저작자표시 - 비영리 - 변경금지 2.0 대한민국 이용자는아래의조건을따르는경우에한하여자유롭게 이저작물을복제, 배포, 전송, 전시, 공연및방송할수있습니다. 다음과같은조건을따라야합니다 : 저작자표시. 귀하는원저작자를표시하여야합니다. 비영리. 귀하는이저작물을영리목적으로이용할

Transcription

1 저작자표시 - 비영리 - 변경금지 2.0 대한민국 이용자는아래의조건을따르는경우에한하여자유롭게 이저작물을복제, 배포, 전송, 전시, 공연및방송할수있습니다. 다음과같은조건을따라야합니다 : 저작자표시. 귀하는원저작자를표시하여야합니다. 비영리. 귀하는이저작물을영리목적으로이용할수없습니다. 변경금지. 귀하는이저작물을개작, 변형또는가공할수없습니다. 귀하는, 이저작물의재이용이나배포의경우, 이저작물에적용된이용허락조건을명확하게나타내어야합니다. 저작권자로부터별도의허가를받으면이러한조건들은적용되지않습니다. 저작권법에따른이용자의권리는위의내용에의하여영향을받지않습니다. 이것은이용허락규약 (Legal Code) 을이해하기쉽게요약한것입니다. Disclaimer

2 Phase evolution study of lithiumrich composite oxide for Li-ion batteries

3 Abstract Phase evolution study of lithiumrich composite oxide for Li-ion batteries Byungjin Choi Department of Materials Science and Engineering Seoul National University The composite material Li x Ni 0.25 Co 0.10 Mn 0.65 O (3.4+x)/2 (x=1.6, 1.4, 1.2, 1.0, 0.8) were synthesized and characterized for their structural, morphological, and performance as cathode materials in Li-ion batteries. The Rietveld refinement results indicate the presence of two phases at high lithium levels (x=1.6 and 1.4): Li 2 MnO 3 (C2/m) and LiMO 2 (M = Ni, Co, Mn) (R3 m); the latter contains Ni 2+ and Ni 3+. At low lithium levels (x=1.2, 1.0, and 0.8) an additional spinel phase LiM 2 O 4 (Fd3 m) emerges, which is known to affect the electrochemical performance of the oxide. Structural analysis reveals that the spinel phase contains mixed transition metals Ni, Co, and Mn as [Li +,Co 2+ ][Ni 2+,Co 3+,Mn 4+ ] 2 O 4. A low lithium level is found to induce primary particle growth, as well as Co and Ni segregation within the secondary particles. These results are expected to contribute to material i

4 optimization and commercialization of lithium-rich oxide cathodes. The composite material Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiM 0.5 Mn 1.5 O 4 (M = Mn, Ni, Co) were synthesized and characterized for their structural, morphological, and performance as cathode materials in Li-ion batteries. XRD analysis indicates the presence of Li 2 MnO 2 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m), and spinel phase LiM 0.5 Mn 1.5 O 4 (M = Mn, Ni, Co) (Fd3 m). In LiM 0.5 Mn 1.5 O 4 (M = Mn) (Fd3 m) composition spinel LiMn 2 O 4 phase is embedded. At 20mol% embedding additionally LiNi 0.5 Mn 1.5 O 4 phase is also detected. Rocksalt NiO phase is formed in LiM 0.5 Mn 1.5 O 4 (M = Ni) composition even in oxygen atmosphere. LiM 0.5 Mn 1.5 O 4 (M = Co) composite composition is homogeneously synthesized even in 20mol% embedding. By embeddinglico 0.5 Mn 1.5 O 4 phase the electrochemical performance in full cell using graphite anode is improved. Spinel embedding in lithium-rich composite oxide can improve the electrochemical performance through structural stability. The phase content, crystal size and lattice parameters were analyzed through the Rietveld refinement in LiM 0.5 Mn 1.5 O 4 (M = Co) composite composition. Spinel embedding induces primary particle growth during heat-treatment, as well as Co and Ni segregation within the secondary particles. Ab initio calculation shows that spinel embedding in lithium-rich composite oxide can lower the formation energy by stabilizing the structure. The phase evolution process was analyzed during high temperature XRD method. Keywords: Li-ion batteries, lithium-rich oxides, spinel structure, Co segregation, Li 2 MnO 3 ii

