prepared by mixing 90 wt% active materials, 7.5 wt% graphite, 7.5% acetylene black, and 5 wt% polyvinylidene fluoride-co-hexafluoropropylene (PVDF-HFP) in N-methyl-2-pyrrolidone (NMP). The slurry then was coated on the aluminum foil with a thickness of 50 μm and density of 2 mg/cm2. The olivine electrodes were cut with a diameter of 10 mm and dried at 130°C under vacuum for 24 hours. The two-electrode Swagelok-type cells were assembled in an Ar-filled glovebox, and the cell consisted of an olivines electrode as the positive electrode, lithium foil as the negative electrode, three glass Whatman microfibers as the seperator, and 100 mL of 1 M LiPF6 in a mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC) at a volumetric ratio of 2 : 1 as the electrolyte.
The electrochemical behaviors of Mn-doped olivines were evaluated by cyclic voltammetry (CV) and galvanostatic cycling test using the apparatus VSP (BioLogic, France). The cyclic voltammetry (CV) was carried out in a potential range of 2.5–4.5 V (versus Li+/Li) in the various rates from 20 µV/s to 120 µV/s. The cycling tests were conducted at rate C/10.
3. Results and Discussion
3.1. Structure and Morphology
The X-ray diffraction patterns of two Mn-doped olivines (Figures 1(a) and 1(b)) can be indexed in an orthorhombic structure with the space group Pnmb. Results of Rietveld refinement, gathered in Table 1, are in good agreement with the original olivine structure (JSPDF: 01-083-2092). The Li atoms are located at 4a sites; the Fe, Mn, and P atoms are in 4c sites; and the O atoms are in 8a sites. The fractions of Mn : Fe are 0.1 : 0.9 and 0.2 : 0.8, which approximately match the EDX results. The X-ray diffraction patterns indicate the presence of a pure and very well crystallized single phase without the impurity phase (e.g., MnO2, Li2O, and Fe2O3). It is obvious that partially replacing Fe with Mn amplifies the size of the olivines lattice unit, which can be explained by the fact that the ionic radius of Mn2+ (80 pm) is superior to that of Fe2+ (78 pm). This result agrees with previous reports about Mn-doped olivine [10–13].
(a)
(b)
Morphology of two Mn-doped olivines was analyzed by scanning electron microscopy. The hydrothermal preparation led to a homogeneous distribution. It can be seen (Figure 2) that the particle grains of both samples point out the polyhedral shapes at submicrometric scale (approximately 200–400 nm). According to EDX data in Figure 3 and Table 2, the elemental composition of Li, Fe, Mn, and P in the samples corresponds to the stoichiometric relationship which is suitable to the theoretical composition of LiMn0.1Fe0.9PO4 and LiMn0.2Fe0.8PO4.
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
Figure 4 shows the TEM images of two Mn-doped olivines, and we observed clearly that the particles’ size was in the range of 100-200 nm. The TEM images were coherent to SEM images. The Raman spectra of two Mn-doped olivines in Figure 5 show two fingerprint signals of carbon in the high-frequency region at 1385 cm−1 (D-band) and 1583 cm−1 (G-band) that confirm the composites between Mn-doped olivine and carbon LiMnxFe1−xPO4@C [14–17].
(a)
(b)
Thermal gravity analysis (TGA) is useful to quantify the proportion of coated carbon in Mn-doped samples. Figure 6 demonstrates the TGA curves of two Mn-doped olivines acquired in air at 10°C/min in the temperature range of 25–800°C. The increase in mass from 300°C to 400°C is caused by the oxidation of Fe and Mn elements in the samples to form metal oxides. At a temperature higher than 400°C, the mass of each sample shows a steady decline that results from the reaction between carbon and oxygen [18]. The amount of carbon in each sample is the difference in mass percentage measured at 400°C and 800°C. Figure 6 infers that the carbon content in LiMn0.1Fe0.9PO4 (6.86%) is higher than that in LiMn0.2Fe0.8PO4 (4.14%).
3.2. Electrochemical Behaviors
The Li insertion into the original olivine phase is described by the domino-cascade mechanism between two phases LiFePO4 and FePO4 with the characteristic voltage at 3.45 V (vs. Li+/Li), corresponding to the redox couple Fe3+/Fe2+. The Mn doping into olivine can add a signal of redox couple Mn3+/Mn2+ around 4 V, and the intensity of signal depends on the Mn content in olivine, in the potential range of 2.5–4.5 V with the scan rate s. Figures 7(a) and 7(b) demonstrate the cyclic voltammograms of LiMn0.1Fe0.9PO4 and LiMn0.2Fe0.8PO4 in various scan rates from 20 to 120 μV/s. The strong and shaped peaks at ∼3.5 V in all CV curves are evidenced as the original redox reaction, while the signals at 3.8–4.0 V can be assigned to the doping element reaction. The higher the scanning rate, the higher the peak separation and intensity of LiMn0.2Fe0.8PO4; the Mn signal looks more significant that proposes more capacity contribution in the galvanostatic test.
