Enhancement of battery performance of LiMn₂O₄: correlations between electrochemical and magnetic properties

Abstract

We report the results of a systematic investigation of structural, electrical transport, magnetic, and electrochemical properties of LiB_xMn_2−xO₄ (where x = 0.0–1), synthesized via a one-step solid state reaction technique. We find that the parent compound (x = 0) has the non-stoichiometric Li_1.05Mn₂O₄ phase and the B ions successfully incorporate the structure for x ≤ 0.5. The resistivity anomaly of the parent compound below 100 K is attributed to the antiferromagnetic correlations. The linear part of the χ⁻¹–T data are used to calculate the effective magnetic moments and to determine the total number of Mn³⁺ and Mn⁴⁺ ions. Magnetization measurements show that Jahn–Teller active Mn³⁺ ions have high spin configuration. Average valance of Mn ions increases with increasing boron content. Electrochemical studies show improved capacity retention for x = 0.125 for 100 charge–discharge cycles. This improvement is attributed to subtle modifications in the structural and magnetic properties upon substitution.

---

Introduction

Li-based batteries have gained importance in the last two decades due to their high capacity and high energy density. Among the cathode materials, spinel LiMn₂O₄ have many advantages such as environmental safety, low cost, high stability and less toxicity.^1–3 However, spinel LiMn₂O₄ cathode suffers from a significant capacity loss after repeated cycling. The main reasons for the capacity loss are reported as Jahn–Teller distortion of [MnO₆] octahedron, dissolution of Mn ions in electrolyte, the structural degradation during lithiation and delithiation and high self-discharging during the storage.^4–6 Mainly two methods have been employed to overcome these problems. One of them is the surface coating to minimize the contact between the electrolytes and the active materials.^7,8 The second method is the modification of the intrinsic properties via partial substitution of Mn.^9–11 So far, many substitutions (S, F, Br, Si, B, Mg, Al, Bi, Cu, Ni, Ru, Ga, Ag, etc.^12–24) have been studied to increase the cycle life of LiMn₂O₄ cathode. The main idea behind this approach is to suppress the Jahn–Teller distortion via increasing the valance state of Mn ions and to stabilize the spinel structure during cycling.²⁵ In an octahedral environment, 3d electrons occupy five orbitals, e_g douplet, which lie along the x, y, z axis, and t_2g triplet which lie in between the axis, forming MnO₆ octahedron. The six ligands of the octahedron interact more with the e_g orbitals than with t_2g orbitals. Thus, the degeneracy of the five orbitals is lost and an energy difference between e_g and t_2g orbitals is formed. If the two e_g orbitals have different number of electrons, further degeneracy loss is observed causing geometric distortions also known as Jahn–Teller effect.²⁶

The crystal structure of LiMn₂O₄ have a spinel-type cubic structure with Fd₃m space group above a Verwey transition at ∼285 K.^27,28 Li ions are located in the 8a tetrahedral sites. The Mn ions, 50% Mn³⁺ and 50% Mn⁴⁺, are distributed randomly over the octahedral sites (16d) with an average valance state of 3.5.²⁹ The Mn⁴⁺ and Jahn–Teller active Mn³⁺ ions have t³_2ge⁰_g and t³_2ge¹_g configurations, respectively.^30,31 A single electron in e_g orbital of Mn³⁺ is expected to cause a strong Jahn–Teller distortion.³² Such distortion is preferred in order to reduce both the orbital degeneracy and overall electron–electron repulsion energy. The e_g electrons in LiMn₂O₄ can be localized causing a semiconducting type conductivity.^33–36 A structural transition from cubic to tetragonal with I4₁/amd space group is driven by a partial charge ordering between Mn³⁺ and Mn⁴⁺ among the distinct Mn sites. The transition is attributed to cooperative Jahn–Teller effect of Mn³⁺ ions.^37,38 On the other hand, LiMn₂O₄ has interesting magnetic properties related closely with the battery performance. It is a geometrically frustrated antiferromagnet, since the magnetic sublattice consists of a network of vertex sharing tetrahedra.

The electrical conduction mechanism of LiMn₂O₄ is another important phenomena in terms of the battery performance and understanding the nature of LiMn₂O₄.³⁹ There are two types of conductivity mechanism in LiMn₂O₄ as the ionic conductivity of Li ions and the non-adiabatic small polaron of e_g electrons of Mn³⁺ ions.⁴⁰ Iguchi et al. suggested that the ionic conductivity due to Li ions is not probable primary mechanism and the conductivity can be mainly due to nonadiabatic small polarons.⁴¹