5 Student Number: iii

6 Contents Abstract... i List of Tables... vii List of Figures... ix Chapter 1. Introduction (Theoretical Review) Lithium ion battery 1.2 Cathode material for lithium ion battery Chapter 2. Phase evolution of lithium-rich oxide Introduction 2.2 Experimental Procedure Synthesis Instrumental Characterization Computational Methods Electrochemical Measurement iv

7 2.3 Results and discussion Material Characterization Identification of Composite phases and Their Growth Mechanism Conclusions Chapter 3. Spinel phase composite of lithium-rich oxide Introduction 3.2 Experimental Procedure Synthesis Characterization and evaluation Computational Methods Results and Discussion Spinel LiMn 2 O 4 composite Spinel LiNi 0.5 Mn 1.5 O 4 composite Spinel LiCo 0.5 Mn 1.5 O 4 composite v

8 3.4. Conclusions Chapter 4. Conclusion References vi

9 List of Tables Table 2.1 Elemental analysis for the materials under study Table 2.2 The results of the Rietveld refinement for the materials under study Table 2.3 XAFS analysis at the Ni K edge (d), Co K edge (e), and Mn K edge (f) for Li 1.60, Li 1.40, Li 1.20, Li 1.00, and Li Table 2.4 Rate capability tests of (1) Li1.60, (2) Li1.40, (3) Li1.20, (4) Li1.00, (5) Li0.80 at different currents (C rates) in the potential range V in coin type cells. Li counter electrodes. Table 2.5 Composition calculation for 5 compositions with different lithium levels. Labels (1)-(4) refer to the spinel phase types LiNi 0.5 Mn 1.5 O 4, LiCo 0.5 Mn 1.5 O 4, LiMn 2 O 4 and [Li +,Co 2+ ][Ni 2+,Co 3+,Mn 4+ ] 2 O 4 respectively. Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, (4) 10, and (5) 20mol% heat-treated at 900degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes. Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, (4) 10, and (5) 20mol% heat-treated at 750degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes. Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, (4) 10, and (5) 20mol% heat-treated at 900degC in air at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes. Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, and vii

10 (4) 10mol% heat-treated at 900degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes. Table The results of the Rietveld refinement for the materials under study Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, and (4) 10mol% heat-treated at 750degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes. Table Summary of spinel composites. viii

11 List of Figures Figure 1.1 Schematic drawing of lithium ion battery operation. Figure 1.2 Initial charge-discharge profiles measured upon galvanostatic cycle. Figure 1.3 Schematics of operational voltage difference in various cathode materials. Figure 1.4 Compositional phase diagram of lithium-rich Li(Ni,Co,Mn)O 2. Figure 1.5 Voltage decay of lithium-rich Li(Ni,Co,Mn)O 2 during chargedischarge cycling. Figure 1.6 Spinel phase evolution in lithium-rich Li(Ni,Co,Mn)O 2. Figure 2.1 (a) Panels (1)-(5) are XRD patterns of Li1.60, Li1.40, Li1.20, Li1.00 and Li0.80 respectively, after heat-treatment at 900 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks. Figure 2.2 (a) High-resolution TEM image taken from the Li1.00 composition. (b) and (c) Local structural information from Fast-Fourier Transformations of regions 1 region 2, respectively. The extra row of spots indicated by the arrow in (c) show a spinel structure. Figure 2.3 Rietveld refinement for the phase contents in Li1.60-Li0.80 compositions. The fractions are calculated with the composite notation like Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiM x Mn 2-x O 4 (M=Ni, Co, Mn). Figure 2.4 XANES spectra of Li1.60-Li0.80 compositions at (a) the Ni K- edge, (b) Co K-edge, and (c) Mn K-edge. XAFS spectra at (d) the Ni K-edge, (e) Co K-edge, and (f) Mn K-edge. ix

