Blending Lithium Nickel Manganese Cobalt Oxide with Lithium Iron Manganese Phosphate as Cathode Materials for Lithium-ion Batteries with Enhanced Electrochemical Performance

Abstract

The effects of LiMn_0.7Fe_0.3PO₄/C (LMFP) species on the electrochemical performance of the blended cathodes of LiNi_0.5Mn_0.3Co_0.2O₂ (NMC) and LMFP are examined. Two types of LMFPs are synthesized by the hydrothermal method (LMFP-1) and the solid-state reaction (LMFP-2) leading to different physical characteristics and uniformity of primary particles. The blended cathodes of NMC and LMFP-1 show higher discharge capacity, gravimetric energy density, and rate capability than those of NMC and LMFP-2 for the same blending ratio (10, 20, 30, 40, and 50 wt% LMFP to NMC) because of the differences in the electronic conductivity, specific surface area, mean particle size, and uniformity of primary particles between LMFP-1 and LMFP-2. The discharge capacity and gravimetric energy density of NMC : LMFP-1 = 9 : 1 and 8 : 2 at 0.2 C-rate are comparable to those of NMC. Further, the rate capability is the highest at the blending ratio of NMC : LMFP-1 = 7 : 3. An optimal range for the blending ratio of LMFP to NMC is revealed based on the discharge capacity, energy density, and rate capability of the blended cathode. Moreover, LMFP has a considerable impact on the electrochemical characteristics of the blended cathodes of NMC and LMFP.

1. Introduction

Lithium-ion batteries are used in portable electronic devices, electric vehicles, and power storage facilities for solar and wind power generation, and they are expected to be deployed in a variety of applications in the future. To address the increasing demand for lithium-ion batteries with enhanced energy density, cathode active materials have been extensively investigated.^1,2 One of the promising candidates for cathode materials is high-nickel lithium nickel manganese cobalt oxides (LiNi_xMn_yCo_zO₂, NMCs).^3,4 However, increasing the Ni content of NMCs enhances the energy density while decreasing the cycling performance and thermal stability.⁵ In addition, the low thermal stability of NMCs increases the risk of thermal runaway of Li-ion batteries, which have been circumvented by studying thermal management, battery management, electrolyte additives, separator design, and cathode materials.^6,7 In cathode materials, the blends of NMC and lithium iron manganese phosphate (LiMn_xFe_yPO₄, LMFP) have attracted attention as they provide high energy density and thermal stability^8,9 because LMFP can compensate for the low thermal stability of NMC. Zhang et al. reported that the thermal stability of LiNi_0.6Mn_0.2Co_0.2O₂ was improved by the blended cathode of LiNi_0.6Mn_0.2Co_0.2O₂ with 30 wt% LiMn_0.8Fe_0.2PO₄/C.⁸ This blended cathode also achieved good cycling performance. 10 Ah pouch cells of the blended cathode of LiNi_0.6Mn_0.2Co_0.2O₂ with 5 wt% LiMn_0.7Fe_0.3PO₄/C passed nail penetration and overcharge tests and exhibited the same energy density for LiNi_0.6Mn_0.2Co_0.2O₂.⁹ Furthermore, LMFP is known to possess different physical characteristics and electrochemical performances based on the synthesis method used.¹⁰ LMFP has been synthesized by solid-state,¹¹ co-precipitation,¹² sol-gel,¹³ hydrothermal,¹⁴ and solvothermal methods.¹⁵ These methods can offer specific advantages; for example, uniform primary nanoparticles can be synthesized by the hydrothermal method. However, the impacts of the different physical and electrochemical characteristics of LMFPs on the blended cathodes of NMC and LMFP have not been adequately studied. Moreover, the effect of the blending ratio of NMC and LMFP on the battery performance has not been reported in detail. Herein, we study the electrochemical performances of the blended cathodes of LiNi_0.5Mn_0.3Co_0.2O₂ with LiMn_0.7Fe_0.3PO₄/C synthesized by solid-state and hydrothermal methods. The electrochemical performances of the blended cathodes are investigated in a wide range of the blending ratio, from 10 wt% to 50 wt% LiMn_0.7Fe_0.3PO₄/C.

2. Experimental

2.1 Cathode preparation

LiNi_0.5Mn_0.3Co_0.2O₂ (NMC) was purchased from M K Impex Corp. Two types of LiMn_0.7Fe_0.3PO₄/C were synthesized by the hydrothermal method (LMFP-1) and employing the solid-state reaction (LMFP-2).^11,16 Composite cathode slurries using a mixture of NMC and LMFP were fabricated by mixing 90 wt% active material, 5 wt% acetylene black (Denka Co., Ltd., Denka Black) as the conducting agent, and 5 wt% polyvinylidene difluoride (Kureha Corp., KF#9305) as the binder in N-methyl-2-pyrrolidone (Wako Pure Chemical Ind. Ltd., special grade). For LMFP-1 and LMFP-2 cathodes, composite cathode slurries were fabricated by mixing 70 wt% active material, 20 wt% acetylene black, and 10 wt% polyvinylidene difluoride. The prepared slurry was coated on an Al foil current collector (Hohsen Corp., thickness: 20 µm) using a doctor blade and dried at 80 °C for 2 h in a vacuum oven. The loading weight of the cathode material on Al foil was ∼1.2 mg cm⁻² for LMFP-1 and LMFP-2 and 9–10 mg cm⁻² for NMC and the blended cathode of NMC and LMFP. The disk-shaped composite cathodes (14 mm in diameter and 30 µm in thickness) were fabricated by roll-press.