Effects of substitution on cell performance of LiMn₂O₄ can be studied in many aspects such as the mass, size, valance state or magnetic properties of the substituted elements. There are many reasons to use boron as a substitution element. Despite the conflicting reports,⁴² substitution of Mn by elements heavier than the atomic weight of manganese is expected to decrease the specific capacity of doped LiMn₂O₄.⁴³ In addition, Mn³⁺ ions are Jahn–Teller active and it is essential to decrease the number of Mn³⁺ ions and hence to increase the structural stability. Thus, substitution Mn³⁺ ions with a non-magnetic element with 3+ valance state is expected to increase the structural stability by suppressing the Jahn–Teller distortion.⁴⁴ Substitution by ions with smaller ionic radius than the ionic radius of Mn is expected to decrease the cell parameters and make the structure more stable. However, there are many conflicting reports on this issue. For example substitution of Mn³⁺ (ionic radius: 0.58 Å, atomic mass: 54.93 amu) ions by 1% Y³⁺ (0.9 Å, 88.9 amu) shrinks the cell by ∼2%.⁴³ On the other hand, substitution by Al³⁺ (0.53 Å, 26.98 amu) slightly expands the cell.⁴¹ Non-magnetic B³⁺ with smaller ionic radius and smaller mass than Mn³⁺ is a good choice for substitution. Besides, there is limited number of studies regarding the effects of B substitution in LiMn₂O₄ on its electrochemical performance. These studies can be classified in two categories; Lithium Boron Oxide (LBO) coated on the LiMn₂O₄ and direct doping of B in Mn sites of LiMn₂O₄.^12–25,45 S. H. Choi et al. investigated the (LBO) glass modified LiMn₂O₄ cathode materials and they found that LBO coated LiMn₂O₄ has a better capacity retention (86% with 100 cycles) than that of the pure LiMn₂O₄ (65% with 100 cycles), however, coating LiMn₂O₄ with 5 wt% LBO decreases the initial capacity from 143 to 123 mA h g⁻¹.⁴⁶ H. W. Chan et al. also investigated the LBO coating effect on LiMn₂O₄ and found a capacity enhancement for the 0.4 wt% LBO coated LiMn₂O₄ cathode materials.¹⁶ H. Şahan et al. reported that LBO coated LiMn₂O₄ have better cyclability after 30 cycles than that of the uncoated LiMn₂O₄.⁴⁵ R. Thirunakaran et al. synthesized LiB_xMn_2−xO₄ (where x = 0–0.4) and reported improved electrochemical performance for x = 0.1 sample.⁴⁷ In another study, J. H. Lee et al. investigated the effects of different doping elements as Li, B, Al, Co and Ni and they found that LiB_0.05Mn_1.95O₄ has the worst battery performance compared to other compounds examined in Lee et al.'s study.⁴⁸

In this work, we examined the effects of B doping on the structural, magnetic and electrochemical properties of LiMn₂O₄. We found that the cell parameters and the cell volume increases with increasing B content. Electrochemical measurements show that 6.25% B substitution (x = 0.125 sample) improves the capacity retention significantly for 100 cycles. We revisited the magnetic and electrical transport properties of LiMn₂O₄ and studied the correlations between the electrochemical properties and structural, electrical, and magnetic properties of LiMn_2−xB_xO₄.

Experimental

Nominal compositions of LiB_xMn_2−xO₄ (where x = 0.0–1) were prepared by conventional solid state reaction technique. Appropriate amounts of reagent grade Li₂O, Mn₂O₃, and B₂O₃ were mixed in agate mortar for 1 h in glove box. We used 8% Li rich composition due to high temperature fast evaporation of Li as indicated in ref. 49 The mixture was heated in an α-Al₂O₃ crucible at 750 °C for 24 h under air and cooled down to room temperature.

Crystal structure investigation was carried out by X-ray diffraction (XRD). Automated Rigaku RadB Dmax X-ray diffractometer with Cu-k_α (λ_Cukα = 1.5405 Å) radiation was used for the XRD analysis with a scan rate of 2° min⁻¹.

Temperature dependences of resistivity (ρ–T) of the samples were measured by closed cycle Leybold LT-10 cryostat system combined SRS AC bridge system with 17.7 Hz frequency using four probe technique. The magnetic properties of the samples were measured simultaneously from 300 K to 2 K by Quantum Design-Physical Properties Measurement System (PPMS) with VSM attachment.

To fabricate the coin cell batteries, LiB_xMn_2−xO₄ active materials, acetylene black and polyvinylidene fluoride (80% : 10% : 10%) were mixed into slurries in N-methyl-2-pyrrolidone into an agate mortar. They were then cast on aluminum foils with a micrometer adjustable film applicator blade and then dried at 110 °C for 10 hours to obtain several electrode laminates. After these transactions, the laminates were put into the vacuum at ∼25 °C for 24 h. Discs (diameter 1/4 14 mm) of the laminates were punched and transferred to an argon filled glove box (MBraun UNILab Pro Modular Glove Box). A typical loading of the active material was approximately 5.0–6.0 mg. Afterwards, CR-2032 type coin-cells with pure Li metal as the counter electrode were assembled in the glove box with the LiPF₆ electrolyte. The cells were cycled by 8 channel battery analyzer system MTI-(BST8-MA) in the voltage window from 3.0 to 4.2 V at 1C current rate. Also the cyclic voltammograms (CV) of the cells were measured with MTI-BST8-STAT-EIS-LD Single Channel Potentiostat/Galvanostat workstation from 3.5 to 4.5 V at a scan rate of 0.2 mV s⁻¹. All measurements of the cells were conducted at room temperature.

Results and discussion

Structural properties

The XRD data (Fig. 1) show no impurity phases in the LiMn₂O₄ and the peak positions match well with the standard PDF card (35-0782) of LiMn₂O₄. The lattice parameter of LiMn₂O₄ is calculated as a = 8.2186 (89) Å by using Rietvelt-Refinement technique. M. Kopec et al. investigated the effects of Li concentration on the structural properties of LiMn₂O₄ and they found that Li content can change the lattice parameters significantly.⁵⁰ Our unsubstituted sample's lattice constants are closer to the lattice constants of Li_1.05Mn₂O₄,⁵⁰ considering the 8% extra Li addition during the synthesis. XRD pattern of the LiMn_2−xB_xO₄ x = 0.0–0.5 samples are almost identical with x = 0 sample without major impurity peaks indicating that boron is doped into the lattice of LiMn₂O₄ successfully. Impurity phase (Mn₃(BO₃)₂) is visible for x > 0.5 samples. The lattice parameters obtained via Rietveld refinement are listed in the Table 1.

Fig. 1 XRD patterns of LiMn_2−xB_xO₄ (from bottom to top, x = 0, 0.125, 0.25, 0.5, 0.75 and 1). Red (top) and blue (bottom) tick marks represent the theoretical peak positions of LiMn₂O₄ and Mn₃(BO₃)₂ phases, respectively. Left panel shows the slight change in peak position of (111) Bragg reflection upon substitution.