12 Figure 2.5 (a) XRD patterns of LiMn 2 O 4 and LiNi 0.5 Mn 1.5 O 4 heat-treated at 900degC. (b) and (c) Selected XRD angular range highlighting the difference between LiMn 2 O 4 and LiNi 0.5 Mn 1.5 O 4. Figure 2.6 FESEM images of (a0) Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 precursor and (a)- (e) Li1.60-Li0.80 compositions heat-treated at 900 C. Figure 2.7 Cross-sectional EPMA mapping and images of (a) Li1.40, (b) Li1.20, and (c) Li1.00. A scale bar is located at the lower right corner. Figure 2.8 a)-(e) Initial charge-discharge profiles measured upon galvanostatic cycle of Li1.60-Li0.80 compositions. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes. Figure 2.9 Plots of specific capacity vs. cycle number of (1) Li1.60, (2) Li1.40, (3) Li1.20, (4) Li1.00, (5) Li0.80 at 1C rate (230mA/g) in the potential range V in coin type cells. Li counter electrodes. Figure 2.10 Initial charge-discharge profiles measured upon galvanostatic cycle of (a) Li1.60, (b) Li1.40, (c) Li0.80 at 23mA/g (C/10) rate in the potential range of V in coin type cells, Li counter electrodes. Figure 2.11 Schematic drawing showing the proposed phase evolution process for high and low lithium compositions. Figure Synthesis procedure for spinel composite oxide Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C. Selected angular ranges x

13 highlight the (b) layered and (c) spinel peaks. Figure FESEM images of spinel composites (a) 0, (b) 2, (c) 5, (d) 10, and (e) 20mol% heat-treated at 900 C. Figure (a)-(d) Initial charge-discharge profiles measured upon galvanostatic cycle of spinel 0-20mol% compositions heat-treated at 900degC. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes. Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks. Figure FESEM images of spinel composites (a) 0, (b) 2, (c) 5, (d) 10, and (e) 20mol% heat-treated at 750 C.. Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C in air. Selected angular ranges highlight the (b) layered and (c) spinel peaks. Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C in air. Selected angular ranges highlight the (b) layered and (c) spinel peaks. Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C in oxygen. Selected xi

14 angular ranges highlight the (b) layered and (c) spinel peaks. Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C in oxygen. Selected angular ranges highlight the (b) layered and (c) spinel peaks. Figure (a)-(d) Initial charge-discharge profiles measured upon galvanostatic cycle of spinel 0-20mol% compositions heat-treated at 900degC in air. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes. Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks. Figure Rietveld refinement for the phase contents in (1- x)[0.4li 2 MnO 3 0.6Li(Ni,Co,Mn)O 2 ] xlico 0.5 Mn 1.5 O 4 compositions. The fractions are calculated with the composite notation like Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiM x Mn 2-x O 4 (M=Ni, Co, Mn). Figure FESEM images of spinel composites (a) 0, (b) 5, and (c) 20mol% heat-treated at 900 C. Figure Cross-sectional EPMA mapping and images of spinel 20mol% composite heat-treated at 900degC. A scale bar is located at the right corner. Figure (a)-(c) HAADF image and EDS mapping taken from spinel xii

