Raman spectral signature of Mn-rich nanoscale phase segregations in carbon free LiFe_1−xMn_xPO₄ prepared by hydrothermal technique

Abstract

Mn-rich nanoscale secondary phases were identified in LiFe_1−xMn_xPO₄, despite the known complete solubility for the LiFePO₄–LiMnPO₄ system and the observed linear increase in the lattice parameters of LiFe_1−xMn_xPO₄ with increasing Mn concentration. Carbon free LiFe_1−xMn_xPO₄ (x = 0, 0.05, 0.10, 0.25) were prepared by the sequential precipitation of Li₃PO₄ and (Fe_1−xMn_x)₃(PO₄)₂, followed by hydrothermal treatment. At low doping concentration (x ≤ 0.05), Li–Mn–O secondary phases were discerned by Raman spectra, which corroborated with the inductively coupled plasma elemental analysis. Though energy dispersive elemental mapping with scanning transmission electron microscopy do not show segregation of Mn at low concentrations, Mn-rich phases were clearly discerned at high doping concentration (x = 0.25). The kinetics of Mn-rich phase formation during hydrothermal synthesis of carbon free LiFe_1−xMn_xPO₄, which was attributed to the difference in the solubility constant of the intermediate products of Li₃PO₄ and (Fe_1−xMn_x)₃(PO₄)₂, and its implications on the capacity of LiFe_1−xMn_xPO₄ cathode material were discussed. Our results present how de-convoluted Raman peaks show clear signatures of nanophase impurity segregations and how an increase in the lattice constant with Mn doping concentration can be decisive.

---

1. Introduction

Phospho-olivines (LiMPO₄; M = Fe, Mn, Co) and their doped derivatives are intensively studied due to their application as cathodes in lithium ion batteries for electric and hybrid electric vehicles. The LiMPO₄ structure has good thermal and chemical stability.¹ However, some of the main limitations of these compounds are their low intrinsic electronic conductivity (<10⁻⁹ S cm⁻¹) and low Li-ion diffusion coefficient (10⁻¹¹ to 10⁻¹³ cm² s⁻¹), which hinder their application in high power batteries.² To circumvent these drawbacks, several approaches, including carbon coating, reduction in particle size, and doping with other transition metals, have been extensively explored.^1,3

A mixed metal solid solution of olivine compounds is one such possibility that can considerably enhance electrochemical characteristics such as power density and specific capacity.⁴ Recently, significant research activities exploring various combinations of metal cations in LiMPO₄ have been reported.⁵ The main focus of such studies includes finding compounds with optimized electrochemical properties similar to the studies on the optimization of layered compounds,^5a which led to the famous LiMn_1/3Co_1/3Ni_1/3O₂ (1/3 1/3 1/3) cathode materials.⁶ LiFe_1−xMn_xPO₄ is one of the various solid solutions that is being considered as potential cathode material.⁷ The possible enhancement in the mobility of lithium along the b-axis in low concentration Mn doped LiFePO₄ is attributed to widening of 1D-channels along the b-axis and an increase in the diffusion coefficient.⁸ The local distribution of cations in the olivine structure, therefore, is very crucial in determining the electrochemical properties. The phospho-olivine structure of LiMPO₄ (M = Fe, Mn, Co) is composed up of hexagonally close packed oxygen arrays, in which transition metal ions occupy the corner-shared M2 octahedral site, lithium ions occupy the edge-shared M1 octahedral site and phosphorus ions are located in tetrahedral sites.⁹ Li-ion diffusion is energetically favored along the b-axis (i.e. 1D transport) and can be hindered due to the blockage of 1D pathway by the presence of defects and impurities.¹⁰ Furthermore, formation of Mn-rich impurity phases can hamper the diffusion of lithium ions. These structural factors are significantly influenced by the synthesis method.¹¹

LiFePO₄ and its derivatives have been synthesized by various solid state and wet chemical techniques. Among these, hydrothermal method is considered as economical and viable for large-scale production. A review on the synthesis of phospho-olivine compounds is reported by Jugović et al.¹¹ It has been found that hydrothermal process parameters, such as temperature, duration, pH of the solution, starting chemicals and the sequence of mixing, greatly influence the microstructure, defects, impurities, and local structural features, which in turn affect the electrochemical characteristics.¹¹ In the majority of the structural studies on LiFe_1−xMn_xPO₄, Mn substitution at the Fe site has been shown by a systematic increase in the lattice parameters.¹² However, the presence of Fe_Li anti-site defect (excess Fe occupying the Li site) is shown to be frozen-in due to the Mn substitution and does not seem to get removed even at high temperature or extended annealing conditions.¹³ Further, the solvothermal preparation of LiFe_1−xMn_xPO₄ is shown to contain mixed phases of LiFePO₄ and LiMnPO₄ instead of the LiFe_1−xMn_xPO₄ solid solution.¹⁴ While the presence of defects or phase segregation can be investigated using state-of-the-art sophisticated characterization techniques, they cannot be used in quality control for mass production. Moreover, when secondary phases exist in very small quantity or in an amorphous form, it is a challenging task to identify the secondary phases present. A combination of characterization techniques, such as FTIR and Raman spectra, magnetometry, elemental mapping,¹⁵ and overall elemental concentration as identified by inductively coupled plasma¹⁶ are required in identifying the secondary phases.¹⁷

