**2. Olivine phosphate cathode**

Iron source is abundant in earth, it is inexpensive and less toxic than cobalt and nickel, using iron redox in cathode is promising to reduce cathode material cost. Inspired and encouraged by the successful development of LiCoO2 cathode, layered LiFeO2 prepared by ion-exchange from α-NaFeO2 presents iron redox activity [3]. But it is metastable and the performance was not attractive due to anion issues [4]. Comparing with other iron compounds, Fe3+/Fe2+ redox energy and potential can be adjusted by different anion groups. As consequence, polyanions were considered to build stable framework structure and improve redox stability. Due to Fe3+/Fe2+ redox potential is influenced by the corresponding anion group. A series of iron polyanion compounds were screened to search stable cathodes for lithium ion batteries [5]. In **Figure 1a**, Fe3+/Fe2+ redox energy is compared in different phosphates. The lower redox energy below Fermi level can achieve higher voltage compared with lithium. Fe3+/Fe2+ redox couple demonstrates the lower redox energy in LiFePO4 and higher redox potential at 3.5 V (*vs*. Li+ /Li). This voltage matches with stable window of carbonate electrolytes very well and accelerate the development of lithium iron phosphate. It delivered 110 mAh/g capacity firstly reported by Nobel laureate J. B. Goodenough at 1997 and has been successfully commercialized with decades of research and development.

#### **2.1 Lithium iron phosphate**

The ordered olivine structure of LiFePO4 is constructed by P-O framework with space group Pnma. Oxygen atoms are hexagonal-close-packed stacking order. Phosphorus atoms occupy tetrahedral sites, iron and lithium atoms locate at octahedral 4a and 4c position in Pnma space group. PO4 tetrahedral share one edge with FeO6 octahedron and two edges are shared with LiO6 octahedron. The FeO6

**Figure 1.**

*(a) Iron redox energy and potential can be adjusted by polyanion groups [5]. (b) LiFePO4 crystal structure [6].*

*Revisiting Olivine Phosphate and Blend Cathodes in Lithium Ion Batteries for Electric Vehicles DOI: http://dx.doi.org/10.5772/intechopen.99931*

octahedra are linked through corner in bc-plane and LiO6 octahedra form edgesharing chains in b direction [6]. Li resides in chains of edge-shared octahedra and connecting as the Li diffusion channel. PO4 polyanion framework is very stable in thermal dynamics as P-O bonding energy is high. During heating up to 350°C in N2 or O2 atmosphere [7], LiFePO4 and delithiated FePO4 structure were not changed which contributes to the high safety performance.

Li diffusion pathway in olivine LiFePO4 is one dimensional channel along [010] direction due to LiO6 octahedra is only continuous along b-axis. Although theoretical intrinsic ionic diffusion coefficient is 10−8–10−7 cm2 /s for both LiFePO4 and FePO4 [8, 9] the tested lithium ion diffusion coefficient is lower than 10−12 cm2 /s. As structure defects existing, such as Li/Fe anti-site defect, stacking fault and impurities, one-dimensional channels would be block and the ionic diffusion coefficient will be dramatically decreased [10]. In ordered olivine structure, corner-shared FeO6 does not form a continuous 3D network, and transition metal-d-oxygen-p hybridization in phosphate is weak [11]. Accordingly, the electronic conductivity of LiFePO4 at room temperature is only ~10−10 S/cm [12], which is usually thought as semiconductor [13]. Slow kinetics of lithium ion diffusion and low electronic conductivity lead to the poor rate capability of LiFePO4. Thus, numerous works have been devoted to overcome this drawback. Such as reducing particle size, element substitution, surface coating et al.

#### **2.2 Reaction mechanism**

Phase transformation mechanism in cathode during delithiation and lithiation is critical for electrochemical performance in lithium ion battery. When lithium ion extracting from LiFePO4, olivine LiFePO4 host will transform to FePO4 which have same structure. Even all active lithium ions are extracted out from LiFePO4, lattice volume only vary 6.5–6.8%, which demonstrate high crystal structure stability. Although the two phases reaction mechanism in lithiation/delithiation is widely accepted, the specific transformation route is reported as several models, which are highly related with material morphology, particle size, even experimental conditions [14, 15].

