Magnetic Properties of Li₃V₂(PO₄)₃/Li₃PO₄ Composite

Abstract

Here, we present the investigation of the magnetic properties of Li₃V₂(PO₄)₃/Li₃PO₄ composites, which can be potentially used as a cathode material in lithium-ion batteries. Li₃V₂(PO₄)₃/Li₃PO₄ was synthesized by the thermal hydrolysis method and has a granular mesoporous structure. Magnetic properties of the composite were investigated using magnetometry and electron spin resonance methods. Based on magnetization measurements, the simultaneous existence of the paramagnetic phase with antiferromagnetic interactions between spins of V³⁺ ions and magnetically correlated regions was suggested. Most probably, magnetically correlated regions were formed due to anti-site defects and the presence of V⁴⁺ ions that was directly confirmed by electron spin resonance measurements.

1. Introduction

The interest in alternative energy sources has been increasing every year over the past two decades. Among the different energy alternatives, the conversion into chemical energy is the most convenient, and rechargeable batteries are the most popular energy storage devices. Currently, one can see the exciting developments in new positive electrode (cathode) materials for these batteries including nickel-rich layered oxides, lithium-rich layered oxides, high-voltage spinel oxides, and high voltage polyanionic compounds [1,2]. Among them, the monoclinic Li₃V₂(PO₄)₃ with promising electrochemical properties including excellent cycling stability, high theoretical capacity, low synthetic cost, improved safety characteristic, and low environmental impact has emerged as a highly suitable candidate for use in lithium-ion batteries (LIB) [3,4]. Recently, significant achievements in electrochemical performance in Li₃V₂(PO₄)₃ have been achieved by utilizing advanced nanotechnologies, for example, nanostructured Li₃V₂(PO₄)₃ hybrid cathodes such as nanoparticles, nanowires, nanoplates, nanospheres, and others [5], and on the other hand, by introducing conductive carbon additives (carbon coated particles) [6,7,8,9] or some oxides such as SiO₂, ZrO₂, and others [3,10,11,12,13]. Another way to improve the charge transfer in the Li₃V₂(PO₄)₃ system is the substitution of both the transition element and its ligands [3,7,13,14].

Generally, it is known that the transition element in Li₃V₂(PO₄)₃ changes its valence state in the lithium intercalation/deintercalation process [15,16], indicating that the movement of lithium ions and the transfer of electrons can take place simultaneously. This mechanism can be realized in the case of the localized electron model. Moreover, the preference of the localized electrons model is proven by the fact that many cathode materials are poor conductors [17] and their conductivity is most likely caused by the excited carriers than the band conduction electrons [18]. In this model, the charge transfer is possible between transition ions of different valences directly or more often indirectly through the nearest ligand environment. The stable pathway of the electron transfer is realized due to the hybridization of d- and p-electronic orbitals of the transition elements and ligand surrounding them, respectively. At the same time, this hybridization can lead to the presence of the magnetic interaction between ions of transition elements and the appearance of different types of magnetic ordering. Any fluctuations, the crystal structure deformation due to the doping process, fluctuations of the chemical composition in composites and others, can lead to a change in the degree of orbital overlap and as a result, to simultaneously change the electron mobility and magnetic properties. Moreover, the fluctuation of the chemical composition can lead to enhanced ion mobility [19].

Therefore, the study of the local fluctuations in chemical composition and their effect on magnetic properties becomes important in connection with the experimentally observed relationships between electrical and magnetic properties. The investigation of the influence of the localized electron spin direction and the magnetic structure on the electron transfer between ions of different valences is of particular interest. In this regard, experiments on the study of the magnetic properties of cathode materials using highly sensitive local methods such as magnetometry and electron spin resonance are of great importance. Here, we present the magnetic properties study of the Li₃V₂(PO₄)₃/Li₃PO₄ system using the magnetometry and electron spin resonance methods. The investigated Li₃V₂(PO₄)₃/Li₃PO₄ composite system was obtained by the thermal hydrolysis method and, among others, this system is promising in terms of improving the electrochemical properties of cathode materials using Li₃PO₄ as an additional phase [20,21].