2.2 Characterization

The crystalline phases of the samples were identified using powder X-ray diffraction (XRD) with a Bruker D8 Advance diffractometer with CuKα (λ = 0.15406 nm) radiation operating at 35 kV and 350 mA. The XRD patterns were collected in the 2θ range from 15° to 65° using a scanning step length of 0.0234° and scanning speed of 0.13 s/step. The lattice parameters were calculated using the Rietveld method and TOPAS3.0 software (Bruker). The particle morphology of the samples and cross section of the blended cathodes of NMC and LMFP were observed using a field emission scanning electron microscope (FE-SEM, JSM-7001F) employing a secondary electron image detector and accelerating voltage of 15 kV. The average particle sizes of the samples were measured using the laser diffraction/scattering particle size distribution method with an MT3300EXII (MicrotracBEL Corp.). The specific surface area of the samples was measured using the Brunauer–Emmett–Teller (BET) method with a FlowSorb III 2305 (Shimadzu Corp.). The electronic conductivity of the samples was measured using the powder resistivity measurement system MCP-PD51 (Nittoseiko Analytech Co., Ltd.). The carbon content of LMFP-1 and LMFP-2 was measured using the carbon/sulfur analyzer EMIA220V2 (HORIBA, Ltd.).

2.3 Electrochemical measurements

The electrochemical properties of the fabricated cathodes were examined using CR2032 coin-type cells. Lithium foil (Honjo Metal Corp.) was used as the counter electrode. The electrolyte solution comprised 1 mol dm⁻³ LiPF₆ in a 3 : 7 mixture of ethylene carbonate and ethyl methyl carbonate, respectively. The separators comprised porous polypropylene (Cellgard 2400). The CR2032 coin-type cells were constructed in an argon-filled glove box (dew point < −90 °C, O₂ level < 0.2 ppm).

The galvanostatic charge–discharge profiles and discharge rate capabilities were measured for the CR2032 coin-type cells at 30 °C using a battery testing system (Hokuto Denko Co., model: HJ1001SM8A). The cells fabricated using a blended cathode of NMC and LMFP were tested in a voltage range of 3.0–4.25 V using a constant current charge at a current rate of 34 mA g⁻¹ (0.2 C). The discharge behavior was tested at 3.0 V using a constant current protocol at current rates of 34, 170, 510, 850, and 1700 mA g⁻¹ (0.2, 1, 3, 5, and 10 C). For the cells comprising LMFP-1 or LMFP-2 cathodes, the electrochemical tests were performed in a voltage range of 2.0–4.5 V using a constant current and voltage charge (charging continued at 4.5 V until the current reached a value corresponding to 1/10th of the value of the charging current rates for a constant current protocol) at a current rate of 34 mA g⁻¹ (0.2 C). The discharge behavior was tested at 2.0 V using a constant current protocol at current rates of 34, 170, 510, 850, and 1700 mA g⁻¹ (0.2, 1, 3, 5, and 10 C).

3. Results and Discussion

3.1 Physical properties of NMC, LMFP-1, and LMFP-2

The XRD patterns of LMFP-1 and LMFP-2 are shown in Fig. 1. The crystalline phase of both samples is identified to be the LiMn_0.7Fe_0.3PO₄ phase, and no impurities are detected from the XRD patterns. This crystal structure is assigned to the orthorhombic space group Pnma (No. 62).¹⁷ The intensities of the LMFP-2 diffraction peaks are smaller than those of LMFP-1, indicating that LMFP-1 has higher crystallinity than LMFP-2.

Figure 1.

XRD patterns of (a) LMFP-1 synthesized using the hydrothermal method and (b) LMFP-2 synthesized using the solid-state reaction.

The particle size and morphology of NMC, LMFP-1, and LMFP-2 are observed using FE-SEM. The secondary particles of NMC, LMFP-1, and LMFP-2 comprising the nano-sized primary particles can be observed in Fig. 2. The secondary particle size of NMC is larger than that of LMFP-1 and smaller than that of LMFP-2 (Figs. 2a, 2b, and 2d). The secondary particles of NMC and LMFP-1 are spherical. On the other hand, the secondary particles of LMFP-2 have an irregular shape. This morphological difference between LMFP-1 and LMFP-2 is most likely caused by the difference in the synthesis methods of LMFPs. As shown in Figs. 2c and 2e, the primary particles of LMFP-1 are more uniform than those of LMFP-2. The primary particles of LMFP-2 are generally smaller than those of LMFP-1, although some primary LMFP-2 particles are larger than those of LMFP-1. The differences in the primary and secondary particle morphologies between LMFP-1 and LMFP-2 result from the different synthesis methods used.

Figure 2.

SEM images of (a) NMC, (b, c) LMFP-1, and (d, e) LMFP-2.