Core-shell (shrinking-core) model was proposed when LiFePO4 was initial reported by J. B. Goodenough at 1997 [7]. In lithiation process, lithium ions will be reserved from surface to particle core in FePO4 phase, two-phase boundary moves accordingly. In delithiation process, phase boundary will move reversibly from core to surface. In order to have deep and clear understanding on reaction mechanism, advanced operando/ex-situ characterization technologies are involved to study the structure changes furtherly. Beside two phases coexistence phenomenon (two phases distribution may be different), lithium deficient Li1-xFePO4 was seizing especially in nano size particle and high rate tests. Solid solution mechanism is reported to support the fast lithium transportation [16].

No matter phase transformation would undergo which mechanism in detail, the rigid P-O tetrahedral in structure provides rigid framework and contributes to highly reversible and stable delithiation/lithiation.

#### **2.3 Carbon coating**

Limited by the intrinsic low electronic conductivity, lithium utilization ratio in LiFePO4 is low (~0.6) and decays fast. Carbon coating was developed to improve material conductivity. This approach can be achieved through simple process by mixing carbon precursors with active materials followed by calcination. However, the electrochemical performance is influenced by the carbon coating quality.

Numerous works were reported to study and optimize carbon coating for LiFePO4 from different carbon sources, coating methods, carbon structure and carbon composite [17]. Beside enhancement of electronic conductivity, carbon coating can also be used to control LiFePO4 particle growth in calcination.

Beside conventional carbon coating, advanced carbon materials such as carbon nanotube, graphene and carbon fibers are introduced to form composite for pursuing high performance. Benefitting by excellent electronic conductivity from carbon coating layer and carbon composite, LiFePO4 material can be prepared into thick electrode without sacrifice performance [18].

#### **2.4 Lithium manganese phosphate**

Encouraged by well development of LiFePO4, pure phase of lithium manganese olivine is expected to improve energy density due to the high redox potential (4.1 V *vs*. Li+ /Li) of Mn3+/Mn2+ couple. However, lithium ion diffusion kinetics and electronic conductivity in LiMnPO4 is even worse than LiFePO4 and hard to prepare the pure phase with high performance [19, 20]. Morphology control and particle size reduction are effective solutions to improve the sluggish kinetics property referencing from LiFePO4 development [21]. A facile polyol synthesis approach is developed to prepare well-crystallized LiMnPO4 with ~30 nm thick nanoplates (as shown in **Figure 2**) [22]. High ratio of (020) plane which orientated in a-c plane and has short length along b-axis provides morphology to optimize kinetics property. Pure phase LiMnPO4 prepared by polyol method can deliver 159 mAh/g reversible capacity at 50°C, and retained 95% over 200 cycles. Elemental substitution was applied on LiMnPO4 to modify the olivine structure. Fe, Ni and Mg substitution made contribution to improve electronic conductivity, but only Fe substitution presents the positive function on electrochemical performance enhancement [24]. Poly synthesis method was further developed and cheap solvent DMSO was used to replace polyhydric alcohols in co-precipitation reaction. 50–100 nm Fe substituted LiMn0.8Fe0.2PO4 presents promising capability and cycling stability as presented in **Figure 2** [23, 25]. Considering kinetics of LiFePO4 is better than LiMnPO4, a coreshell structure of LiMn0.8Fe0.2PO4/C which has Fe rich on surface is prepared by adjusting co-precipitation process. LiFePO4 is controlled to grow on LiMnPO4 shell [26]. Comparing with LiMn0.8Fe0.2PO4 solid solution cathode, lithium ion diffusion kinetics is further enhanced due to the lower charge transfer resistance is achieved in this structure modification.