(a)
(b)
(c)
During the Li intercalation process, Li+ ions migrated following the 1D channels along the (010) plane in the orthorhombic structure; thus, Li mobility (also called kinetic of intercalation process) affects the electrode performance in the galvanostatic cycling test and habitually determined through the diffusion coefficient (). In our case, kinetics of Li intercalation process into the Mn-doped olivines was assessed by cyclic voltammetry. The scattered plot of cathodic peak current versus square root of scan rates was built up based on CV results and presented in Figure 7(c). Diffusion coefficients of the Li+ insertion process () were determined by using the Randles–Sevcik equation:where A is the surface area of the cathode (0.785 cm2); C is the concentration of Li+ ions in the material (cm3/mol); and n is the number of transferred electron (n = 1). Figure 7(c) shows a linear proportion between and with the correlation coefficient . values calculated from the Randles–Sevcik equation are listed in Table 3. value of LiMn0.2Fe0.8PO4 ( cm2/s) is four times higher than that of LiMn0.1Fe0.9PO4 ( cm2/s). According to previous publications, diffusion coefficients of LiFePO4 were obtained between cm2/s [15, 19, 20]. Therefore, partial Mn substitution in the olivine structure benefits the ionic conductivity through the soar of , compared with the original olivine phase.
Cycling performances of two Mn-doped olivines were executed at rate C/10 in the voltage range of 2.5–4.5 V (vs. Li+/Li). Figure 8 compares the first galvanostatic curve of two doping olivines which demonstrate two Li insertion regions: (i) well-defined plateau at 3.5 V and (ii) the short one at 3.7–4.0 V. We observed the Mn-couple capacity contribution in LiMn0.2Fe0.8PO4 (∼30 mAh/g) was higher than its in LiMn0.1Fe0.9PO4 (∼10 mAh/g). These results are coherent with the CV′ results. The discharged capacities are found out to be 158 mAh/g for LiMn0.1Fe0.9PO4 and 160 mAh/g for LiMn0.2Fe0.8PO4.
As shown in Figure 9(a), the discharge curves of LiMn0.1Fe0.9PO4 practically superimpose without any hysteresis, and the capacity is stable at 158 mAh/g upon 50 cycles. In case of LiMn0.2Fe0.8PO4, the Mn capacity contribution is faded cycle by cycle; after 50 cycles, the Mn capacity region at 3.7 V seems to disappear, leading to the decline of discharge capacity to 140 mAh/g. The fading capacity is interpreted by the release of Mn during the charge-discharge process. The Coulombic efficiencies of over 97% (Figure 9(c)) are retained during the cycling test, which indicates a reversible Li intercalation into the olivine hosts.
(a)
(b)
(c)
(d)
The rate capability performance of Mn-doped olivines is demonstrated in Figure 9(d), in which current density was set up from C/10 to 5C (1C corresponds to 165 mAh/g). The Mn-doped olivines show a stable capacity, for instance LiMn0.1Fe0.9PO4, 155, 147, 139, 130, 115, and 87 mAh/g at C/10, C/5, C/2, 1C, 2C, and 5C; an initial capacity of 155 mAh/g is remained at C/10 after 30 cycles. The discharge capacity of LiMn0.2Fe0.8PO4 still seems to be higher, but a slightly fading capacity is also observed. The rate capability performance compares very well with the previous reports on Mn-doped olivines [10, 13].
4. Conclusions
In brief, two olivine-type materials LiMn0.1Fe0.9PO4 and LiMn0.2Fe0.8PO4 were successfully synthesized via the hydrothermal route, which present an orthorhombic structure, and their lattice parameters acceded to the preceding reports. Doping Mn can increase the diffusion of Li+ ions in two orders of magnitude as compared with the original phase LiFePO4. Two Mn-doped olivines showed a stable cycling performance upon 50 cycles as well as rate capability. For the next step, the influence of Mn2+ on the activation energy of Li intercalation will be deeply considered.
Data Availability
The data used to support the findings of this study are included within the article.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This work was supported by Vietnam National University Ho Chi Minh (VNUHCM) through grant number C2018-18-11.