In this report, we present the effect of Mn doping on the structural properties of LiFePO₄ synthesized by hydrothermal technique. Despite the observation of a linear increase in the lattice parameter with Mn substitution, we show using Raman and Fourier transform infrared spectroscopy the signatures of Mn-rich secondary phases [Li₂MnO₃ or LiMnO₂ and Mn₃(PO₄)₂] in LiFe_1−xMn_xPO₄, which is further substantiated by several techniques, including X-ray diffraction (XRD), magnetometry, inductively coupled plasma-optical emission spectrometry (ICP-OES), high-resolution transmission electron microscopy (HRTEM) and scanning transmission electron microscope-energy dispersive spectroscopy (STEM-EDS) elemental mapping. These Mn-rich phases are identified in small concentrations as nano-sized impurity phases present within the crystalline lattice of hydrothermally synthesized LiFe_0.75Mn_0.25PO₄. We further discuss the implications of such Mn-rich inclusions on the electrochemical properties.

2. Experimental

Co-precipitation or sequential precipitation methods are commonly used to prepare multicomponent phases. In this study, we used a sequential method by beginning with the required ratio of metal cation to anions in the final component. Carbon free LiFe_1−xMn_xPO₄ (x = 0, 0.05, 0.1, 0.25) was synthesized by hydrothermal technique using chemicals LiOH·H₂O, FeSO₄·7H₂O, (NH₃)₂HPO₄ and MnSO₄·H₂O. An aqueous (NH₃)₂HPO₄ solution (1 M) was added to an aqueous solution of LiOH·H₂O (1 M) with constant stirring under a N₂ atmosphere giving rise to a white precipitate. To this white precipitate, a 1 M solution mixture of xFeSO₄·7H₂O + (1 − x) MnSO₄·H₂O was added, which resulted in the formation of a green colored thick gel. The pH of the final solution was 7.5 and after stirring for 30 minutes, the homogenous solution was transferred to a Teflon autoclave and heated for 10 h at 200 °C. The precipitate formed after hydrothermal treatment was filtered and washed several times with de-ionized water until it was free of sulphate ions. The final filtrates were dried under vacuum in a desiccator. For this study four samples of LiFe_1−xMn_xPO₄ with different Mn concentration, x = 0, 0.05, 0.1, and 0.25, were synthesized and further characterized for structural properties. The phase formation in the as-prepared samples was confirmed using an X-ray powder diffractometer (X'pert-Pro, PANalytical) with a 1.2 kW Cu Kα (λ = 1.5406 Å). To understand the effect of Mn doping on the formation of the impurity phases, Fourier transform infrared spectroscopy (FTIR) was performed using a Perkin-Elmer spectrometer (Spectrum One) in the range from 400 to 4000 cm⁻¹ after dispersing the samples in an anhydrous KBr pellet. Furthermore, the presence of impurities was studied using Raman spectroscopy (Horiba Jobin-Yvon, HR 800UV) in the range between 200 and 1500 cm⁻¹ using a 488 nm Ar⁺-ion laser source. The magnetic behavior of the samples was studied using a physical property measurement unit (PPMS, Quantum Design) operating under the VSM mode. The isothermal magnetization studies were carried out with a magnetic field varying in the range from +7 to −7 T. Susceptibility studies under a constant magnetic field were studied from 5 K to 300 K. Elemental analyses were carried out using ICP-OES. The morphology of the prepared samples was observed using a Hitachi S-4300 field emission scanning electron microscope (FE-SEM, Hitachi Co. Ltd. S-4300) operating at 20 kV. To prevent charging of the samples, the samples were coated with a thin gold layer by a sputtering process for 1 min. Transmission electron microscopy (TEM) was performed using a Technai G2 F30 TEM and a Titan G2 80-200 TEM with a probe corrector. Energy dispersive X-ray spectroscopy (EDS) was conducted with a Titan G2 80-200 TEM using a Super-X EDX detector. EDS maps were constructed using Fe-Kα, Mn-Kα, O-Kα, and P-Kα. For the electrochemical characterization, cathodes were prepared by mixing active material, acetylene black as a conducting carbon, and polyvinylidene fluoride as binder at a weight ratio of 80 : 10 : 10, respectively, using N-methyl pyrrolidone as a solvent. The resulting slurry was coated uniformly on Al foil, dried at 120 °C for 12 h and disk electrodes were cut. Coin cells with lithium as counter electrode were fabricated in an argon filled glove box with less than 1 ppm oxygen and moisture. 1 M LiPF₆ was used as electrolyte dissolved in ternary solvent, ethylene carbonate (EC) and dimethyl carbonate (DMC) with the ratio of 50 : 50. Galvanostatic charge and discharge studies were performed at a current of 0.1 mA with cut-off voltages of 2.2–4.4 V at room temperature using Arbin instruments (Model: BT2000).