#### **Figure 2.**

*(a) (020) Oriented LiMnPO4 prepared by polyhydric alcohols approach [22]. (b) Developed co-precipitation method for synthesis of Fe substituted LiMnPO4 [23].*

*Revisiting Olivine Phosphate and Blend Cathodes in Lithium Ion Batteries for Electric Vehicles DOI: http://dx.doi.org/10.5772/intechopen.99931*

#### **Figure 3.**

*(a) XRD patterns of the LMP and LVP sample prepared by different ratios. (b) Influence of LVP amount on reversible capacity of LMP-LVP composite [29]. (c) XRD pattern of LMFP and LVP composite cathode [30]. (d) Rate capability of LMP and LVP composite cathodes [29].*

#### **2.5 Composite cathode**

Although the Mn doped LiFe1-xMnxPO4 can deliver higher energy density contributed by the higher redox potential of Mn3+/Mn2+, rate performance is geared down when increasing manganese ratio. NASICON structured phosphate Li3V2(PO4)3 has open lattice framework which guarantee the fast Li ion transportation in cathode bulk phase [27]. As consequence, high rate NISICON cathode is integrated with olivine phosphate to improve Li diffusion kinetics [28]. In LiMnPO4-LVP composite cathode prepared by solid state approach, both the NISSCON and olivine phase diffraction peaks and plateau character could be identified even LVP ratio is below 3% [29]. As the composite LVP ratio is higher than 20%, the capacity contributed from Mn2+/Mn3+ plateau is increased and indicates that LVP composite can improve activity of Mn redox in olivine phase. Checking the lattice paraments, olivine phase LiMnPO4 is substituted by a small amount of vanadium [30]. The composite cathode presents enhanced rate capacity (**Figure 3**).

#### **3. Blend cathodes**

The present commercial cathode layered oxide, spinel oxide and olivine phosphate have their own advantage and have already been successfully utilized in different lithium ion battery designs according to the application scenarios. The blend of different cathode is a facile method to tailor the properties and performance of electrodes for lithium ion batteries.

#### **3.1 Blend cathode – olivine phosphate and layered oxide**

Layered structure oxide contained nickel cobalt manganese (NMC) has much higher theoretical capacity (~270 mAh/g) than olivine phosphate cathodes (~170 mAh/g). Limited by the delithiated structure stability at high voltage and electrolyte decomposition window [31, 32], only <0.8 Li in layered oxide cathode will participate in charge–discharge process. Although NMC cathode can be used to prepare high energy density lithium ion batteries for electric vehicles, the safety concerns arise comparing with the one using olivine phosphate cathodes [33, 34]. Blend cathode including both olivine and layered oxide cathode would a moderate strategy to integrate.

LiMn0.8Fe0.2PO4/C was mixed directly with LiNi0.88Co0.09Al0.03O2 (NCA) to improve the comprehensive performance. The reasonable ratio of LiMn0.8Fe0.2PO4/C can extend cycling life of Ni rich cathode without sacrificing capacity at electrode level [35]. Adding 1% and 2% LiMn0.8Fe0.2PO4/C, electrodes present similar discharge capacity. But the capacity retention ratio is increased from 77% using pristine NCA to 88% using 2% olivine addition. Using LiNi0.5Mn0.3Co0.2O2 (NMC532) and LiFe0.15Mn0.85PO4 blend cathode, cycling stability is enhanced when LFMP/C ratio in blend cathode is less than 10% [36].

LiCoO2 presents more stable structure stability with adding LiFePO4, both the particle cracking and irreversible phase transformation are inhabited even cycling at high cutoff voltage. Mixing with 12 wt% LiFePO4, discharge capacity of blend cathode is decreased to 177.9 mAh/g at 0.2C charging to 4.5 V (*vs*. Li+ /Li). However, overpotential of cobalt redox is reduced and presents enhanced cycling stability [37].

Redox dynamics in LiFePO4 and LiCoO2 blend cathode is different with individual component. At low C-rate, each components of the blend cathode can independently present their redox reaction. When working at high C-rate, the effective C-rate for each constituent is higher than nominal value. However, the rate performance of blend cathode is significantly improved [38]. Buffer effect is proposed that the internal pathway would be constructed by component with fast reaction kinetics as shown in **Figure 4**. Lithium is redistributed among the constituents to reach an equilibrium potential during relaxation.