Table 1 Lattice parameters, unit cell volume, and Mn–Mn distances of LiMn_2−xB_xO₄ samples obtained from XRD data using Jade Rietveld refinement program. All samples have Cubic, Fd3̄m symmetry

x	a (Å) = b (Å) = c (Å)	RMn–Mn (Å)	V (Å^3)
0	8.21861 ± 0.00895	2.905 ± 0.005	555.1305
0.125	8.22520 ± 0.00265	2.907 ± 0.002	556.4669
0.25	8.22859 ± 0.00632	2.909 ± 0.004	557.1553
0.50	8.23268 ± 0.01243	2.910 ± 0.006	557.9865
0.75	8.24501 ± 0.00988	2.915 ± 0.005	560.4973
1.00	8.24838 ± 0.00454	2.916 ± 0.003	561.1849

---

There are controversial reports regarding the effects of B substitution on lattice volume of the LiMn₂O₄. A. Veluchamy et al.⁵¹ and S. H. Choi et al.⁴³ determined a decrease in lattice volume with B substitution.

On the other hand, J. H. Lee et al.,⁴⁸ R. Thirunakaran et al.⁴⁷ and Ebin et al.⁵³ reported that B substitution increases the cell parameters and expands the cell volume of LiMn₂O₄. Ebin et al. claim that some of the B³⁺ ions occupy the 8a tetrahedral sites instead of the 16d octahedral sites causing an expansion in the unit cell volume. In our data, the cell parameters increase with increasing B content, consistent with Ebin et al.'s results. In spite of this increase in the cell parameters, we observed that B substitution does not significantly affect the morphology of the samples (see ESI†).

Magnetic properties

LiMn₂O₄ has attracted attention from the scientists due to not only its battery performance but also its rich magnetic properties. Network of vertex sharing tetrahedra causes geometric frustration of magnetism in LiMn₂O₄. In these kind of materials, the magnetic ground state has a large number of degeneracy since the moments at the vertexes of a tetrahedron cannot have a definite spin structure. In the case of cubic spinels, however, magnetic exchange interactions between neighboring atoms have the ability to break the degeneracy of the ground state manifold. Depending on the balance between the exchange pathways, the system can have a unique magnetic ground state.²⁸ Previous literature holds conflicting reports about the magnetic properties of LiMn₂O₄ at low temperatures. Massarotti et al.⁵⁴ reported existence of long-range magnetic order below 65 K while Oohara et al.⁵⁵ reported only short-range correlations down to 8 K via neutron diffraction studies. In our LiMn₂O₄, although we did not observe any anomalies in our magnetic measurements showing antiferromagnetic correlations, we observed a resistivity anomaly below 100 K which is discussed in the following section.

The magnetic properties of LiMn₂O₄ are very sensitive to Li content in the structure. Excess Li ions replace with the Mn ions in the lattice and can cause changes in the electronic and magnetic properties.^50,52,56,57 Y. Shimakawa et al. reported that the cell parameters (from 8.24 Å to 8.21 Å for Li content of 1 to 1.1), Curie–Weiss constant (from 2.44 to 2.12 for per Mn) and the Mn³⁺/Mn⁴⁺ ratio (average Mn valance from +3.5 to 3.78) decrease with excess Li.^58,59 Neutron scattering measurements⁶⁰ show that excess Li suppresses the Jahn–Teller distortion down to 4 K increasing the stability of the cubic spinel structure. Boron substitution is another reason for decrease in Mn³⁺/Mn⁴⁺ ratio since the Mn³⁺ ions are substituted with B³⁺ ions. In this case, e_g electrons of the Mn³⁺ ions are the main reasons for the Jahn–Teller distortion. Any decrease in the number of Mn³⁺ ions is expected to suppress such structural instability.

The χ–T (magnetic susceptibility vs. temperature) curves of LiB_xMn_2−xO₄ for x = 0, 0.25, 0.5, 0.75 and 1 under 50 kOe (Fig. 2) show that the room temperature χ values decreases with increasing B content. Both excess Li and B substitution in our samples may suppress Jahn–Teller distortion and increase the structural stability with decreasing number of Mn³⁺ ions.^61,62

Fig. 2 χ–T data for LiB_xMn_2−xO₄ where x = 0, 0.25, 0.5, 0.75 and 1 under 50 kOe applied field.

On the other hand, below 50 K, magnetic susceptibility of x = 1 sample is higher than that of x = 0.75 sample. This is attributed to the increased amount of impurity phases as the boron content increases. XRD data in Fig. 1 shows that intensity of the Mn₂(BO₃)₂ impurity for x = 1 sample is higher than that of x = 0.75 sample. Thus, increase in the magnetic susceptibility of x = 1 sample below 50 K can be related with Mn₂(BO₃)₂ impurity phase.

In order to quantitatively understand the change of the susceptibility due to B substitution in LiMn₂O₄, one can concentrate on the Curie–Weiss form of the susceptibility as;

where χ₀ is the temperature independent susceptibility term, C is the Curie–Weiss constant and θ is the Weiss temperature. The χ₀ is a very small value and can be disregarded as mentioned in ref. 63.

Therefore, only the linear part of the χ⁻¹–T curves are taken into account for Curie–Weiss fittings as (Fig. 3);

Fig. 3 χ⁻¹–T data for LiB_xMn_2−xO₄ where x = 0, 0.25, 0.5, 0.75 and 1. The solid blue lines indicate the Curie–Weiss fitting.

The fitting parameters are listed in Table 2. The C and θ values for LiMn₂O₄ are determined as 4.518 emu K mol⁻¹ and −253 K, respectively, in good agreement with the literature.^64–66 The negative θ value shows that the antiferromagnetic interactions are dominant in LiMn₂O₄.