3. Results

Fig. 1 shows the morphology of LiFe_1−xMn_xPO₄ (x = 0, 0.05, 0.1, 0.25) samples studied by FESEM. All the samples showed well faceted, oriented crystallites with a uniform size distribution. The XRD of the as-prepared carbon free LiFe_1−xMn_xPO₄ (x = 0, 0.05, 0.1, 0.25) powder samples (Fig. 2a) reveals the formation of a phospho-olivine phase where all the peaks were indexed according to the standard ICDD pattern 98-003-5125. The lattice parameters and the unit cell volume show a linear increase with Mn doping concentration, as shown in Fig. 2b, which was in accordance to the reported change in lattice parameters with Mn doping.¹² Few of the very low intensity peaks at 2θ = 14.6°, 25.32°, 31.76°, and 35.09°, seen clearly in the x = 0 sample, correspond to the sarcopside phase of Fe₃(PO₄)₂. This sarcopside phase¹⁸ is equivalent to phospho-olivine with no Li cations. Furthermore, one half of the Li sites (M1) in the phospho-olivine structure is occupied by Fe or other ions of transition metals, such as Mn, Cr, Co, and the other half of M1 site is vacant, i.e. M₃(PO₄)₂ can be visualized as M_0.5MPO₄ (M = Fe, Mn, and Cr). With Mn doping and increase in the concentration of Mn substitution, no noticeable changes in the intensity of XRD peaks corresponding to sarcopside [Fe₃(PO₄)₂/(Fe_1−yMn_y)₃PO₄, x ≠ y] was observed. The discharging capacity (not shown) of LiFe_1−xMn_xPO₄ was found to be approximately 60 mA h g⁻¹, which was low compared to the theoretical capacity of LiFePO₄ (170 mA h g⁻¹) at 0.1 C rate. It should be noted, in general, that due to poor electronic and ionic conductivity of phospho-olivine compounds the capacity exhibited by the cathode would be lower than the theoretical capacity. This is generally improved by substantial size reduction and achieving proper conductive coating on the nanoparticles. In our samples, the particle size was large (∼1 μm) and there was no carbon coating. However, to understand and address phase-segregation related decrease in capacity of LiFe_1−xMn_xPO₄, various structural characterizations were carried out. To investigate the presence of possible amorphous phases and lithium deficiency, ICP-OES was carried out and the Li : Fe : Mn : P ratios for these samples are listed in Table 1. Elemental analysis of undoped LiFePO₄ sample suggests that compared to the stoichiometric LiFePO₄, the sample contains an excess of 18 at% of phosphor and an excess of 24 at% of iron with respect to lithium. This excess concentration of Fe and P corresponds to ∼10 mol% of Fe₃(PO₄)₂ in LiFePO₄. However, very low intensity of Fe₃(PO₄)₂ peaks in XRD indicates that not all excess Fe and P could exist as Fe₃(PO₄)₂, and part of the excess Fe can also reside at the Li site in LiFePO₄. With the incorporation of Mn, though phosphor concentration decreases, there was no change in the Li concentration, suggesting the formation Li–Mn–O phases. The absence of any X-ray reflections corresponding to Li–Mn–O phases suggests the possible presence of the amorphous phase or very low concentration of nanocrystalline Li–Mn–O phases.

Fig. 1 Field emission scanning electron micrograph of LiFe_1−xMn_xPO₄ where (a) x = 0, (b) x = 0.05, (c) x = 0.10, and (d) x = 0.25.

Fig. 2 (a) X-ray diffraction pattern of (x = 0, 0.05, 0.10, 0.25). (b) The unit cell volume as a function of Mn concentration (x).

Table 1 The compositional results of LiFe_1−xMn_xPO₄ samples from ICP-OES analyses

Nominal composition of the sample	Elemental concentration from ICP-OES
	Li	Fe	Mn	P
LiFePO4	0.76	1	0	0.94
LiFe0.95Mn0.05PO4	0.67	0.92	0.08	0.56
LiFe0.90Mn0.1PO4	0.73	0.85	0.15	0.58
LiFe0.75Mn0.25PO4	0.65	0.72	0.28	0.55

---

Magnetometric measurements¹⁹ were carried out to detect the presence of small concentrations of magnetic impurities such as Fe₂P or Fe₂O₃. Fig. 3a and b shows the temperature-dependent inverse magnetic susceptibility of LiFe_1−xMn_xPO₄ samples depicting the antiferromagnetic-paramagnetic transition with Néel temperature around 50 ± 2 K. The μ_eff calculated from the paramagnetic region^17b of χ⁻¹ vs. temperature plot is given in Table 2. Mn substitution in LiFePO₄ increases μ_eff from 5.3 μ_B to 5.8 μ_B (Fig. 3d). The change in μ_eff is similar to the trend reported by Yamada et al.^17b suggesting the contribution from spin and angular momentum. The temperature dependent magnetic susceptibility reveals the absence of magnetic transitions due to other phases, particularly the antiferromagnetic transition at ∼120 K characteristic of FeP and a ferromagnetic transition at ∼265 K of Fe₂P impurities. This confirms that the hydrothermally synthesized LiFe_1−xMn_xPO₄ are free of FeP and Fe₂P impurities.²⁰ The absence of magnetic impurities, such as Fe₂P or γ-Fe₂O₃, was further confirmed by the linear dependence of magnetization on magnetic field measured at various temperatures (Fig. 3c).