2. Experimental

2.1. Sample Synthesis and Characterization

In this paper, the investigated Li₃V₂(PO₄)₃/Li₃PO₄ (LVPO/LPO) solid solution was obtained by the thermal hydrolysis method with subsequent annealing in an Ar atmosphere. The synthesis was carried out according to the following scheme:

(i) Chemically pure lithium carbonate Li₂CO₃, phosphoric acid H₃PO₄, vanadium (IV) oxide V₂O₅, and oxalic acid C₂H₂O₄ were used in stoichiometric molar ratios as starting materials. An excess amount of lithium carbonate Li₂CO₃ and phosphoric acid H₃PO₄ was used to obtain the second phase (Li₃PO₄) in the amount of 7.5 mol. %. The reagents, in stoichiometric molar ratio, were mixed in a Teflon beaker for thermal hydrolysis using 20 mL of distilled water for the homogenization of the reaction mixture.

(ii) The annealing was carried out at 180 °C for 36 h. The cooling was carried out at room temperature without air access.

(iii) The resulting green-gray precursor was dried at 350 °C in an argon flow to remove moisture.

(iv) The dried precursor was calcined in two steps at 850 and 875 °C in the presence of carbon (5 mas. %) in an argon flow during 5 h. During the synthesis, the carbon reduced vanadium ions accompanied by the formation of gaseous carbon monoxide and carbon dioxide, which were removed from the reaction zone with an argon flow. Carbothermal reduction, as a rule, is realized at temperatures above 800 °С, while the temperatures of the phase formation of Li₃V₂(PO₄)₃, depending on the synthesis method, can vary over a wide range—from 700 to 1200 °С [22]. The final removal of carbon from the reaction mixture was confirmed by the complete dissolution of the synthesis product in nitric acid.

The composition of the obtained sample was controlled using a Shimadzu XRD-7000 S automatic diffractometer with of 0.03° step in the 10°–70° range with an exposure of 2 s at a point. The phase analysis of the reaction products was performed using the crystallographic database “Database of Powder Standard—PDF2” (ICDD, Newtown Square, PA, USA, Release 2009). X-ray pattern processing was performed according to the Rietveld method using the FULLPROF-2018 software. According to the x-ray diffraction data (Figure 1), the resulting product was a Li₃V₂(PO₄)₃ (92.5 mol. %)/Li₃PO₄ (7.5 mol. %) composite. The crystal structure of the Li₃V₂(PO₄)₃ phase belongs to the monoclinic space group P2₁/n (#14) with unit cell parameters a = 8.606(1) Å, b = 8.587(4) Å, c = 12.032(1) Å, α = γ = 90°, β = 90.554°(1), and V = 889.1(2) ų. The crystal structure of the Li₃PO₄ phase belongs to the orthorhombic space group Pnma (#62), a = 6.146 Å, b = 10.453 Å, c = 4.913 Å, α= β = γ 90°, and V = 315.64 ų. In comparison, the mesoporous sample of Li₃V₂(PO₄)₃/C (LVPO/C) was synthesized by the soft-template method [16,23], similarly to Na₃V₂(PO₄)₃ [24,25,26]. The unit cell parameters for the Li₃V₂(PO₄)₃/C material were a = 8.6095 Å, b = 8.6041 Å, c = 12.0560 Å, and β = 90.490°and its cell volume was estimated to be 893.044 ų (space group P2₁/n) [23]. The observed here difference between the crystal structure parameters for the two samples is quite common in the literature, so one can see that the lattice parameters can slightly differ depending on the synthesis process and monoclinic axis selection [27,28,29,30,31,32].

The morphology of Li₃V₂(PO₄)₃/Li₃PO₄ solid solution was investigated using scanning electron microscopy (SEM) via an EVO 50 XVP scanning electron microscope. SEM images of the Li₃V₂(PO₄)₃/Li₃PO₄ surface are shown in Figure 2. It was interesting to compare the results of the electron microscopy measurements of our samples and samples with carbon additives (LVPO/LPO (Figure 2) and LVPO/C composites (see Figure 2 in [23]), respectively). From the SEM images, one can see that composites had a granular structure with closed values of the average grain sizes.

2.2. Experimental Details

The magnetization was measured using a commercial PPMS-9 platform (Quantum Design, USA) in the temperature range from 5 to 305 K in field-cooled (FC) and zero field-cooled (ZFC) regimes. The magnetic hysteresis loops were measured in the magnetic field range of 1 T.