Table 2 The fitting parameters obtained from eqn (2), (4) and χ–T curve

x	C	χ	μeff from χ–T	μeff from eqn (4)^a	z in eqn (4)	Mean valence of Mn ions	Chemical formula of compounds extracted from eqn (4)
a Using equal number Mn^3+ and Mn^4+ for HS of Mn^3+ and HS of Mn^4+.
0	4.518	−253	6.03665	6.16541	0.41081	3.589	Li1.05Mn^3+0.8Mn^4+1.15O4
0.25	4.424	−277	5.97358	5.75664	0.66572	3.334	Li1.05Mn^3+1.13Mn^4+0.57B^3+0.25O4
0.5	3.698	−247	5.46173	5.31653	0.61948	3.379	Li1.05Mn^3+0.9Mn^4+0.55 B^3+0.5O4
0.75	3.182	−273	5.06631	4.83654	0.70991	3.292	Li1.05Mn^3+0.85Mn^4+0.35B^3+0.75O4
1.00	2.486	−251	4.64451	4.30334	0.85574	3.147	Li1.05Mn^3+0.81Mn^4+0.14 B^3+1.0O4

---

Our Curie Weiss parameters for LiMn₂O₄ are consistent with the Curie Weiss parameters of Li_1.05Mn_1.95O₄ reported in the literature.^61–63 In the paramagnetic region of χ–T data, the Curie constant can be written as;

where N_A is the number of ions, μ_eff is the effective magnetic moment, μ_B is the Bohr magneton and k_B is the Boltzmann constant. The Mn³⁺ ions can have either HS (high spin) state with μ_eff = 4.9 μ_B or LS (low spin) state with μ_eff = 2.83 μ_B depending on the strength of the crystal field splitting of d-orbitals. Mn⁴⁺ ions have three unpaired electrons in an octahedral environment with μ_eff = 3.87 μ_B. The μ_eff values obtained from Curie–Weiss fitting decrease with increasing boron content as shown in Fig. 4. LiMn₂O₄ with excess Li can be written as Li_1+yMn_2−yO₄ as explained in detail in ref. 67. Effective magnetic moment of Li_1+yMn_2−yO₄ can be expressed as:

μ_eff² = (2 − y)[zμ_eff²(Mn³⁺) + (1 − z)μ_eff²(Mn⁴⁺)] (4)

where y is the amount of excess Li which is taken as 0.05.

Fig. 4 Boron content dependence of μ_eff. The green and black lines indicate the theoretical μ_eff values determined by eqn (4) for Mn³⁺(HS), Mn⁴⁺(HS) and Mn³⁺(LS), Mn⁴⁺(HS), respectively. The solid circles represent the experimental μ_eff values extracted from χ⁻¹–T data using Curie–Weiss law.

Using eqn (4), we theoretically calculated the μ_eff value depending on the number of Mn³⁺ and Mn⁴⁺ ions with different spin states, shown in Fig. 4. The theoretical values for HS state of Mn³⁺ ions match well with the experimental values indicating that the HS state is more probable than the LS state for Mn³⁺ ions.

Schimel et al. reported via inelastic neutron scattering measurements that Mn³⁺ ions are in low spin state in Li [Mn_1.94Li_0.04]O_3.94.⁶⁸ However, they also admitted that small compositional differences may affect the spin state of Mn and occurrence of Jahn–Teller distortions due to their dynamic nature.⁶⁸

We calculated the theoretical effective magnetic moment of the pure LiMn₂O₄ as μ_eff = 6.16 μ_B for the whole sample assuming that the number of Mn³⁺ and Mn⁴⁺ are equal. We also extracted the experimental effective magnetic moment from the χ–T curve as = 6.04 μ_B for the whole sample which is slightly lower than the theoretical one. Experimental μ_eff values decrease with increasing boron content as seen in Fig. 4 and Table 2. Average valence state and the mole fractions of Mn ions are also listed in Table 2. Average valance state of Mn ions decrease with increasing boron content, consistent with the results obtained by J. H. Lee et al.⁴⁸ We have determined the Mn³⁺/Mn⁴⁺ ratio in Li_1.05Mn_1.95O₄ as 0.76 (=∼0.8/∼1.15). From the Curie–Weiss fitting of susceptibility, μ_eff is found as 3.015 μ_B/Mn ion which is a slightly smaller value than the one obtained from eqn (4).

All these small differences between the theoretical and experimental values can be due to the replacement of the Mn³⁺ ions with excess Li and B and/or the spin–orbit coupling. It is well known that Mn³⁺ for a d⁴ configuration of transition metals, the high spin state is f³_2ge¹_g in an octahedral environment. In the absence of the octahedral field, the configuration of the four d- electrons of the Mn³⁺ ion in the incomplete 3d shell is ⁵D ground term. A weak crystal field causes this term to split into two states that ⁵D term splits into the orbital doublet ⁵E_g as the ground sub-term and excited orbital triplet ⁵T_2g.⁶⁹ The Jahn–Teller distortion of Mn³⁺ in a weak field has ⁵E_g, the ground state which has an imbalance in the filling electrons in orbitals. As a result, a significant distortion from the octahedral symmetry takes place.