Fig. 3 (a) Magnetization as a function of the magnetic field, (b) inverse magnetic susceptibility vs. temperature plots, (c) magnified plot of (b) in the low temperature regime, and (d) the change in μ_eff and Neel temperature as a function of Mn (x = 0–0.25) of LiFe_1−xMn_xPO₄.

Table 2 Neel temperature and effective magnetic moment of LiFe_1−xMn_xPO₄ samples

x in LiFe1−xMnxPO4	TN (K)	μeff
0	50.55	5.32 μB
0.5	53.07	5.70 μB
0.10	52.35	5.62 μB
0.25	50.87	5.77 μB

---

Though the XRD patterns and magnetometry analyses suggest close to the phase purity of the LiFe_1−xMn_xPO₄ sample except for (Fe_1−xMn_x)₃PO₄, the compositional analysis shows significant offset from the expected Li : Fe : Mn : P stoichiometry. This indicates the presence of other possible impurity phases mainly comprising of amorphous or nanocrystalline forms of Li–Mn–O phase. Raman spectroscopy can be used to identify such phases, as it is very sensitive, down to the molecular level (Fig. 4). However, one should be cautious and aware of the decomposition of LiFePO₄ due to laser heating, posing a challenge to the phase analysis.²¹ In this study, we used a 488 nm laser with a power density of 1 mW per 20 μm² to avoid the decomposition of the sample while acquiring the spectra and further ascertained all the vibrations were characteristic of the LiFe_1−xMn_xPO₄ sample investigated. Because the relative intensity of Raman spectral modes at ∼950 cm⁻¹ is very high, and the features below 800 cm⁻¹ are small, for clarity the Raman spectra of LiFe_1−xMn_xPO₄ from wavenumber 100 to 800 cm⁻¹ and 800 to 1300 cm⁻¹ are given in separate figures Fig. 4a and b, respectively. A vibrational spectrum of LiFePO₄ is generally classified into internal modes corresponding to vibrations in the range 400–1200 cm⁻¹ and external modes from 100 to 400 cm⁻¹. The relative intensity of the vibrational features match with Raman spectral data reported by Paraguassu et al.,²² and there were no discernible changes due to Mn substitution in LiFePO₄ in the external modes of vibration. However, we observed a clear influence of Mn doping on internal modes, which originated due to the intramolecular vibrations of the PO₄³⁻ polyanions. To see this effect, the bending modes of PO₄³⁻ peaks are fitted to mixed Gaussian and Lorentzian shape profiles and the corresponding fitted plots in the range from 400 to 700 cm⁻¹ are given in Fig. 4c. The broad mode appearing at ∼440 cm⁻¹ can be fitted into two distinct peaks for pure LiFePO₄; however, the best fit in Mn doped LiFePO₄ could be obtained with either 3 or 4 peaks. The vibrational mode at 405 cm⁻¹ is attributed to the external mode of LiFePO₄, whereas the mode at ∼440 cm⁻¹ is due to the overlapping bending modes of LiFePO₄. The Raman bands at 573, 591 and 629 cm⁻¹ are due to bending modes of PO₄³⁻. From the fitted peak, it is evident that the intensity of the Raman signal at 556 cm⁻¹ (shown with arrow) increases with Mn concentration. We attribute the additional bands at 426 and 556 cm⁻¹, that are observed only with Mn substitution in LiFePO₄, to vibrations corresponding to Li–Mn–O phases.²³ Although it is difficult to comment on the specific phase that gives rise to these vibrational modes, such modes are characteristic of Li–Mn–O phases, as reported in the previous studies.²⁴ The Raman spectra of pure and Mn substituted LiFePO₄ samples in the range of 900–1200 cm⁻¹, have a strong peak at ∼947 cm⁻¹ due to the symmetric stretching mode and peaks at 997 and 1066 cm⁻¹ due to the asymmetric stretching mode of PO₄³⁻. The low intensity shoulder peaks at 926 and 971 cm⁻¹, whose relative intensity increases with an increasing in Mn concentration, are attributed to modes from Fe₃(PO₄)₂/Mn₃(PO₄)₂/(Fe_1−xMn_x)₃PO₄.²⁵

Fig. 4 Raman spectra of LiFe_1−xMn_xPO₄ (x = 0, 0.05, 0.10, 0.25), (a) in the 100–800 cm⁻¹ range, (b) in the 800–1300 cm⁻¹ range, and (c) the peak fit for the Raman spectra in the 350–700 cm⁻¹ range. The peaks marked with # corresponds to Fe₃(PO₄)₂/Mn₃(PO₄)₂/(Fe_1−xMn_x)₃PO₄ and * corresponds to Li–Mn–O phases.