The physical properties of LMFP-1, LMFP-2, and NMC are shown in Table 1. The electronic conductivity of LMFP-1 is more than 10 times as high as that of LMFP-2, and the electronic conductivity of NMC is much higher than that of LMFP-1. This is owing to the difference in the synthesis method between LMFP-1 and LMFP-2, which may cause the differences in the morphologies of carbon and the primary and secondary particles. The carbon content of LMFP-1 and LMFP-2 is 1.3 wt% and 4.5 wt%, respectively. To obtain good electrochemical performances, LMFP-2 requires a larger amount of carbon than LMFP-1. This might be caused by the differences in the crystallinity or particle size between LMFP-1 and LMFP-2. On this point, more detailed analysis is needed in the future. The difference of the carbon content between LMFP-1 and LMFP-2 can lead to make the difference in the positional relationship of carbon and LMFP particles between LMFP-1 and LMFP-2, and this may cause the difference of the electronic conductivity. The specific surface area of LMFP-1 is much smaller than that of LMFP-2. This result indicates that the average primary particle size of LMFP-1 is larger than that of LMFP-2 as shown in Fig. 2. Moreover, the specific surface areas of LMFP-1 and LMFP-2 may be affected not only by the primary particle size but also carbon contents and their morphology.¹⁸ The specific surface area of NMC is much smaller than that of both LMFP-1 and LMFP-2. The mean particle size diameter (D₅₀) of NMC, LMFP-1, and LMFP-2 is 12.34, 10.61, and 23.92 µm, respectively. The D₅₀ value of NMC is larger than that of LMFP-1 and smaller than that of LMFP-2. This corresponds to the secondary particle size shown in Figs. 2a, 2b, and 2d. The true densities of LMFP-1, LMFP-2, and NMC are 3.4, 3.2, and 4.6 g cm⁻³, respectively. The true density of NMC substantially differs from those of LMFPs, while LMFP-1 and LMFP-2 show similar true densities.

Table 1. Physical properties of LMFP-1, LMFP-2 and NMC.

Sample	Electronic conductivity (S cm−1)	Specific surface area (m2 g−1)	D50 (µm)	True density (g cm−3)
LMFP-1	4.44E-06	20	10.61	3.4
LMFP-2	1.13E-07	39	23.92	3.2
NMC	1.82E-03	0.3	12.34	4.6

3.2 Electrochemical performance of NMC, LMFP-1, and LMFP-2

The galvanostatic charge–discharge profiles and the corresponding dQ/dV curves of NMC, LMFP-1, and LMFP-2 are shown in Fig. 3. As shown in Fig. 3a, the voltage of NMC changes continuously between 4.25 V and 3.00 V with charge or discharge. The 0.2 C discharge capacity of NMC is 154.3 mAh g⁻¹, and the initial coulombic efficiency is 85.6 %. On the other hand, the charge curve of LMFP-1 shows two distinct plateaus at approximately 3.5 and 4.1 V at a current rate of 0.2 C. These plateaus correspond to the extraction of lithium ions from LMFP at these voltages for Fe³⁺/Fe²⁺ and Mn³⁺/Mn²⁺ redox couples, respectively.¹⁹ Similarly, the discharge curve of LMFP-1 shows the corresponding two plateaus at 3.5 and 4.0 V at a current rate of 0.2 C. The charge and discharge curves of LMFP-2 show two distinct plateaus same as those of LMFP-1, but these plateaus are shorter than those of LMFP-1. The 0.2 C-rate discharge capacities of LMFP-1 and LMFP-2 are 149.5 and 129.6 mAh g⁻¹, respectively. The 0.2 C-rate discharge capacity of LMFP-1 is 1.15 times higher than that of LMFP-2.

Figure 3.

Charge–discharge curves and corresponding dQ/dV curves of (a, b) NMC, (c, d) LMFP-1, and (e, f) LMFP-2.

Furthermore, the differential capacities (dQ/dV) are calculated to determine the redox behaviors. As the discharge rate increases, the peaks of the dQ/dV curves of NMC, LMFP-1, and LMFP-2 shift to lower voltages during discharge. This is due to resistive and diffusion polarization effects. The dQ/dV curves of NMC during charge and discharge have one peak at 3.81 V and 3.72 V at a current rate of 0.2 C, respectively. These peaks are assigned to the oxidation and reduction of the nickel ions of NMC.^20,21 The dQ/dV curves of LMFP-1 have two distinct peaks at around 3.54 V and 4.13 V during charge and around 3.47 V and 4.00 V during discharge at a current rate of 0.2 C, which correspond to Fe³⁺/Fe²⁺ and Mn³⁺/Mn²⁺ redox couples, respectively. The dQ/dV curves of LMFP-1 show two peaks at all current rates. As the current rate increases from 0.2 C-rate to 5 C-rate, these two peaks of discharge dQ/dV curves shift to lower voltages. The dQ/dV curves of LMFP-2 also have two distinct peaks during charge at around 3.54 V and 4.12 V and during discharge at around 3.47 V and 3.94 V at a current rate of 0.2 C. As the discharge rate increases, these peaks of the dQ/dV curves of LMFP-2 during discharge shift to lower voltages, and the intensities of the peaks diminish. The peaks are barely observed during discharge at current rates of 3 C and 5 C. For LMFP-1, the voltage difference between the oxidation and reduction peaks at a current rate of 0.2 C is 70 mV for Fe³⁺/Fe²⁺ and 130 mV for Mn³⁺/Mn²⁺, respectively. For LMFP-2, this difference is 70 mV for Fe³⁺/Fe²⁺ and 180 mV for Mn³⁺/Mn²⁺, respectively. These results indicate that the voltage difference between the oxidation and reduction peaks for Fe³⁺/Fe²⁺ are almost the same between LMFP-1 and LMFP-2, but for Mn³⁺/Mn²⁺, the difference is smaller in the case of LMFP-1 than that of LMFP-2. In addition, the intensity of the peaks of the dQ/dV curves of LMFP-2 during discharge is much smaller than those of LMFP-1 at all discharge rates. The smaller voltage difference between charge and discharge and the higher peak intensities of the dQ/dV curves indicate faster kinetics for electrodes during lithiation and delithiation, resulting from the combined effects of coated carbon morphology and the size, uniformity, and crystallinity of primary LMFP particles.^22–27