Normally, the ⁵E_g ground state is paramagnetic from four unpaired electrons with μ_spin-only = 4.9 μ_B. However, the spin-only magnetic moment is not always observed. Although the ground state has no orbital contribution, the excited electronic states with same spin multiplicity can be mixed with the ground state. This may be the case for high spin state of Mn³⁺ ions and the effective magnetic moment will deviate from the spin-only formula value causing small discrepancies between theoretical and experimental results.⁷⁰

Electrical transport properties

LiMn₂O₄ has almost equal number of Mn³⁺ and Mn⁴⁺ ions and they order below 285 K with a Verwey transition.²⁸ The crystal field splits the d-electrons of Mn ions as t_2g and e_g configurations. The e_g levels of Mn³⁺ and Mn⁴⁺ in the lattice have one and zero electrons, respectively. The e_g electrons in Mn³⁺ does not go to the neighboring Mn⁴⁺ sites causing insulating type conductivity.⁷¹ The resistivity value of our LiMn₂O₄ at room temperature is determined as 5 kohm cm which is consistent with the above explanation. Room temperature resistivity values of LiMn_2−xB_xO₄ increases with increasing B content as shown in Fig. 5. Here, B³⁺ ions replace the Mn³⁺ ions and atomic energy levels of B and Mn are significantly different from each other causing such increase in room temperature resistivity values. The conduction mechanism of LiMn₂O₄ depends on the hoping electrons between Mn ions with different valance states.^72,73 The overlapping of the wave functions due to the Mn–Mn distances can affect the conduction mechanism of the samples as explained in ref. 74. The distance of Mn–Mn atoms is given⁷⁴ as where a is the lattice constant of LiMn₂O₄. While the distance is lower than 2.743 Å, the sample shows metallic conductivity, otherwise it exhibits non-conducting behavior.⁷⁵ The calculated values of Mn–Mn distances, listed in Table 3, increase with increasing B content which is consistent with the resistivity data. Yuan et al. reported a separation of ZFC (zero field cooling) and FC (field cooling) curves at 65 K in the magnetization data since spin–spin correlations are effective above 65 K and the structural transition was observed at around 100 K.³³ M. Kopec et al. reported the Neel temperature of LiMn₂O₄ as 65 K which is related to the onset of a completed charge ordering.⁵⁰ Neutron diffraction studies on LiMn₂O₄ show short range correlations above 100 K.²⁸ All these antiferromagnetic correlations are expected to appear as a resistivity anomaly below 100 K.⁵⁰ Fig. 6 shows the temperature dependent resistivity data for pure LiMn₂O₄. The sample has semiconducting type conductivity down to 97.8 K and then the resistivity gives a small deep at 66.5 K before starting to increase with decreasing temperature.

Fig. 5 Room temperature resistivity values of LiMn_2−xB_xO₄.

Table 3 The data obtained from the voltammogram of the samples. The differences between the anodic and cathodic peaks are ΔE_i = ith anodic peak − ith cathodic peak (P_i)

x (content)	Anodic peaks		Cathodic peaks		ΔE1 (V)	ΔE2 (V)
	P1 (V)	P2 (V)	P1 (V)	P2 (V)
0	4.11	4.24	3.88	4.00	0.23	0.24
0.125	4.06	4.19	3.90	4.07	0.16	0.12
0.25	4.12	4.28	3.84	4.01	0.28	0.27
0.375	4.06	4.19	3.90	4.05	0.16	0.14
0.5	4.06	4.22	3.91	4.05	0.15	0.17

---

Fig. 6 ρ–T and dM/dT curves of pure LiMn₂O₄.

The differential change of magnetization with temperature has two critical points as shown in the Fig. 6. The first deviation point from linearity in the dM/dT is related to start of the antiferromagnetic correlations and the second point is 65.2 K which corresponds to the resistivity minimum on ρ–T curve. It is possible that this resistivity anomaly region corresponds to the transition from short-range correlations to long range antiferromagnetic ordering.^76,77 These results will be confirmed by ongoing neutron scattering studies.

Electrochemical properties

The electrochemical properties of the B doped LiMn₂O₄ are investigated systematically. The cycling voltammograms, in the potential region of 3.5–4.5 V at the sweep rate of 0.2 mV s⁻¹, of the electrodes fabricated by different boron content are given in Fig. 7. There are two redox current peaks at 4.11/3.88 and 4.24/4.01 V for LiMn₂O₄ electrode. The two peak structure in voltammogram is attributed to the intercalation reaction of lithium ions occupying 8a sites in the LiMn₂O₄.^36,78–80 The first peak in the voltammogram corresponds to the removal of Li-ions from the tetrahedral sites where Li–Li interactions are dominant and the second peak corresponds to the removal of lithium ions again from the tetrahedral sites but where Li–Li interactions are insignificant.⁴

Fig. 7 Cyclic voltammograms of (a) x = 0, (b) x = 0.125, (c) x = 0.250, (d) x = 0.375 and (e) x = 0.5 samples for first 3 cycles.

The reactions during the charge and discharge of electrodes can be expressed as LiMn₂O₄ ↔ Li_0.5Mn₂O₄ ↔ λ-MnO₂.⁸¹ It is expected that the potential differences between anodic and cathodic current should be as low as possible due to the reversible process of electrodes.

The obtained differences in the voltammogram are within these limits. The peak positions of the LiMn_2−xB_xO₄ cathodes for various B contents (Table 3) are almost identical. As mentioned above, substitution of Jahn–Teller active Mn³⁺ ions with B³⁺ ions suppresses the Jahn–Teller distortion and makes the structure more stable during charge–discharge cycles. Almost identical peak positions of cathodes with various B contents (Fig. 7) confirm this structural stability upon B substitution.