Despite electrochemical performance is improved by blending, LiFePO4 and layered oxide cathode blend cathode may have heterogeneity issue as material property such as particle size and density are different. 3D X-ray tomography technology is

#### **Figure 4.**

*a) Cyclic voltammetry of a blend electrode illustrating the contributions of LiFePO4 and LiCoO2. (scanning rate 25* μ*V/s). b) and c) potential profile and specific current during discharge and subsequent relaxation of a LiFePO4 and LiCoO2 blend at 0.2C and 5C [38].*

*Revisiting Olivine Phosphate and Blend Cathodes in Lithium Ion Batteries for Electric Vehicles DOI: http://dx.doi.org/10.5772/intechopen.99931*

**Figure 5.** *Schematic figure of typical features in a) NMC, b) LFP, and c) NMC/LFP blend electrodes [39].*

used to check the composition distribution in prepared blend electrodes [39, 40]. Small NMC, nano sized LiFePO4 and carbon additives are enriched in layer close to current collector and large particle NMC is depleted in this region. Although electrode composition is close to design value above the NMC depleted layer, the mean NMC particle size presents gradient distribution toward top surface, as demonstrated in **Figure 5**. Beside local morphology heterogeneities, LiFePO4 agglomeration and cavities are observed in NMC rich zones. The heterogeneity issue may can be resolved by technique and equipment optimization, but the reported results reveal that materials properties differences should be considered to prepare uniform electrodes with high quality especially for blend cathodes.

#### **3.2 Blend cathode – olivine phosphate and spinel oxide**

Spinel oxide cathode has advantage of low cost, high thermal stability and high rate capability. However, spinel oxide suffers from fast capacity fading in long cycling test, especially at high temperature. Olivine phosphate has excellent cycling stability and could blend with spinel to pursue the comprehensive performance.

LiFexMn1-xPO4 (LFMP) and spinel oxide blend cathode demonstrates linear changing on tap density, reversible capacity, energy density and power density at low C-rate by adjusting blend ratio between two components. However, synergetic effect was observed in blend cathode at high C-rate [41]. Discharge curves of blend cathode at 3C presents lower polarization for Mn3+/Mn2+ plateau as shown in **Figure 6**. Structure change was studied by in-situ XRD characterization during 3C pulse discharge and the following relaxation. (004) diffraction peak of spinel cathode shifts toward lower angle during discharge and moves to high angle in relaxation period. This result indicates that spinel component is reoxidized in relaxation step. As lithium diffusion in LFMP is slower than spinel LiMn2O4, the higher delithiated state LFMP will provide lithium ions to spinel for reaching a common

#### **Figure 6.**

*(a) Discharge curve comparison of blend cathode and individual component [42]. (b) Schematic illustration of buffer effect in blend cathode during pulse power test [41].*


 *Characteristic properties of electrodes with an aerial capacity of approximately 2.5 mAh cm−2 and a porosity of 30% calculated from the electrodes' thickness. Mass loading includes active materials, binder, and conductive carbon. Specific capacity is based on 93% active material (AM) content. The proportions indicated in the second column are based on the total AM [45].* *Revisiting Olivine Phosphate and Blend Cathodes in Lithium Ion Batteries for Electric Vehicles DOI: http://dx.doi.org/10.5772/intechopen.99931*

equibium potential state (demonstrated in **Figure 6**). This buffer effect reduce the electrode polarization and improves the power pulse capability. Energy density and power density at high C-rate are enhanced contributed by the reduction of electrode polarization by blend ratio optimization [42].

Manganese dissolution is notorious for lithium ion batteries using spinel LiMn2O4 cathodes [43]. Dissolution mechanism is generally ascribed to the presence of HF generated by the reaction between hexafluorophosphate anion and water impurity. Dissolved manganese will be reduced/deposited on anode side and involved in SEI formation in graphite surface. Spinel cathode will loss reversible capacity by manganese dissolution and anode impedance will be dramatically increased by manganese migration and deposition, as result spinel cathode performance in lithium ion battery is poor, especially calendar life and cycling stability at high temperature. This issue can be alleviated by using blend cathode design. Layered structure cathodes such as LiCoO2 and NMC reported that can work as proton scavenger through Li ion and proton exchanging in blend cathodes [44]. In LFMP and spinel oxide cathodes, dissolved manganese may be precipitation on LFMP particle surface instead of migrating to anode side. This results is also approved in LiFe0.3Mn0.7PO4 and LiMn1.9Al0.1O4 blend cathode that manganese dissolution from spinel cathode is dramatically reduced [42].