Table 2 shows the discharge capacities of LMFP-1 and LMFP-2 at current rates of 0.2, 1, 3, and 5 C. LMFP-1 delivers higher discharge capacities than LMFP-2 at all current rates. The rate performances (5 C-rate/0.2 C-rate discharge capacity ratio) of LMFP-1 and LMFP-2 are 0.93 and 0.62, respectively. Regarding LMFP-2, the polarization of the discharge curves increases with the discharge current rate (Figs. 3c and 3e). On the other hand, the polarization of the discharge curves of LMFP-1 increases smaller than that of LMFP-2 (Figs. 3c and 3e). These results indicate that LMFP-1 shows rate capability superior to that of LMFP-2, likely due to the differences in electronic conductivity and lithium-ion conductivity between LMFP-1 and LMFP-2 derived from the size and uniformity of primary particles, coated carbon morphology, and crystallinity on the surface of LMFP primary particles.

Table 2. Discharge capacities of LMFP-1 and LMFP-2 at current rates of 0.2, 1, 3, and 5 C.

Discharge rate	LMFP-1	LMFP-2
0.2 C	149.5	129.6
1 C	145.6	116.2
3 C	141.5	95.2
5 C	139.3	80.9

3.3 Characterization of the blended cathodes of NMC and LMFP

Before electrochemical performance is examined, we analyze the cross section of the cathodes using FE-SEM to confirm that the active materials are sufficiently dispersed in the cathode active material layer. The cross-sectional SEM images of the blended cathodes of NMC and LMFP are shown in Fig. 4. From the SEM images, the white spheres and the gray regions represent NMC particles and LMFP particles, respectively. Secondary LMFP-1 particles are spherical when synthesized, but they are deformed by pressing during cathode fabrication. Acetylene black and polyvinylidene difluoride are present in the black area in the gaps between NMC and LMFP particles. NMC and LMFP particles are well dispersed in the cathode active material layer. The thickness of the cathode active material layer is almost the same between the blended cathodes of NMC and LMFP-1 and that of NMC and LMFP-2 at the same blending ratio of LMFP to NMC.

Figure 4.

Cross-sectional SEM images of blended cathodes at (a) NMC : LMFP-1 = 9 : 1, (b) NMC : LMFP-2 = 9 : 1, (c) NMC : LMFP-1 = 8 : 2, (d) NMC : LMFP-2 = 8 : 2, (e) NMC : LMFP-1 = 7 : 3, (f) NMC : LMFP-2 = 7 : 3, (g) NMC : LMFP-1 = 6 : 4, (h) NMC : LMFP-2 = 6 : 4, and (i) NMC : LMFP-1 = 5 : 5, (j) NMC : LMFP-2 = 5 : 5.

In addition to cross-sectional SEM images, the densities of the cathode active material layer are calculated by loading weight and thickness. The densities of the cathode active material layer of the blended cathodes of NMC and LMFP are shown in Table 3. As shown in Table 3, there are no differences in loading weight, thickness, and density between the blended cathodes of NMC and LMFP-1 and that of NMC and LMFP-2 at the same blending ratio of LMFP to NMC. The thickness of the blended cathodes is around 40–50 µm, and this value corresponds to the thickness shown in Fig. 4. As the weight ratio of LMFP increases in blended cathodes, the densities slightly decrease in the blended cathodes of NMC with LMFP-1 and LMFP-2. This is due to the larger true density of NMC particles than those of LMFP particles, as shown in Table 1.

Table 3. Loading weights, thicknesses, densities of the blended cathodes of NMC and LMFP.

LMFP ratio (wt%)	LMFP-1			LMFP-2
	Loading weight (g cm−2)	Thickness (µm)	Density (g cm−3)	Loading weight (g cm−2)	Thickness (µm)	Density (g cm−3)
10	0.93	42	2.2	0.81	41	2.0
20	1.05	48	2.2	1.01	47	2.1
30	1.04	47	2.2	0.95	45	2.1
40	1.00	50	2.0	0.92	47	2.0
50	0.90	47	1.9	0.90	47	1.9

3.4 Electrochemical performance of blended cathode of NMC and LMFP

The blended cathodes of NMC and LMFP are examined for the electrochemical properties. The galvanostatic charge–discharge curves of blended cathodes of NMC and LMFP-1 and that of NMC and LMFP-2 are shown in Fig. 5. As the blending ratio of LMFP to NMC increases, two distinct plateaus of charge curves and discharge curves at around 4.0 V and 3.5 V become longer in the blended cathodes of NMC and LMFP-1. The plateau at 4.0 V corresponds to the Mn³⁺/Mn²⁺ redox couple of LMFP, and the plateau at 3.5 V corresponds to the Fe³⁺/Fe²⁺ redox couple of LMFP. For the blended cathodes of NMC and LMFP-2, the plateaus are not as clear as those in the blended cathodes of NMC and LMFP-1, and the voltage changes rapidly along with charge or discharge at all blending ratios of NMC to LMFP-2. As the discharge rate becomes higher, the voltage drops rapidly along with discharge.