Electron spin resonance (ESR) spectrum of the Li₃V₂(PO₄)₃/Li₃PO₄ composite was measured on an ER 200 SRC (EMX/plus) spectrometer (Bruker, Germany) at the frequency of 9.4 GHz at room temperature using a double rectangular x-band resonator ER 4105DR. This equipment allows for the simultaneous detection of the electron spin resonance spectrum of the investigated sample and benchmark spectrum. The measured spectra were approximated using an open-source MATLAB toolbox for simulating and fitting a wide range of electron paramagnetic resonance spectra—EasySpin software package [33].

3. Results and Discussion

3.1. Magnetization

The magnetization of the Li₃V₂(PO₄)₃/Li₃PO₄ (LVPO/LPO) sample as a function of temperature (M-T curve) was measured in magnetic field H = 0.1 T in the FC regime (Figure 3). In all investigated temperature ranges, the magnetization of Li₃V₂(PO₄)₃/Li₃PO₄ decreased with an increase in temperature. At temperatures above T > 120 K, the inverse magnetic susceptibility is linear and can be well fitted by the Curie–Weiss law χ = C/(T-θ_CW), where C is the Curie constant and θ_CW is the Curie–Weiss temperature (inset in Figure 3). The high temperature approximation of the experimental data by the Curie–Weiss law gives the negative values of the Curie–Weiss temperature θ_CW = −68 К, which suggests the antiferromagnetic nature of the exchange interactions between spins in the investigated sample. Higher absolute value of the Curie–Weiss temperature in LVPO/LPO indicates stronger magnetic interactions in this sample compared to others. The known from the literature value of the Curie–Weiss temperature θ_CW = −37 К was obtained by Cahill et al. [34], proving the antiferromagnetic nature of exchange interactions in Li₃V₂(PO₄)₃ regardless of the synthesis method. The inset in Figure 3 shows inverse dependences of molar susceptibility, calculated taking into account the compositions of the composites (7.5 mol. % of Li₃PO₄ or 8% of carbon [23]). The approximation of inverse dependences of molar susceptibility gives the value of Curie constants С = 1.9 emu∙mol⁻¹Oe⁻¹ for LVPO/LPO and С = 0.84 emu∙mol⁻¹Oe⁻¹ for LVPO/C, respectively. The effective magnetic moment can be obtained from the Curie constants as:

$${\mu }_{eff}=\sqrt{3k_{B}C/N_A}$$

(1)

where k_B is the Boltzmann constant; C is the Curie constant; and N_A is the Avogadro constant. The obtained μ_eff was equal to 3.9 μ_B for LVPO/LPO and 2.59 μ_B for LVPO/C, respectively. Taking into account that the magnetic ion V³⁺ has a 3d² electronic configuration and ground state ³F with spin S = 1, let us estimate the effective magnetic moment μ_eff as:

$${\mu }_{theor}=g·\sqrt{Z·S·\left(S+1\right)}·{\mu }_B$$

(2)

where μ_B is the Bohr magneton; g is the Lande g-factor; Z is the number of magnetic ions in a unit cell; and S is the spin. Taking into account that g = 1.95 [16] for vanadium ions, one obtains the effective magnetic moment per mole of μ_theor = 3.9 μ_B. Theoretical and experimental values of the effective magnetic moment for the LVPO/LPO composite coincided very well, which confirms the valence state of the vanadium ions as V³⁺. The difference between the theoretical and experimental values of the effective magnetic moment for the LVPO/C composite is most likely associated with the higher carbon content in the sample compared to the data from [23]. In this case, the composite had a lower molar mass, which was used in calculating the data in the inset of Figure 3 and, therefore, the Curie constant and effective magnetic moment.

In addition to the above-mentioned experiments, the magnetization measurements in the ZFC-FC regimes in low magnetic field H = 5 mT (Figure 4) were performed to determine the temperature below which magnetic correlations become dominant over thermal fluctuations. The ZFC-FC splitting was observed in LVPO/LPO below T_split = 120 K, while in LVPO/C, this splitting was not observed. This splitting proves the above-mentioned suggestion of the presence of more significant short-range magnetic correlations in LVPO/LPO compared to LVPO/C. Isothermal magnetization measurements as a function of the external magnetic field below 120 K in LVPO/LPO (Figure 5) showed that in the investigated field range up to H = 1 T, the M-H curves were linear without any tendency to saturation. Most probably, the magnetic ordering was realized below T = 5 K.