15 20mol% heat-treated at 900degC. (d) High-resolution TEM image taken from spinel 20mol% heat-treated at 900degC. Figure XPS data taken from spinel 20mol% heat-treated at 900degC. Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks. Figure FESEM images of spinel composites (a) 0, (b) 5, and (c) 20mol% heat-treated at 750 C. Figure Cross-sectional EPMA mapping and images of spinel 20mol% composite heat-treated at 750degC. A scale bar is located at the right corner. Figure (a)-(c) Initial charge-discharge profiles measured upon galvanostatic cycle of spinel 0-10mol% compositions heat-treated at 750degC. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes. Figure Full cell cycle life (a) discharge capacity retention (b) voltage decay in type full cells with graphite anode. Figure (a) Panels (1)-(4) are XRD patterns of spinel 0mol%, 20mol% LiMn 2 O 4, 20mol% LiNi 0.5 Mn 1.5 O 4 and 20mol% LiCo 0.5 Mn 1.5 O 4 respectively, after heat-treatment at 900 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks. xiii

16 Figure Formation energy of spinel embedded composite. Figure High temperature XRD of spinel embedded composite. Figure Schematic drawing showing the proposed phase evolution process for spinel embedded composite. xiv

17 Chapter 1. Introduction (Theoretical Review)

18 1.1 Lithium ion battery Lithium ion battery operates at high voltage 3-4V range compared with other battery systems. High voltage operation can deliver more energy to consumer electronics and electric vehicles. The battery components are cathode, anode, electrolyte, and separator. Lithium ions intercalate between cathode and anode. Cathode material can accept the lithium ion in its crystal structure reversibly. LiCoO 2 as cathode and graphite as anode system was commercialized by Sony at 1991.[1, 2] During charging lithium ions are extracted from LiCoO 2 cathode and intercalated into graphite anode in Figure 1.1. The capacity means the amount of lithium which can be extracted from cathode, and the operational voltage is the difference in chemical potential between cathode and anode in Figure Cathode material for lithium ion battery For high capacity and voltage cathode materials are required to have light weight and high chemical potential. LiMO 2 (M = transition metal) layered structure is used for lithium ion battery cathode material. The 3d transition metals are used for M in LiMO 2. LiCoO 2 and LiNiO 2 are wellknown for layered structure. Spinel structure LiM 2 O 4 and olivine structure LiMPO 4 are also used for lithium ion battery cathode, but the applications are limited due to lower capacity and rate performance. The operational voltage in different cathode materials is shown in Figure 1.3.

19 Among layered cathode material recently Ni-rich Li(Ni,Co,Mn)O 2 and Li-rich and Mn-rich Li(Ni,Co,Mn)O 2 are researched for high energy density battery. In Ni-rich Li(Ni,Co,Mn)O 2 the high capacity is achieved Ni 2+/4+ redox reaction during charge-discharge. In Li-rich and Mn-rich Li(Ni,Co,Mn)O 2 the high capacity is achieved through activation of oxygen in lithium-rich Li 2 MnO 3 phase during high voltage activation above 4.5V in Figure The aim of this thesis The demand for Li-ion batteries (LIBs) has increased rapidly owing to their relatively high energy density and design flexibility. LIBs are attractive energy sources for applications ranging from mobile devices to large-scale products such as electric vehicles (EV) and energy-storage systems (ESS).[3] However, the limited capacities from typical cathode materials such as LiCoO 2 and LiMn 2 O 4 cannot satisfy the high energy density requirements for high-power LIBs used in EV and hybrid electric vehicles (HEV). Lithium ion battery is widely used for consumer electronics, electric vehicles, and ESS application. Higher energy density is the key requirement to expand the application. Layered oxides are the most balanced material to meet the high energy and safety requirement. Lithium-rich Li(Ni,Co,Mn)O 2 has been researched for commercialization because it shows high capacity above 250mAh/g.[4-6] The bottleneck for commercialization is the phase stability during charge-discharge condition.