Figure 5.

Charge–discharge curves of blended cathodes at (a) NMC : LMFP-1 = 9 : 1, (b) NMC : LMFP-2 = 9 : 1, (c) NMC : LMFP-1 = 8 : 2, (d) NMC : LMFP-2 = 8 : 2, (e) NMC : LMFP-1 = 7 : 3, (f) NMC : LMFP-2 = 7 : 3, (g) NMC : LMFP-1 = 6 : 4, (h) NMC : LMFP-2 = 6 : 4, and (i) NMC : LMFP-1 = 5 : 5, (j) NMC : LMFP-2 = 5 : 5.

The blended cathodes of NMC and LMFP-1 achieve a higher discharge capacity than the blended cathodes of NMC and LMFP-2 at all discharge rates and blending ratios. The 0.2 C-rate discharge capacity of the blended cathodes of NMC and LMFP-1 is 154, 154, 148, 143, and 141 mAh g⁻¹ at the blending ratio (NMC : LMFP-1) of 9 : 1, 8 : 2, 7 : 3, 6 : 4, and 5 : 5, respectively. On the other hand, the 0.2 C-rate discharge capacity of the blended cathodes of NMC and LMFP-2 is 139, 136, 126, 104, and 92 mAh g⁻¹ at the blending ratio (NMC : LMFP-2) of 9 : 1, 8 : 2, 7 : 3, 6 : 4, and 5 : 5, respectively. The decrease in discharge capacity with the increase in the blending ratio of LMFP to NMC is because the discharge capacity of LMFP is smaller than that of NMC. In addition, as the blending ratio of LMFP increases, the differences in the 0.2 C-rate discharge capacity between the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 increase. Considering these results, the discharge capacities of the blended cathodes of NMC and LMFP are greatly affected by the electrochemical properties of LMFP, as shown in Figs. 3c and 3e.

The differential capacities (dQ/dV) are calculated to observe the redox behaviors in the blended cathodes of NMC and LMFP. The dQ/dV curves of the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 are shown in Fig. 6. As the discharge rate becomes higher, the dQ/dV curves of the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 during discharge have their reduction peaks at lower voltages for all blending ratios. For dQ/dV curves during charge, the voltages do not change as greatly as those during discharge. This is due to the constant charge current rate of 0.2 C. For the blended cathodes of NMC and LMFP-1, the two distinct reduction peaks appear in NMC : LMFP-1 = 9 : 1 during discharge at 0.2 C current rates, and only one peak appears at 1, 3, and 5 C discharge rates. At a current rate of 0.2 C, two peaks exist at 3.98 V and 3.69 V during discharge, which are attributed to the Mn³⁺/Mn²⁺ redox reaction in LMFP and the redox reaction of the nickel ion in NMC, respectively. At 1, 3, and 5 C discharge rates, the peak corresponds to the redox reaction of the nickel ion in NMC. For NMC : LMFP-1 = 8 : 2, 7 : 3, 6 : 4, and 5 : 5, three distinct peaks are observed during discharge at a current rate of 0.2 C. At NMC : LMFP-1 = 6 : 4, there are three peaks at around 3.48 V, 3.71 V, and 4.03 V during discharge, which are attributed to the Fe³⁺/Fe²⁺ redox reaction in LMFP, the redox reaction of the nickel ion of NMC, and the Mn³⁺/Mn²⁺ redox reaction in LMFP, respectively. At current rates of 1, 3, and 5 C, the peak corresponding to the Fe³⁺/Fe²⁺ redox reaction at around 3.3–3.4 V is barely observed. As the blending ratio of LMFP-1 increases, the intensity of the peak representing the redox reaction of the nickel ion in NMC at around 3.7 V decreases, and those corresponding to Fe³⁺/Fe²⁺ and Mn³⁺/Mn²⁺ redox couples at around 3.5 V and 4.0 V increase. In NMC : LMFP-2 = 9 : 1 and 8 : 2, only one peak appears during discharge at current rates of 1, 3, and 5 C. This peak corresponds to the redox reaction of the nickel ion in NMC. At a current rate of 0.2 C, one additional peak is observed at 4.16 V during discharge corresponding to the Mn³⁺/Mn²⁺ redox reaction in LMFP at NMC : LMFP-2 = 9 : 1 and 8 : 2, but this peak is very small. In NMC : LMFP-2 = 7 : 3, 6 : 4, and 5 : 5, this peak at around 4.16 V is observed more clearly during discharge at all current rates, but the peak is still much smaller than that of the blended cathodes of NMC and LMFP-1. As for the blended cathodes of NMC and LMFP-2, the peak corresponding to the redox reaction of the nickel ion in NMC becomes small as the blending ratio of LMFP-2 to NMC increases. The peak corresponding to the Fe³⁺/Fe²⁺ redox couple in LMFP barely appears at all blending ratios. There is no significant voltage difference between the oxidation and reduction peaks at around 3.7 V corresponding to the redox reaction of the nickel ion in NMC in both the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 at all blending ratios. However, the voltage differences between the oxidation and reduction peaks of the Mn³⁺/Mn²⁺ redox couple in the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 at a current rate of 0.2 C are around 110 mV and 170 mV, respectively, at the blending ratio of NMC : LMFP = 8 : 2, 7 : 3, 6 : 4, and 5 : 5. These results indicate that the decrease in the discharge capacity of the blended cathode of NMC and LMFP-2 along with the increase in the blending ratio of LMFP-2 to NMC is ascribed to not only the decrease in the blending ratio of NMC but also the lower electrochemical performance of LMFP-2 than that of LMFP-1. The analysis of the dQ/dV curves suggests that the electrochemical performance of LMFP largely influences the electrochemical performance of the blended cathodes of NMC and LMFP.