#### **3.3 Ternary blend cathodes**

Based on the development of binary blend cathode, ternary blend cathode composited by olivine phosphate, layered oxide and spinel oxide are proposed to study. Mixing three components at even ratio (33% for each), the ternary blend electrode presents highest electronic conductivity when all the other electrode parameters are controlled at same level (as list in **Table 1**) [45]. Redox activities studies reveal that each components in blend cathode can work independently and contributes to reduce the over-potential [46]. This is consistent with other reports.

### **4. Safety**

Along energy density of lithium ion battery increasing, safety concerns are serious raised. Safety performance and evaluation is one of the critical criteria for the practical utilization of novel battery materials/technologies. Although battery management systems (BMS) are equipped on electrical vehicles to monitor all the cells and ensure the safety operation, battery still has thermal runway risk and would cause disaster results. Many countries published the strict compulsory standards for battery safety tests before releasing to commercial utilization on EVs [47]. Therefore, it is important to pay attention on safety development as well as energy/ power density improvement.

Lithium ion battery is a comprehensive system and sensitive to temperature. When battery temperature rising to the cathode decomposition triggering point, the exothermic reaction will abruptly release heat and accelerate thermal runaway which is the detrimental safety issue. Thermal studies of materials components in lithium ion batteries can be investigated by differential scanning calorimetry (DSC) or accelerating rate calorimetry (ARC) [48]. As cathode material property are different, the onset reaction temperature and released heat are varied. In brief, the onset temperature for cathodes follow the order: LiFePO4 > LiNixCoyMnzO2 > LiCoO2 as compared in **Figure 7** [47]. Polyanionic phosphate cathodes have robust P-O covalent constructing stable three-dimensional framework, it reduces the reactivity between cathode and electrolyte.

**Figure 7.**

*The energy release diagram of different cathode in lithium ion batteries [47].*

#### **Figure 8.**

*(a–d) Schematic description of Li redistribution among the components induced by a temperature change [38]. (e) Thermal performance comparison in LFMP/C and NMC523 blend cathode [36]. (f) Differential scanning calorimetry of fully charged blend electrodes (red) comparing with individual cathode at different ratios [45].*

Although spinel is reported that onset temperature is little higher than LiFePO4 [49], but the exothermic reaction enthalpy is higher than LiFePO4 and the peak shape is more sharp which means more heat will be generated in shorter time.

*Revisiting Olivine Phosphate and Blend Cathodes in Lithium Ion Batteries for Electric Vehicles DOI: http://dx.doi.org/10.5772/intechopen.99931*

Although electrochemical performance of layered oxide cathodes have been improved a lot by material modification [33, 50, 51], they are still hard to compete with LiFePO4 on safety performance evaluation. Blend cathode is a reasonable method to balance the comprehensive performance. By adding olivine phosphate cathode LiFe0.15Mn0.85PO4/C in NMC523 cathode, the exothermic reaction enthalpy is significantly decreased [36]. But there is no linearly relationship between LFMP adding ratio and thermal performance (**Figure 8**).

## **5. Summary**

Olivine phosphate has been successfully commercialized and used in lithium ion batteries to equip on electrical vehicles. Although energy density is not competitive comparing with using layered oxide cathode especially Ni rich cathode, olivine phosphate present excellent safety performance. Manganese substitution and composited with lithium vanadium phosphate is carried out to improve energy density and power density. Beside this, blend cathode is an effective strategy to improve energy density without sacrificing safety advantage. Binary and ternary blend cathode present the possibility to tailor cathode property and performance. Blend cathodes show synergetic effect on rate capability and thermal stability tests which is higher than nominal value. Structure and electrochemistry studies reveal that buffer effect in blended components is contributed to the improvement. This chapter provides opinions from material science and electrochemistry viewpoints to understand the requirement of lithium ion battery on cathode materials development.

### **Conflict of interest**

The authors declare no conflict of interest.

### **Author details**

Yujing Bi1 \* and Deyu Wang1,2\*

1 Ningbo Institute of Materials Technology and Engineering, CAS, Ningbo, China

2 Jianghan University, Wuhan, Hubei, China

\*Address all correspondence to: biyujing15@hotmail.com and wangdeyu@jhun.edu.cn

© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