Fig. 8 shows the charge–discharge curves of the batteries with LiMn_2−xB_xO₄ cathodes for various B content up to 100 cycles. The battery cells are activated until 4.2 V and then they are cycled between 3.0 and 4.2 V. There are two plateaus during the charge (at 4 V and 4.14 V) and discharge (at 4.1 V and 3.8 V) processes. The first plateau is due to the extraction of Li ions from half of tetrahedral site with Li–Li interaction and the second is due to the extraction of the other part of Li ions in the charging curve. The discharge process is the inverse of the charging process as explained in ref. 82. The numbers of Mn⁴⁺ and Mn³⁺ ions are important parameters for high cell capacity. If the number of Mn⁴⁺ ions is more than that of the Mn³⁺, lattice parameters may become smaller due to ionic radii of Mn ions. As a result of smaller lattice parameters Mn–O bonds get stronger while Li–O interactions weaken.⁸³ Thus, the number of Li ions contributing the charge and discharge process increase.^49,82 On the other hand, Mn³⁺ ions are Jahn–Teller active and they can cause the lattice parameters to expand and the capacity to fade. Yamada et al. reported that the Jahn–Teller instability due to Mn³⁺ ions can be reduced by 2–3% substitution of Mn ions by other elements.⁸⁰ In this study, the highest number of Mn³⁺ ions are obtained in x = 0.25 sample (Table 2) which gives the lowest initial cell capacity.

Fig. 8 Charge–discharge profiles of (a) x = 0, (b) x = 0.125, (c) x = 0.250, (d) x = 0.375 and (e) x = 0.5 samples.

The cycling performance of the cells fabricated with LiMn_2−xB_xO₄ cathodes up to 100 cycles are shown in Fig. 9. Although B substitution decreases the maximal capacity of the cells with increasing boron content, LiMn_1.875B_0.125O₄ (6.25% substitution) shows higher cycling performance after 60 cycles than that of the pure LiMn₂O₄ (Table 4).

Fig. 9 Capacity retention of LiMn_2−xB_xO₄.

Table 4 LiMn_(2−x)B_(x)O₄ discharge capacity values for various cycles

x (content)	1 cycle (mA h g^−1)	25 cycle (mA h g^−1)	50 cycle (mA h g^−1)	75 cycle (mA h g^−1)	100 cycle (mA h g^−1)
0	122.23558	103.96484	96.33511	82.30601	77.36307
0.125	122.22481	103.4814	95.48102	91.42203	87.36305
0.25	94.56857	85.57071	79.99714	79.96071	78.32143
0.375	105.4476	73.7613	75.7044	66.7233	68.3451
0.5	98.4555	40.88696	31.92052	30.84455	30.48589

---

Electrochemical performance of LiMn₂O₄ spinel is highly affected by Jahn–Teller distortion.⁷⁹ Although Mn³⁺ ions are in HS configuration which induces this distortion, replacing these Jahn–Teller active ions with non-magnetic B³⁺ ions helps suppressing the Jahn–Teller effect and enhances the electrochemical performance. Especially x = 0.125 sample has a higher capacity retention (71% for 100 cycle) than the x = 0 sample (63% for 100 cycle) showing that 6.25% B substitution makes LiMn₂O₄ structurally more stable. We attribute this result to a fine balance of the interatomic distances and the Mn³⁺/Mn⁴⁺ ratio upon substitution. As explained in the above sections B³⁺ ions with small ionic radius (27 pm) replace with the Mn³⁺ ions and some of them may go into the 8a tetrahedral Li sites. Average valance of Mn ions change with increasing B content.

Thus, B³⁺ substitution affects the magnetic and structural properties and especially 6.25% substitution is a promising cathode for Li-ion batteries. Initial capacities of x = 0.15, 0.375 and 0.5 samples are significantly lower than that of the x = 0 and 0.125 samples. Moreover, there are no detectable impurity phases in the XRD data (Fig. 1) of these samples. This result supports that some of the B³⁺ ions are placed in the 8a Li sites and decrease the number of electrochemically active Li⁺ ions and the initial capacities. Further investigations are ongoing to determine the structural, magnetic and electrochemical properties of x < 0.125 samples with finer intervals of x.

Conclusions

The structural, magnetic and electrical properties of LiMn_2−xB_xO₄, 0 ≤ x ≤ 1 are studied in detail. The structural studies show that there are no impurity phases in x ≥ 0.5 samples. Unit cell parameters increase with increasing boron content, indicating that some of the boron ions are located in 8a tetrahedral sites instead of 16d octahedral sites. The observed anomaly in the temperature dependent resistivity data is attributed to the transition from short range correlations to long range antiferromagnetic ordering. The μ_eff values of the samples are determined from the linear part of the χ⁻¹–T curves. The amount of Mn³⁺ ions, which are in the HS state and Jahn–Teller active mode, decreases with increasing B content. Such decrease is expected to suppresses the Jahn–Teller distortion and make the structure more stable during electrochemical cycling. Electrochemical measurements give promising results for x = 0.125 sample in terms of the capacity retention which is important for the improvement of the Li-ion battery technology. Significantly low initial cell capacities of the samples with high B content (x = 0.25, 0.375 and 0.5) confirms that some of the B³⁺ ions are placed in the Li sites.

Acknowledgements

This study was supported by TUBITAK (The Scientific and Technical Research Council of Turkey) under grant no. TUBITAK 112M487 and IUBAP (Inonu University Scientific Research council)-2014/02.