Figure 6.

dQ/dV curves of (a) NMC : LMFP-1 = 9 : 1, (b) NMC : LMFP-2 = 9 : 1, (c) NMC : LMFP-1 = 8 : 2, (d) NMC : LMFP-2 = 8 : 2, (e) NMC : LMFP-1 = 7 : 3, (f) NMC : LMFP-2 = 7 : 3, (g) NMC : LMFP-1 = 6 : 4, (h) NMC : LMFP-2 = 6 : 4, and (i) NMC : LMFP-1 = 5 : 5, (j) NMC : LMFP-2 = 5 : 5 blended cathodes.

The 0.2 C-rate discharge capacities, the gravimetric energy densities, and capacity retention (5 C-rate/0.2 C-rate) as a function of the LMFP ratio in the blended cathodes of NMC and LMFP are shown in Fig. 7. As for the blended cathodes of NMC and LMFP-1, NMC : LMFP-1 = 9 : 1 and NMC : LMFP-1 = 8 : 2 achieve a 0.2 C-rate discharge capacity of 154 mAh g⁻¹, respectively (Fig. 7a). These are equal to the 0.2 C-rate discharge capacity of NMC, as shown in Fig. 3a. The gravimetric energy densities based on cathode materials of NMC : LMFP-1 = 9 : 1 and NMC : LMFP-1 = 8 : 2 are 583 Wh kg⁻¹ and 589 Wh kg⁻¹, respectively (Fig. 7b). These are also comparable to that of NMC (587 Wh kg⁻¹). However, the discharge capacities and the gravimetric energy densities decrease by increasing the blending ratio of LMFP-1 to NMC in NMC : LMFP-1 = 7 : 3, NMC : LMFP-1 = 6 : 4, and NMC : LMFP-1 = 5 : 5. This tendency that the effect of blending ratio on discharge capacity changes beyond a certain blending ratio has been reported in previous studies.^28,29 This tendency relies on the optimization of percolation networks or the connectivity of active materials, which lowers the electrical resistance of the blended cathode of LiCoO₂ and Li₂RuO₃ reported by Stux et al.²⁸ In other words, this tendency is assumed to be caused by the difference in the electron pathways of NMC and LMFP-1 particles in the cathodes at each blending ratio of NMC and LMFP-1. When the blending ratio of LMFP-1 to NMC is 10 wt% and 20 wt%, LMFP-1 particles are assumed to exist in the gaps formed by NMC particles in the cathode, and this may lead to maintaining electron pathways through the NMC particles. When the blending ratio of LMFP is higher than 30 wt%, the volumetric ratio of LMFP to NMC exceeds 50 vol% calculated by each true density shown in Table 1, and NMC particles are assumed to exist in the gaps formed by LMFP particles in the cathode. This may lead to maintaining electron pathways through the LMFP particles. Due to the lower electronic conductivity of LMFP compared with NMC (Table 1), the resistance of blended cathodes of NMC and LMFP increases with increasing LMFP content, decreasing the discharge capacity and energy density at 30 wt% or higher blending ratio of LMFP.

Figure 7.

(a) 0.2 C-rate discharge capacities, (b) 0.2 C-rate discharge energy densities, and (c) rate capability of blended cathodes of NMC and LMFP as a function of LMFP/(NMC + LMFP) weight ratio in the cathode.

The discharge capacities and energy densities of the blended cathodes of NMC and LMFP-2 are lower than those of NMC and LMFP-1 at the same blending ratios of LMFP to NMC (Figs. 7a and 7b). This is caused by the difference in the discharge capacities and energy densities between LMFP-1 and LMFP-2. This is supported by the results that the difference in the discharge capacities and the energy densities between the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 increase as the blending ratio of LMFP to NMC increases. The 0.2 C-rate discharge capacities and the gravimetric energy densities of the blended cathodes of NMC and LMFP-2 decrease with the increased blending ratio of LMFP-2 to NMC. Especially, when the blending ratio of LMFP-2 is higher than 30 wt%, 0.2 C-rate discharge capacities and the gravimetric energy densities decrease drastically. This tendency would be caused by the same mechanism in the case of the blended cathodes of NMC and LMFP-1.