The IR spectral modes of LiFePO₄ have been reported to show a significant shift in the stretching and bending modes of PO₄³⁻ due to the presence of Fe_Li antisite defects.²⁶ Such Fe_Li antisite defects, which accommodate Li deficiency, have been more commonly reported in the hydrothermally prepared LiFePO₄ samples. The Li deficiency found from ICP analysis of LiFe_1−xMn_xPO₄ also signify the possible presence of such Fe_Li antisite defects. To investigate how LiFe_1−xMn_xPO₄ accommodates lower atomic concentrations of Li from stoichiometric LiFePO₄ and low Li and P atomic concentrations in Mn doped LiFePO₄, FTIR spectroscopy of these samples was carried out (Fig. 5a). The positions of the vibrational modes of all the LiFe_1−xMn_xPO₄ match well the reported values.²⁶ We did not find any change in the vibrational modes in FTIR spectra except for a small increase in the ν₁ vibrational mode positions with increase in the Mn concentration. As depicted in Fig. 5b, ν₁ mode can be resolved into two bands, one ∼940 cm⁻¹ and the second ∼972 cm⁻¹ at x = 0. A shift to high frequency of these band was observed with an increase in the Mn concentration (Fig. 5c), which was probably due to the increase in the Fe_Li antisite defect concentration.²⁶ Shifting of the vibration bands to higher frequency has been reported earlier by Bini et al.²⁷ with Mn substitution in LiFePO₄.

Fig. 5 FTIR spectra of LiFe_1−xMn_xPO₄ (x = 0, 0.05, 0.10, 0.25), (a) in the 400–1300 cm⁻¹ range, (b) the peak fit for the FTIR spectra in the 800–1200 cm⁻¹ range. (c) The deconvoluted peak position of ν₁ absorption mode in b as a function of ‘x’.

To investigate the local structural details and elemental distribution, we carried out HRTEM and STEM-EDS elemental mapping analysis. The representative bright field image, HRTEM and elemental mappings are shown in Fig. 6 and 7 for LiFe_1−xMn_xPO₄ samples with x = 0.05 and x = 0.25, respectively. The HRTEM image of one of the crystallites is shown in Fig. 6b. The compositional analyses studied by EDS-mapping clearly depict the homogeneous distribution of all the elements across the crystal (Fig. 6d) in LiFe_0.95Mn_0.05PO₄. However, STEM investigations of LiFe_0.75Mn_0.25PO₄ reveal Mn-rich regions within the crystallites. The aerial mapping of Fe, Mn, P and O of a part of a crystallite is shown in Fig. 7d depicting the Mn-rich region of ∼50 nm. It should be noted that in this region, though there was a decrease in Fe concentration, phosphor concentration is uniform throughout the crystal. The line-profile obtained from the region marked in Fig. 7d is given in Fig. 7c and clearly shows the local elemental variations leading to Mn-rich regions within the LiFe_1−xMn_xPO₄ crystals. These local structural and elemental characterisations obtained from STEM-EDS mapping substantiate the observations from Raman and FTIR that with increasing Mn concentration in LiFePO₄, the Mn-rich nanoscale phase precipitates within the LiFe_1−xMn_xPO₄ phase.

Fig. 6 (a) Bright-field TEM image and (b) high resolution transmission electron microscopy (HRTEM) image of LiFe_1−xMn_xPO₄ with x = 0.05 sample. (c) The elemental mapping of Fe, Mn, O and P from the region shown in the STEM image.

Fig. 7 (a) Bright-field TEM image and (b) high resolution transmission electron microscopy (HRTEM) image of LiFe_1−xMn_xPO₄ with x = 0.25 sample. The inset in (b) shows the SAD pattern from the crystallite. (c) STEM-EDS line-profile and (d) the elemental mapping of Fe, Mn, O and P of the crystal from the region marked by the black line in STEM image.

4. Discussion

From the abovementioned studies, it is clear that hydrothermal conditions and sequential precipitation conditions that were used in this study, also favored the formation of small concentrations of Li–Mn–O and nanosized Mn-rich sarcopside phase segregation within the lattice of LiFe_1−xMn_xPO₄. Therefore, it is evident from the IR, Raman and STEM investigations of LiFe_1−xMn_xPO₄ that even when there is an increase in the lattice constant with Mn concentration, the presence of impurity phases or segregation of secondary phases within the lattice of LiFePO₄ should not be ruled out.