20 During charge-discharge the structure is changed from layered to spinel. It affects power performance by decreasing average voltage in Figure 1.5. To improve the voltage decay various approaches have been tried. Doping, surface coating, and composite structure improved the performance even though much improvement is strongly needed. Spinel formation is reported for structural stability in Figure 1.6. By optimizing the spinel phase in lithium-rich Li(Ni,Co,Mn)O 2 various approaches are attempted. The spinel phase evolution mechanism has to be clarified and electrochemical performance improvement is needed. In chapter 2, the phase evolution during heat-treatment condition was investigated by controlling lithium content in precursor Ni 0.25 Co 0.10 Mn 0.65 (OH) 2. Structural analysis and ab initio calculation were used for clarifying phase evolution and stable phase structure. In chapter 3, various spinel phases are embedded to improve the electrochemical performance. In precursor Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 the spinel forming transition metal precursors(mnco 3, Co(OH) 2, Ni(OH) 2 ) were added and melt-infiltrated through heat-treatment with Li 2 CO 3. The composite structure and morphology were analyzed. The stable phase formation mechanism was explained. The electrochemical performances were analyzed using coin half-cell with lithium metal and full-cell with graphite.

21 Figure 1.1 Schematic drawing of lithium ion battery operation. [7]

22 Figure 1.2 Initial charge-discharge profiles measured upon galvanostatic cycle.

23 Figure 1.3 Schematics of operational voltage difference in various cathode materials.

24 Figure 1.4 Compositional phase diagram of lithium-rich Li(Ni,Co,Mn)O 2.[8]

25 Figure 1.5 Voltage decay of lithium-rich Li(Ni,Co,Mn)O 2 during chargedischarge cycling.

26 Figure 1.6 Spinel phase evolution in lithium-rich Li(Ni,Co,Mn)O 2. [9]

27 Chapter 2. Phase evolution of lithium-rich oxide

28 2.1 Introduction Lithium-rich layered composite oxide cathode materials, referred to as over-lithiated layered oxides (OLO) in the composite system Li 2 MnO 3 LiMO 2 (M = Ni, Co, Mn), have shown capacities exceeding 250 mah/g at high operating voltages (>3.5 V vs. Li/Li+). However, lithium-rich oxide cathodes have several major drawbacks, including capacity loss during the first cycle, poor rate capability, and decreased cyclic performance. [3, 10] Various approaches, such as surface modification and composition change including doping, have been used to overcome these problems.[11] Surface modification was used to block side reactions between the electrolyte and cathode at high voltage, and suppress phase changes during the cycle. [12-15] Others have focused on optimizing the composition to reduce transition metal migration during cycling in order to maintain the high capacity (>250 mah/g). [16-19] After screening various compositions, our research group has focused on the cobalt-containing Li 1.40 Ni 0.25 Co 0.10 Mn 0.65 O 2.40 baseline composition for EV applications. [20] Similar compositions have been studied by other groups and have been demonstrated to show promising performance.[3, 16-19, 21-25] However, a systematic study of the structure and phase evolution in these compositions is still lacking, [3, 16-19] even though compositions with only Ni and Mn transition metals have been investigated extensively. [26-31] The addition of cobalt to lithium-rich oxides induces structural and morphological changes, especially in the spinel phase formation. [3, 10, 16-19, 21-25] In the layered-layered-spinel (Li 2 MnO 3 LiMO 2 LiM x Mn 2-

29 xo 4 (M = Ni, Co, Mn)) composite structure, the initially embedded spinel phase has been shown to improve the voltage decay by stabilizing the layered-layered (Li 2 MnO 3 LiMO 2 (M = Ni, Co, Mn)) composite.[9, 32] The reported spinel phase in lithium-rich oxide are LiMn 2 O 4 and LiNi 0.5 Mn 1.5 O 4.[9, 26, 29, 33, 34] When cobalt is included, it is important to identify the spinel structure and composition, including the mixed transition metal spinel phase. [35, 36] This work investigates the structural, morphological, and electrochemical performance changes in Li x Ni 0.25 Co 0.10 Mn 0.65 O (3.4+x)/2 (x = 1.6, 1.4, 1.2, 1.0, 0.8). For brevity, they may be referred to as Li1.60 Li0.80 throughout the rest of the article. The analytical methods used here include high-resolution powder diffractometry (HRPD), scanning electron microscope (SEM), transmission electron microscope (TEM), X-ray absorption near edge structure (XANES), extended X-ray absorption fine structure (EXAFS), and electron probe microscopic analysis (EPMA). 2.2 Experimental Procedure Synthesis The Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 precursor was prepared through a hydroxide co-precipitation process. Proper amounts of NiSO 4 6H 2 O, CoSO 4 7H 2 O, and MnSO 4 H 2 O were stirred in deionized water to form a homogeneous solution. The solution was chelated using NH 4 OH and precipitated with NaOH. The co-precipitated Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 after