Based on the above results, there is a clear difference in the tendency of the discharge capacity and gravimetric energy density as a function of the blending ratio of NMC and LMFP between NMC/LMFP-1 and NMC/LMFP-2. As for the blended cathodes of NMC and LMFP-1, the 0.2 C discharge capacities and gravimetric energy densities of them decrease approximately linearly with increase of the ratio of LMFP-1. The discharge capacity and gravimetric energy density of the blended cathode of NMC and LMFP-2 decreases linearly at 30 wt% or lower blending ratio of LMFP-2. However, at 40 wt% or higher blending ratio of LMFP-2, they decrease more rapidly. These results indicate that the physical and electrochemical characteristics of LMFP may be important to bring out the electrochemical performances of the blended cathode of NMC and LMFP. Due to some factors, it can be difficult for LMFP-2 to maintain a good electron pathway in the blended cathode. Identifying these factors is important for improving the electrochemical performance of the blended cathode, so more detailed research in this point is needed in the future.

As in the case of the discharge capacity and energy density, the difference of the trend of the rate capability can be seen between the blended cathode of NMC/LMFP-1 and NMC/LMFP-2. As shown in Fig. 7c, the rate capability is almost the same between the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 at NMC : LMFP = 9 : 1, but the blended cathodes of NMC and LMFP-1 achieve higher rate capability than those of NMC and LMFP-2 at NMC : LMFP = 8 : 2, 7 : 3, 6 : 4, and 5 : 5. As the blending ratio of LMFP increases, the difference in the rate performance between the blended cathodes of NMC/LMFP-1 and NMC/LMFP-2 becomes larger, indicating that the rate capability of the blended cathode is greatly affected by the physical or electrochemical characteristics of LMFP, such as the particle morphology and the electronic conductivity.

For the blended cathodes of NMC and LMFP-1, the rate capability is maintained at the same value of 0.80 at the blending ratio of 30 wt% LMFP-1 or lower. This value is almost equivalent to the rate capability of NMC (0.83) and LMFP-1 (0.83), and this result is consistent with our expectations. However, at NMC : LMFP-1 = 6 : 4 and 5 : 5, the rate capability gradually decreases with the increased blending ratio of LMFP-1 to NMC. The cause of this result is not clear at present, but it is assumed to be caused by the difference in the pathways of the electrons of NMC and LMFP-1 particles in the cathodes as a function of a blending ratios of NMC and LMFP-1, as in the case of the discharge capacity and energy density of the blended cathodes. Thus, in order to maintain the good rate capability, it is presumed to be important to ensure the good electron pathways in the cathodes by optimizing the particle size distribution and particle morphology including carbon coating the cathode materials as in the case of discharge capacity and energy density.

For the blended cathodes of NMC and LMFP-2, the rate capability decreases along with the increase in the blending ratio of LMFP-2 to NMC from 10 wt% to 50 wt%. Especially, when the blending ratio of LMFP-2 to NMC is over 30 wt%, the rate capability significantly decreases. This trend is somewhat similar to that of the blended cathode of NMC and LMFP-1, and it is also probably due to the difference in the electron pathways of NMC and LMFP particles in the cathodes as a function of a blending ratio of NMC and LMFP, as in the case of the blended cathodes of NMC and LMFP-1. However, at 30 wt% or higher blending ratio of LMFP, the rate capability of the blended cathode of NMC/LMFP-2 decreases more rapidly than that of the blended cathode of NMC/LMFP-1. From this result, it can be assumed that the physical or electrochemical characteristics of LMFP have a significant effect on the rate capability of the blended cathode of NMC and LMFP as well as in the case of the discharge capacity and gravimetric energy density.

4. Conclusions

We investigated the effects of the physical properties of LMFP on the electrochemical characteristics of the blended cathodes of NMC and LMFP for lithium-ion batteries. For LMFP, LMFP-1 synthesized by the hydrothermal method and LMFP-2 synthesized by the solid-state reaction were used. NMC : LMFP-1 = 9 : 1, 8 : 2, 7 : 3, 6 : 4, and 5 : 5 showed higher discharge capacities and gravimetric energy densities compared with the same blending ratios of LMFP-2 to NMC at current rates of 0.2, 1, 3, and 5 C. The discharge capacities and gravimetric energy densities of NMC : LMFP-1 = 9 : 1 and 8 : 2 at 0.2 C-rate were comparable to those of NMC. The discharge capacities and gravimetric energy densities of the blended cathodes of NMC and LMFP-2 decreased significantly as the blending ratio of LMFP-2 to NMC increased. The blended cathodes of NMC and LMFP-1 also exhibited high rate capability at NMC : LMFP = 9 : 1, 8 : 2, 7 : 3, 6 : 4, and 5 : 5, and NMC : LMFP = 7 : 3 exhibited the highest rate capability. Further, the rate capability of the blended cathodes of NMC and LMFP-2 decreased drastically with the increase in the blending ratio of LMFP-2 to NMC. Thus, an optimal range for the blending ratio of LMFP to NMC exists based on the discharge capacity, energy density, and rate capability of the blended cathode. The differences in the electrochemical characteristics between the blended cathodes of NMC and LMFP-1 and those of NMC and LMFP-2 are because of the different electronic conductivity, BET surface area, mean particle size, and uniformity of the primary particles of LMFP-1 and LMFP-2. The physical and electrochemical characteristics of LMFP have a considerable impact on the electrochemical characteristics of the blended cathodes of NMC and LMFP.

Acknowledgments

This work was supported by Ms. Shiori Kuga of Hosei Industry Co., Ltd. for cathode fabrication.