In the sequential precipitation of LiFe_1−xMn_xPO₄, when (NH₃)₂HPO₄ was added to LiOH, the pH was 11 and under this pH Li₃PO₄ precipitated as surmised from its white color.^12,14 The addition of Fe precursor to this mixture reduces the pH to 7.5 and favors the formation of Fe₃(PO₄)₂ along with Li₃PO₄ with the molecular level mixing of these two phases in gel form. In earlier investigations using the sequential precipitation method, an excess Li molar ratio, three times higher than PO₄ concentration, was used.²⁸ Due to this, PO₄³⁻ anions were almost completely utilized for the formation of Li₃PO₄, and only a very small quantity of Fe₃(PO₄)₂ was formed when Fe²⁺ solution was added. The concentration of Li₃PO₄ and Fe₃(PO₄)₂ is governed by the solubility limit of the two products. The solubility product (K_sp) at room temperature is of the order of 1.0 × 10⁻³⁶ mol⁵ l⁻⁵ for Fe₃(PO₄)₂, which is considerably smaller than 3.2 × 10⁻⁹ mol⁴ l⁻⁴ for Li₃PO₄, and therefore, the formation of Fe₃(PO₄)₂ is preferred to that of Li₃PO₄.^11,28 In the present investigation, in contrast to earlier studies, equal concentrations of Li, Fe, and P are used, and therefore it is expected to have a molecular level mixture of both Fe₃(PO₄)₂ and Li₃PO₄ in the as-prepared precipitate. The formation of LiFePO₄ from Fe₃(PO₄)₂ can be understood from the close structural similarities between LiFePO₄ and Fe₃(PO₄)₂. The detailed crystal structure similarities are discussed by Moore.¹⁸ The two structures can be visualized as made up of hexagonal close packing of oxygen. In LiFePO₄ the M2 octahedral sites are occupied by Fe, M1 octahedral sites are occupied by Li and tetrahedral sites correspond to PO₄. The analogy of the Fe₃PO₄ structure as sarcopside with olivine LiFePO₄ can be visualized by writing Fe₃(PO₄)₂ as Fe_0.5FePO₄. In sarcopside, half of the M2 sites are occupied by 0.5 Fe atoms leaving half of the M2 sites empty compared to the olivine phase, and the remaining Fe occupies M1 sites and P occupies tetrahedral sites, very similar to the olivine phase. The unit cell lattice parameters are also very close to each other. Therefore, the only difference in the structure of olivine and sarcopside is the occupation of M2 sites. During hydrothermal synthesis, due to the larger solubility product, Li₃PO₄ dissociates to Li⁺ and (PO₄)³⁻, and Li migrates to empty M2 sites of sarcopside and also partially replace the remaining Fe in M2 sites of Fe₃(PO₄)₂ to form LiFePO₄. Thus, the reduction in the Li⁺ concentration in the solution further drives the dissociation of Li₃PO₄. However, it appears that complete replacement of Fe in M1 sites by Li does not take place, which results in the presence of small fractions of Fe₃(PO₄)₂ in the final product. This also leads to a significant loss of Li (∼25 at%) and P (∼5 at%) as identified by ICP-OES.¹⁴

In LiFe_1−xMn_xPO₄ sequential precipitations, similar to the LiFePO₄ preparation, a 1 M solution of (NH₃)₂HPO₄ was added to a 1 M solution of LiOH to form Li₃PO₄. To this mixture, when Mn and Fe precursors are introduced, (Fe_1−xMn_x)₃(PO₄)₂ was formed along with Li₃PO₄ and these two products reacted to give LiFe_1−xMn_xPO₄ during the hydrothermal treatment. By ICP analysis, it was observed that in Mn substituted samples, though the Li concentration remains the same as that of undoped LiFePO₄, the phosphor concentration decreases significantly. This can be understood from the slightly higher solubility product of Mn₃(PO₄)₂ [1 × 10⁻²⁷ mol⁵ l⁻⁵] than that of Fe₃(PO₄)₂ and smaller than Li₃PO₄. When Mn is also present along with Fe, Li gets incorporated into (Fe_1−xMn_x)₃(PO₄)₂ leading to LiFe_1−xMn_xPO₄, very similar to LiFePO₄. At high concentrations, this also leads to local precipitation of Mn-rich regions, probably Mn₃(PO₄)₂, as within the phospho-olivine phase as discerned from STEM studies and further substantiated by Raman and FTIR spectral studies. The dissolved Mn, P, and Li ions in the hydrothermal solutions also favor the precipitation of small concentration of Li–Mn–O, which are identified by Raman and IR spectroscopy. The segregation of secondary phases within the lattice of LiFePO₄ hinders the intercalation and deintercalation of Li⁺ through the lattice.

5. Conclusions

Carbon free LiFe_1−xMn_xPO₄ (x = 0, 0.05, 0.1, 0.25) is synthesized using hydrothermal treatment of Li₃PO₄ and the (Fe_1−xMn_x)₃(PO₄)₂ gel is obtained by sequential addition of Fe to Li and P solution in Li : Fe : P ratio of 1 : 1 : 1. Predominantly phase pure LiFe_1−xMn_xPO₄ is obtained as evidenced from XRD. However, very dilute concentrations of the Mn-rich impurity phases [Li–Mn–O and (Fe_1−yMn_y)₃(PO₄)₂] are discerned by FTIR, Raman spectroscopy and elemental analysis by ICP. These Mn-rich regions are clearly shown to exist as nanoscale segregations within the lattice of LiFe_1−xMn_xPO₄ for x = 0.25. It is difficult to identify the nanoscale segregation from XRD and TEM at low Mn concentrations, but their signature can be identified from FTIR and Raman spectroscopy. Such nanoscale lattice segregations in LiFe_1−xMn_xPO₄ can arise during hydrothermal process, wherein Li from Li₃PO₄ diffuses into the structurally similar phospho-olivine like (Fe_1−xMn_x)₃(PO₄)₂ lattice. Formation of the LiFe_1−xMn_xPO₄ from (Fe_1−xMn_x)₃(PO₄)₂ can be further understood in terms of the difference in the solubility limits of constituents and crystal structure similarities between LiFe_1−xMn_xPO₄ and (Fe_1−xMn_x)₃(PO₄)₂. The nanoscale lattice segregations and associated localized disorder in LiFe_1−xMn_xPO₄ will significantly affect the Li diffusion coefficient and charge capacity in the electrochemical properties.