Obviously, magnetic correlations in the LVPO/LPO sample can lead to the existence of ferromagnetically ordered regions. The appearance of these regions can be explained by the anti-site cation exchange, namely the occupation of transition metal sites by Li and vice versa, which is the typical point defects in crystal lattices and the topic of extraordinary research interest in solid state physics and chemistry. In the case of Li₃V₂(PO₄)₃, the occupation of the V site by the Li ion changes the charge allocation in the vicinity of this anti-site defect and freed electrons redistribute between neighboring ions of vanadium. De Gennes et al. [35] showed that in such a localized state, the small ferromagnetic cluster of the nearest neighboring vanadium ions can be formed due to the indirect exchange interaction between them, which was proven experimentally for the Li_xMn_1-xSe system [36,37]. The simultaneous existence of the paramagnetic phase with antiferromagnetic correlations and magnetically correlated regions, which form due to structural defects and the presence of the mixed-valence magnetic ions, was also previously observed in ytterbium manganites [38,39] and La_xSr_2-xFe_xTi_1-xO₄ compounds [40].

3.2. Electron Spin Resonance

As was above-mentioned, V³⁺ in an even number of electrons in the respective electronic shells and singlet ground-state levels may result so that no ESR is observable. At the same time, we were able to resolve two resonance signals in the ESR spectra of the Li₃V₂(PO₄)₃/Li₃PO₄ composite at room temperature (Figure 6). The first signal was observed in high magnetic fields. Approximation of this signal yielded the best fit of the experimental data for the powder spectrum, corresponding to the paramagnetic centers with effective spin S = 1/2 and anisotropic g-factor g_⊥ = 1.963, g_|| = 1.934 and the value of the linewidth ΔH = 1.64 mT. The observed anisotropic g-factor is the characteristic feature of most polyanionic cathode compounds where the transition element is localized in the tetragonally distorted octahedral crystal field. This was also the case for the Li₃V₂(PO₄)₃ [16]. Thus, the observed ESR spectra is most probably due to a small amount of V⁴⁺ ions (3d¹, S = 1/2). The second signal with g = 2 most likely represents a signal from dangling bonds or radicals and no particular interest. To estimate the number of V⁴⁺ ions, the integral intensity of the Li₃V₂(PO₄)₃/Li₃PO₄ spectrum was compared with the same parameters for the benchmark (inset in Figure 6). The integral intensities ratio of the sample (N_LVPO/LPO) and the reference (N₀) was equal to N_LVPO/LPO/N₀ = 4.81, which corresponds to the number of magnetic centers N_LVPO/LPO = 7.7 × 10¹⁷ (about 10 percent of total vanadium ions). The change in the valence state of vanadium ions and the presence of V⁴⁺ can be associated with insignificant lithium non-stoichiometry or anti-site defects (the occupation of V sites by Li and vice versa) in the investigated compound. The presence of anti-site defects can lead to the formation of ferromagnetically correlated clusters, which is confirmed by the above measurements of magnetization.

4. Conclusions

Here, we present the synthesis details and magnetic properties investigations of the Li₃V₂(PO₄)₃/Li₃PO₄ composite, which can be potentially used as a cathode material in a lithium-ion battery. The Li₃V₂(PO₄)₃/Li₃PO₄ composite was synthesized by the thermal hydrolysis method and contained 7.5 mol. % of Li₃PO₄ phase. It had a granular structure and consisted of nanoscale particles with a monolithic structure. Magnetization measurements of Li₃V₂(PO₄)₃/Li₃PO₄ indicate that above T > 120 K, the investigated sample was in the paramagnetic state and exhibited Curie–Weiss like behavior. The negative value of the Curie–Weiss temperature θ = −68 K suggests the presence of the antiferromagnetic interactions between the spins of vanadium ions. The experimentally obtained value of the effective magnetic moment was 3.9 μ_B for the Li₃V₂(PO₄)₃/Li₃PO₄ composite, which perfectly corresponded to the V³⁺ ions (3d², S = 1). The observed ZFC-FC splitting below T < 120 K suggests the presence of magnetically correlated regions (probably, ferromagnetically correlated clusters), in addition to the paramagnetic phase due to anti-site defects and the presence of V⁴⁺ ions. The existence of V⁴⁺ ions was directly confirmed by electron spin resonance measurements and the number of magnetic spins was estimated (<10%).