CRediT Authorship Contribution Statement

Mayu Shiozaki: Writing – original draft (Lead)

Hiroki Yamashita: Conceptualization (Lead)

Yuko Hirayama: Data curation (Supporting)

Takaaki Ogami: Supervision (Lead)

Kiyoshi Kanamura: Supervision (Equal)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Footnotes

K. Kanamura: ECSJ Fellow

References

• 1)  M. Kotal, S. Jakhar, S. Roy, and H. K. Sharma, J Energy Storage, 47, 103534 (2022).
• 2)  H. Zhang, H. Zhao, M. A. Khan, W. Zou, J. Xu, L. Zhang, and J. Zhang, J. Mater. Chem. A, 6, 20564 (2018).
• 3)  D. Ren, Y. Yang, L. Shen, R. Zeng, and H. D. Abruña, J. Power Sources, 447, 227344 (2020).
• 4)  G. Mao, W. Yu, Q. Zhou, L. Li, Y. Huang, Y. Yao, D. Chu, H. Tong, and X. Guo, Appl. Surf. Sci., 531, 147245 (2020).
• 5)  H. J. Noh, S. Youn, C. S. Yoon, and Y. K. Sun, J. Power Sources, 233, 121 (2013).
• 6)  B. Xu, J. Lee, D. Kwon, L. Kong, and M. Pecht, Renewable Sustainable Energy Rev., 150, 111437 (2021).
• 7)  R. D. McKerracher, J. Guzman-Guemez, R. G. A. Wills, S. M. Sharkh, and D. Kramer, Adv. Energy Sustainability Res., 2, 2000059 (2021).
• 8)  X. Zhang, M. Hou, A. G. Tamirate, H. Zhu, C. Wang, and Y. Xia, J. Power Sources, 448, 227438 (2020).
• 9)  Z. Hong, H. Dong, S. Han, W. Li, Q. Dong, Y. Cao, X. Gao, Y. Zhang, W. Lu, and L. Chen, J. Power Sources, 512, 230505 (2021).
• 10)  L. Yang, W. Deng, W. Xu, Y. Tian, A. Wang, B. Wang, G. Zou, H. Hou, W. Deng, and X. Ji, J. Mater. Chem. A, 9, 14214 (2021).
• 11)  Y. J. Shin, J. K. Kim, G. Cheruvally, J. H. Ahn, and K. W. Kim, J. Phys. Chem. Solids, 69, 1253 (2008).
• 12)  W. Yang, Y. Bi, Y. Qin, Y. Liu, X. Zhang, B. Yang, Q. Wu, D. Wang, and S. Shi, J. Power Sources, 275, 785 (2015).
• 13)  A. M. Hashambhoy and J. F. Whitacre, J. Electrochem. Soc., 158, A390 (2011).
• 14)  D. V. Trinh, M. T. T. Nguyen, H. T. M. Dang, D. T. Dang, H. T. T. Le, H. T. N. Le, H. V. Tran, and C. D. Huynh, Sci. Rep., 11, 12280 (2021).
• 15)  J. Li, C. Guo, Y. Qin, and X. Ning, Mater. Technol., 35, 565 (2020).
• 16)  Y. Hirayama, J. Ikegami, H. Yamashita, T. Ogami, and K. Kanamura, Taiheiyo Cement Kenkyu Hokoku, 177, 41 (2019). [in Japanese]
• 17)  A. Yamada, Y. Kudo, and K. Y. Liu, J. Electrochem. Soc., 148, A1153 (2001).
• 18)  B. Xu, D. Qian, Z. Wang, and Y. S. Meng, Mater. Sci. Eng. Rep., 73, 51 (2012).
• 19)  D. D. Di Lecce and J. Hassoun, J. Phys. Chem. C, 119, 20855 (2015).
• 20)  H. Yu, S. Wang, Y. Hu, G. He, L. Q. Bao, I. P. Parkin, and H. Jiang, Green Energy Environ., 7, 266 (2022).
• 21)  J. U. Choi, N. Voronina, Y. K. Sun, and S. T. Myung, Adv. Energy Mater., 10, 2002027 (2020).
• 22)  W. Xiang, Y. J. Zhong, J. Y. Ji, Y. Tang, H. H. Shen, X. D. Guo, B. H. Zhong, S. X. Dou, and Z. Y. Zhang, Phys. Chem. Chem. Phys., 17, 18629 (2015).
• 23)  R. Chen, T. Zhao, X. Zhang, L. Li, and F. Wu, Nanoscale Horiz., 1, 423 (2016).
• 24)  M. Gaberscek, R. Dominko, and J. Jamnik, Electrochem. Commun., 9, 2778 (2007).
• 25)  H. Li and H. Zhou, Chem. Commun., 48, 1201 (2012).
• 26)  Y. Wang, Y. Wang, E. Hosono, K. Wang, and H. Zhou, Angew. Chem., 47, 7461 (2008).
• 27)  S. K. Martha, J. Grinblat, O. Haik, E. Zinigrad, T. Drezen, J. H. Miners, I. Exnar, A. Kay, B. Markovsky, and D. Aurbach, Angew. Chem., 48, 8559 (2009).
• 28)  A. M. Stux and K. E. Swider-Lyons, J. Electrochem. Soc., 152, A2009 (2005).
• 29)  C. Qiu, L. Liu, F. Du, X. Yang, C. Wang, G. Chen, and Y. Wei, Chem. Res. Chin. Univ., 31, 270 (2015).