Acknowledgements

We would like to acknowledge the financial support received from the Department of Science and Technology, (DST-SERC) Government of India (no. 1R/S3/EU/0001/2011). The authors are thankful to Director, ARCI, for his constant encouragement and discussion during the course of this work. The authors would like to acknowledge Dr Ranjith Ramadurai, IIT Hyderabad, for the magnetic measurements and Prof. K. Hono, Director, Magnetic Materials Unit, National Institute for Materials Science for extended TEM-STEM characterization facilities.

References

1. W.-J. Zhang, Structure and performance of LiFePO₄ cathode materials: A review, J. Power Sources, 2011, 196(6), 2962–2970.
2. X. C. Tang, L. X. Li, Q. L. Lai, X. W. Song and L. H. Jiang, Investigation on diffusion behavior of Li⁺ in LiFePO₄ by capacity intermittent titration technique (CITT), Electrochim. Acta, 2009, 54(8), 2329–2334.
3. L.-X. Yuan, Z.-H. Wang, W.-X. Zhang, X.-L. Hu, J.-T. Chen, Y.-H. Huang and J. B. Goodenough, Development and challenges of LiFePO₄ cathode material for lithium-ion batteries, Energy Environ. Sci., 2011, 4(2), 269–284.
4. G. Hautier, A. Jain, H. Chen, C. Moore, S. P. Ong and G. Ceder, Novel mixed polyanions lithium-ion battery cathode materials predicted by high-throughput ab initio computations, J. Mater. Chem., 2011, 21(43), 17147–17153.
5. (a) T. Muraliganth and A. Manthiram, Understanding the Shifts in the Redox Potentials of Olivine LiM_1−yM_yPO₄ (M = Fe, Mn, Co, and Mg) Solid Solution Cathodes, J. Phys. Chem. C, 2010, 114(36), 15530–15540; (b) R. Malik, F. Zhou and G. Ceder, Phase diagram and electrochemical properties of mixed olivines from first-principles calculations, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79(21), 214201; (c) J. Molenda, A. Kulka, A. Milewska, W. Zając and K. Świerczek, Structural, Transport and Electrochemical Properties of LiFePO₄ Substituted in Lithium and Iron Sublattices (Al, Zr, W, Mn, Co and Ni), Materials, 2013, 6(5), 1656–1687; (d) D.-H. Seo, H. Gwon, S.-W. Kim, J. Kim and K. Kang, Multicomponent Olivine Cathode for Lithium Rechargeable Batteries: A First-Principles Study, Chem. Mater., 2009, 22(2), 518–523.
6. (a) Y.-U. Park, J. Kim, H. Gwon, D.-H. Seo, S.-W. Kim and K. Kang, Synthesis of Multicomponent Olivine by a Novel Mixed Transition Metal Oxalate Coprecipitation Method and Electrochemical Characterization, Chem. Mater., 2010, 22(8), 2573–2581; (b) H. Gwon, D.-H. Seo, S.-W. Kim, J. Kim and K. Kang, Combined First-Principle Calculations and Experimental Study on Multi-Component Olivine Cathode for Lithium Rechargeable Batteries, Adv. Funct. Mater., 2009, 19(20), 3285–3292; (c) K.-W. Nam, X.-J. Wang, W.-S. Yoon, H. Li, X. Huang, O. Haas, J. Bai and X.-Q. Yang, In situ X-ray absorption and diffraction studies of carbon coated LiFe₁/4Mn₁/4Co₁/4Ni₁/4PO₄ cathode during first charge, Electrochem. Commun., 2009, 11(4), 913–916.
7. J. Kim, K.-Y. Park, I. Park, J.-K. Yoo, J. Hong and K. Kang, Thermal stability of Fe–Mn binary olivine cathodes for Li rechargeable batteries, J. Mater. Chem., 2012, 22(24), 11964–11970.
8. K. T. Lee and K. S. Lee, Electrochemical properties of LiFe_0.9Mn_0.1PO₄/Fe₂P cathode material by mechanical alloying, J. Power Sources, 2009, 189(1), 435–439.
9. A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, Phospho-olivines as Positive-Electrode Materials for Rechargeable Lithium Batteries, J. Electrochem. Soc., 1997, 144(4), 1188–1194.
10. C. A. J. Fisher, V. M. H. Prieto and M. S. Islam, Lithium battery materials LiMPO₄ (M = Mn, Fe, Co, and Ni): Insights into defect association, transport mechanisms, and doping behavior, Chem. Mater., 2008, 20(18), 5907–5915.
11. D. Jugović and D. Uskoković, A review of recent developments in the synthesis procedures of lithium iron phosphate powders, J. Power Sources, 2009, 190(2), 538–544.