References

1. C. Wei, J. Shen, J. Zhang, H. Zhang and C. Zhu, RSC Adv., 2014, 4, 44525.
2. B.-S. Liu, Z.-B. Wang, Y. Zhang, F.-D. Yu, Y. Xue, K. Ke and F.-F. Li, J. Alloys Compd., 2015, 622, 902.
3. D.-L. Fang, J.-C. Li, X. Liu, P.-F. Huang, T.-R. Xu, M.-C. Qian and C.-H. Zheng, J. Alloys Compd., 2015, 640, 82.
4. G. Xu, Z. Liu, C. Zhang, G. Cui and L. Chen, J. Mater. Chem. A, 2015, 3, 4092.
5. H. Huang, C. A. Vincent and P. G. Bruce, J. Electrochem. Soc., 1999, 146, 481.
6. M. M. Thackeray, Y. Shao-Horn, A. J. Kahaian, K. D. Kepler, E. Skinner, J. T. Vaughey and S. A. Hackney, Electrochem. Solid-State Lett., 1998, 1, 7.
7. J. M. Zheng, J. Li, Z. R. Zhang, X. J. Guo and Y. Yang, Solid State Ionics, 2008, 179, 794.
8. X. Liu, J. Wang, J. Zhang and S. Yang, Mater. Sci. Eng., A, 2006, 430, 248.
9. J. Cabana, L. Monconduit, D. Larcher and M. R. Palacín, Adv. Mater., 2010, 22, E170.
10. J.-M. Tarascon and M. Armand, Nature, 2001, 414, 359.
11. M. G. Kim and J. Cho, Adv. Funct. Mater., 2009, 19, 1497.
12. Q. Liu, S. Wang, H. Tan, Z. Yang and J. Zeng, Energies, 2013, 6, 1718.
13. D. Zhan, Y. Liang, P. Cui and Z. Xiao, RSC Adv., 2015, 5, 6372.
14. C.-G. Han, C. Zhu, G. Saito and T. Akiyama, RSC Adv., 2015, 5, 73315.
15. A. Iturrondobeitia, A. Goñi, L. Lezama, C. Kim, M. Doeff, J. Cabana and T. Rojo, J. Mater. Chem. A, 2013, 1, 10857–10862.
16. H. W. Chan, J. G. Duh and S. R. Sheen, Surf. Coat. Technol., 2004, 188, 116.
17. P. Singh, A. Sil, M. Nath and S. Ray, Ceram.-Silik., 2010, 54, 38.
18. F. A. Amaral, N. Bocchi, R. F. Brocenschi, S. R. Biaggio and R. C. Rocha-Filho, J. Power Sources, 2010, 195, 3293.
19. D. Q. Liu, X. Q. Liu and Z. Z. He, Mater. Chem. Phys., 2007, 105, 362.
20. Z. D. Peng, Q. L. Jiang, K. Du, W. G. Wang, G. R. Hu and Y. X. Liu, J. Alloys Compd., 2010, 493, 640.
21. Y. Ein Eli, R. C. Urian, W. Wen and S. Mukerjee, Electrochim. Acta, 2005, 50, 1931.
22. Y. Xu, G. Chen, E. Fu, M. Zhou, M. Dunwell, L. Fei, S. Deng, P. Andersen, Y. Wang, Q. Jiad and H. Luo, RSC Adv., 2013, 3, 18441.
23. H. L. Wang, T. A. Tan, P. Yang, M. O. Lai and L. Lui, J. Phys. Chem., 2011, 115, 6102.
24. D. Q. Liu, Z. Z. He and X. Q. Liu, J. Alloys Compd., 2007, 440, 69.
25. H. Zhang, Y. Xu and D. Liu, RSC Adv., 2015, 5, 11091.
26. H. A. Jahn and E. Teller, Proc. R. Soc. London, Ser. A, 1937, 161, 220.
27. D. G. Fauteux, A. Massucco, J. Shi and C. Lampe-Önnerud, J. Appl. Electrochem., 1997, 27, 543.
28. A. S. Wills, N. P. Raju and J. E. Greedan, Chem. Mater., 1999, 11, 1510.
29. M. Bianchini, F. Fauth, E. Suard, J.-B. Leriche, C. Masquelier and L. Croguennec, Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2015, 71, 688.
30. P. Umek, A. Gloter, M. Pregelj, R. Dominko, M. Jagodič and Z. Jagličić, J. Phys. Chem. C, 2009, 113, 14798.
31. S. K. Ranjan and M. S. Sundar, Solid State Sci., 2003, 5, 549.
32. J. A. Alonso, M. J. Martinez-Lope and M. T. Casais, Inorg. Chem., 2000, 39, 917.
33. Y. F. Yuan, H. M. Wu, S. Y. Guo, J. B. Wu, J. L. Yang, X. L. Wang and J. P. Tu, Appl. Surf. Sci., 2008, 255, 2225.
34. D. H. Jang, Y. J. Shin and S. M. Oh, J. Electrochem. Soc., 1996, 143, 2204.
35. Y. Xia, N. Kumada and M. Yoshio, J. Power Sources, 2000, 90, 135.
36. L. Guohua, H. Ikuta, T. Uchida and M. Wakihara, J. Electrochem. Soc., 1996, 143, 178.
37. C. Y. Ouyang, S. Q. Shi and M. S. Lei, J. Alloys Compd., 2009, 474, 370.
38. D. Capsoni, M. Bini, G. Chiodelli, V. Massarotti, C. B. Azzoni, M. C. Mozzati and A. Comin, Phys. Chem. Chem. Phys., 2001, 3, 2162.
39. J. Marzec, K. Świerczek, J. Przewoźnik, J. Molenda, D. R. Simon, E. M. Kelder and J. Schoonman, Solid State Ionics, 2002, 146, 225.
40. K. Hoang, J. Mater. Chem. A, 2014, 2, 18271.
41. E. Iguchi, Y. Tokuda, H. Nakatsugawa and F. Munakata, J. Appl. Phys., 2002, 91, 2149.
42. M. V. Reddy, S. S. Manoharan, J. John, B. Singh, G. V. Subba Rao and B. V. R. Chowdari, J. Electrochem. Soc., 2009, 156, A652.
43. D. K. Lee, S. C. Han, D. Ahn, S. P. Singh, K.-S. Sohn and M. Pyo, ACS Appl. Mater. Interfaces, 2012, 4, 6842.