12. M. Yoncheva, V. Koleva, M. Mladenov, M. Sendova-Vassileva, M. Nikolaeva-Dimitrova, R. Stoyanova and E. Zhecheva, Carbon-coated nano-sized LiFe_1−x Mn_xPO₄ solid solutions (0 ≤ x ≤ 1) obtained from phosphate–formate precursors, J. Mater. Sci., 2011, 46(22), 7082–7089.
13. K. M. Ø. Jensen, M. Christensen, H. P. Gunnlaugsson, N. Lock, E. D. Bøjesen, T. Proffen and B. B. Iversen, Defects in Hydrothermally Synthesized LiFePO₄ and LiFe_1−xMn_xPO₄ Cathode Materials, Chem. Mater., 2013, 25(11), 2282–2290.
14. W. Huang, S. Tao, J. Zhou, C. Si, X. Chen, W. Huang, C. Jin, W. Chu, L. Song and Z. Wu, Phase Separations in LiFe_1–xMn_xPO₄: A Random Stack Model for Efficient Cathode Materials, J. Phys. Chem. C, 2013, 118(2), 796–803.
15. X. Zhou, F. Wang, Y. Zhu and Z. Liu, Graphene modified LiFePO₄ cathode materials for high power lithium ion batteries, J. Mater. Chem., 2011, 21(10), 3353–3358.
16. R. Amin, P. Balaya and J. Maier, Anisotropy of Electronic and Ionic Transport in LiFePO₄ Single Crystals, Electrochem. Solid-State Lett., 2007, 10(1), A13–A16.
17. (a) C. M. Burba and R. Frech, Raman and FTIR Spectroscopic Study of Li_xFePO₄ (0 ≤ x ≤ 1), J. Electrochem. Soc., 2004, 151(7), A1032–A1038; (b) A. Yamada, Y. Takei, H. Koizumi, N. Sonoyama, R. Kanno, K. Itoh, M. Yonemura and T. Kamiyama, Electrochemical, Magnetic, and Structural Investigation of the Li_x(Mn_yFe_1−y)PO₄ Olivine Phases, Chem. Mater., 2006, 18(3), 804–813.
18. P. B. Moore, Sarcopside: Its atomic arrangement, Am. Mineral., 1972, 57, 24–35.
19. (a) C. M. Julien, A. Mauger, A. Ait-Salah, M. Massot, F. Gendron and K. Zaghib, Nanoscopic scale studies of LiFePO₄ as cathode material in lithium-ion batteries for HEV application, Ionics, 2007, 13(6), 395–411; (b) K. Zaghib, N. Ravet, A. Mauger, M. Gauthier, J. B. Goodenough and C. Julien, LiFePO₄ High Electrochemical Performance at 60 °C with Purity Controlled by SQUID Magnetometry, ECS Trans., 2007, 3(27), 119–129.
20. K. Zaghib, A. Mauger, F. Gendron and C. M. Julien, Magnetic studies of phospho-olivine electrodes in relation with their electrochemical performance in Li-ion batteries, Solid State Ionics, 2008, 179(1–6), 16–23.
21. Y. Bai, Y. Yin, J. Yang, C. Qing and W. Zhang, Raman study of pure, C-coated and Co-doped LiFePO₄: thermal effect and phase stability upon laser heating, J. Raman Spectrosc., 2011, 42(4), 831–838.
22. W. Paraguassu, P. T. C. Freire, V. Lemos, S. M. Lala, L. A. Montoro and J. M. Rosolen, Phonon calculation on olivine-like LiMPO₄ (M = Ni, Co, Fe) and Raman scattering of the iron-containing compound, J. Raman Spectrosc., 2005, 36(3), 213–220.
23. C. M. Julien and M. Massot, Lattice vibrations of materials for lithium rechargeable batteries III. Lithium manganese oxides, Mater. Sci. Eng., B, 2003, 100(1), 69–78.
24. (a) C. M. Julien, F. Gendron, A. Amdouni and M. Massot, Lattice vibrations of materials for lithium rechargeable batteries. VI: Ordered spinels, Mater. Sci. Eng., B, 2006, 130(1–3), 41–48; (b) T.-J. Kim, D. Son, J. Cho and B. Park, Enhancement of the electrochemical properties of o-LiMnO₂ cathodes at elevated temperature by lithium and fluorine additions, J. Power Sources, 2006, 154(1), 268–272.
25. L. Zhang and R. K. Brow, A Raman Study of Iron–Phosphate Crystalline Compounds and Glasses, J. Am. Ceram. Soc., 2011, 94(9), 3123–3130.
26. X. Qin, J. Wang, J. Xie, F. Li, L. Wen and X. Wang, Hydrothermally synthesized LiFePO₄ crystals with enhanced electrochemical properties: simultaneous suppression of crystal growth along [010] and antisite defect formation, Phys. Chem. Chem. Phys., 2012, 14(8), 2669–2677.
27. M. Bini, M. C. Mozzati, P. Galinetto, D. Capsoni, S. Ferrari, M. S. Grandi and V. Massarotti, Structural, spectroscopic and magnetic investigation of the LiFe_1−xMn_xPO₄ (x = 0–0.18) solid solution, J. Solid State Chem., 2009, 182(7), 1972–1981.
28. M. H. Lee, T. H. Kim, Y. S. Kim and H. K. Song, Precipitation Revisited: Shape Control of LiFePO₄ Nanoparticles by Combinatorial Precipitation, J. Phys. Chem. C, 2011, 115(25), 12255–12259.

---