44. D. Capsoni, M. Bini, G. Chiodelli, P. Mustarelli, V. Massarotti, C. B. Azzoni, M. C. Mozzati and L. Linati, J. Phys. Chem. B, 2002, 106, 7432.
45. H. Şahan, H. Göktepe, Ş. Patat and A. Ülgen, Solid State Ionics, 2008, 178, 1837.
46. S. H. Choi, J. H. Kim, Y. N. Ko and Y. C. Kang, Int. J. Electrochem. Sci., 2013, 8, 1146.
47. R. Thirunakaran, K.-T. Kim, Y.-M. Kang and J.-Y. Lee, Ionics, 2004, 10, 188.
48. J. H. Lee, J. K. Hong, D. H. Jang, Y.-K. Sun and S. M. Oh, J. Power Sources, 2000, 89, 7.
49. Z. Chang, Z. Chen, F. Wu, H. Tang, X. Z. Yuan and H. Wang, Electrochem. Solid-State Lett., 2008, 11, A229.
50. M. Kopec, J. R. Dygas, F. Krok, A. Mauger, F. Gendron and C. M. Julien, J. Phys. Chem. Solids, 2008, 69, 955.
51. A. Veluchamy, H. Ikuta and M. Wakihara, Solid State Ionics, 2001, 143, 161.
52. C. Masquelier, M. Tabuchi, K. Ado, R. Kanno, Y. Kobayashi, Y. Maki, O. Nakamura and J. B. Goodenough, J. Solid State Chem., 1996, 123, 255.
53. B. Ebin, G. Lindbergh and S. Gürmen, J. Alloys Compd., 2015, 620, 399.
54. V. Massarotti, D. Capsoni, M. Bini, P. Scardi, M. Leoni, V. Baron and H. Berg, J. Appl. Crystallogr., 1999, 32, 1186.
55. Y. Oohara, J. Sugiyama and M. Kontani, J. Phys. Soc. Jpn., 1999, 68, 242.
56. K. Yamaura, Q. Huang, L. Zhang, K. Takada, Y. Baba, T. Nagai, Y. Matsui, K. Kosuda and E. Takayama-Muromachi, J. Am. Chem. Soc., 2006, 128, 9448.
57. C. M. Julien, A. Ait-Salah, A. Mauger and F. Gendron, Ionics, 2006, 12, 21.
58. Y. Shimakawa, T. Numata and J. Tabuchi, J. Solid State Chem., 1997, 131, 138.
59. J. M. Tarascon, E. Wang, F. K. Shokoohi, W. R. McKinnon and S. Colson, J. Electrochem. Soc., 1991, 138, 2859.
60. H. G. Schimmel, W. Montfrooij, G. J. Kearley, V. W. J. Verhoeven and I. M. de Schepper, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 63, 214409.
61. J. Sugiyama, T. Atsumi, A. Koiwai, T. Sasaki, T. Hioki, S. Noda and N. Kamegashira, J. Phys.: Condens. Matter, 1999, 9, 1729.
62. R. Englman and B. Halperin, Phys. Rev. B: Solid State, 1970, 2, 75.
63. G. A. Bain and J. F. Berry, J. Chem. Educ., 2008, 85, 532.
64. Y.-I. Jang, F. C. Chou, B. Huang, D. R. Sadoway and Y.-M. Chiang, J. Phys. Chem. Solids, 2003, 64, 2525.
65. M. Tabuchi, C. Masquelier, H. Kobayashi, R. Kanno, Y. Kobayashi, T. Akai, Y. Maki, H. Kageyama and O. Nakamura, J. Power Sources, 1997, 68, 623.
66. K. Mukai and J. Sugiyama, Chem. Lett., 2009, 38, 944.
67. T. Eriksson, A. M. Andersson, A. G. Bishop, C. Gejke, T. Gustafsson and J. O. Thomas, J. Electrochem. Soc., 2002, 149, A69.
68. D. Wu, G. L. Liu, C. Jing, Y. Z. Wu, D. Loison, G. S. Dong, X. F. Jin and D.-S. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 63, 214403.
69. R. J. Radwanski and Z. Ropka, J. Magn. Magn. Mater., 2004, 272, e259.
70. D. Gatteschi, R. Sessoli and J. Villain, Molecular Magnets, Oxford University press, Newyork, 2006.
71. M. Park, X. Zhang, M. Chung, G. B. Less and A. M. Sastry, J. Power Sources, 2010, 195, 7904.
72. J. M. Amarilla, K. Petrov, F. Pico, G. Avdeev, J. M. Rojo and R. M. Rojas, J. Power Sources, 2009, 191, 591.
73. K. Tateishi, D. du Boulay and N. T. Ishizawa, Appl. Phys. Lett., 2004, 84, 529.
74. J. Guan and M. Liu, Solid State Ionics, 1998, 110, 21.
75. J. Blasco, J. García, J. M. de Teresa, M. R. Ibarra, P. A. Algarabel and C. Marquina, J. Phys.: Condens. Matter, 1996, 8, 7427.
76. J. Cho, J. Mater. Chem., 2008, 18, 2257.
77. R. Basu and R. Seshadri, J. Mater. Chem., 2000, 10, 507.
78. S. Chitra, P. Kalyani, T. Mohan, R. Gangadharan, B. Yebka, S. Castro-Garcia, M. Massot, C. Julien and M. Eddrief, J. Electroceram., 1999, 3, 433.
79. Y. Xia, Y. Zhou and M. Yoshio, J. Electrochem. Soc., 1997, 144, 2593.
80. J. Yamada, Solid State Chem., 1996, 122, 160.
81. B. Ammundsen, G. R. Burns, M. S. Islam, H. Kanoh and J. Rozière, J. Phys. Chem. B, 1999, 103, 5175.
82. S. Chen, C. Mi, L. Su, B. Gao, Q. Fu, X. Zhang, M. Zhao, L. Yao and T.-Y. Zhang, J. Appl. Electrochem., 2009, 39, 1943.
83. B. Ying, W. Chuan, W. Feng and W. Guo-qing, Nonferrous Met., 2006, 16, 402.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra05032h

---