30 drying was mixed with Li 2 CO 3 to form the composite material with the average composition Li x Ni 0.25 Co 0.10 Mn 0.65 O (3.4+x)/2 (x = 1.6, 1.4, 1.2, 1.0, 0.8). [21, 29, 37] Their specific formulae are as follows: Li 1.60 Ni 0.25 Co 0.10 Mn 0.65 O 2.50 (or Li Ni Co Mn O 2 ) Li 1.40 Ni 0.25 Co 0.10 Mn 0.65 O 2.40 (or Li Ni Co Mn O 2 ) Li 1.20 Ni 0.25 Co 0.10 Mn 0.65 O 2.30 (or Li Ni Co Mn O 2 ) Li 1.00 Ni 0.25 Co 0.10 Mn 0.65 O 2.20 (or Li Ni 0.25 Co 0.10 Mn 0.65 O 2 ) Li 0.80 Ni 0.25 Co 0.10 Mn 0.65 O 2.10 (or Li Ni Co Mn O 2 ) The chemical formula in the parenthesis is the layered structure nomenclature of solid solution. The mixed powders were then calcined at 900 C for 10 h in flowing air Instrumental Characterization The crystal structures of the powder samples were determined by synchrotron HRPD performed at 9B HRPD beamline at Pohang Light Source II (PLS-II, Pohang, Korea) using the wavelength λ= The data were refined using the FULLPROF program. XANES and EXAFS measurements were carried at 7D-XAFS beamline in Pohang Accelerator Laboratory, Korea. The XANES and EXAFS data were analyzed by established methods using the ATHENA software package. [38] Morphology changes of the powders were determined using SEM (S-4700N, Hitachi). The inductively coupled plasma technique (ICP-AES) was used to determine the ratios of Li, Ni, Co, and Mn elements in each sample. In order to observe the transition metal distributions in the secondary particles, cross sections of the powder particles were prepared by Ar ion milling on a LN2 cooled stage

31 and measured by an Electron Probe MicroAnalyzer (EPMA, JEOL JXA- 8530F). The atomic-level structure and local phases were identified by diffraction and high-resolution TEM (FEI, Titan-cubed ) Computational Methods The first principle calculations were performed using the Vienna ab initio simulation package (VASP) [39, 40] with the Projector-Augmented- Wave (PAW) method. [41] The exchange correlation interactions were included with the generalized gradient approximation Perdew-Burke- Ernzerhof (GGA-PBE) functional, [42] and the plane wave cutoff energy was set to 500 ev. The structure relaxations were carried out with a criteria of 10-4 ev for the total energy, and 0.02eV/Å for the forces on each atom. The effective on-site Hubbard U eff corrections were 6.885, 5.95 and 5.0eV on the 3d electrons for Ni, Co and Mn atoms, respectively. [43] The supercells selected in this work contained between 8 and 28 unit cells, depending on the sample composition and structure Electrochemical Measurement The electrodes were prepared by making a slurry of 92 wt% active material (Li x Ni 0.25 Co 0.10 Mn 0.65 O (3.4+x)/2), 4 wt% conductive Denka Black, and 4 wt% polyvinylidene difluoride (PVDF) binder in N-methyl-2-pyrrolidone (NMP) as a solvent. The slurry was coated using doctor-blade method onto Al foil as a current collector. The electrodes were then dried at 120 C in vacuum and pressed. Metallic Li was used as the anode. The electrolyte solution was 1.3 mol L -1 LiPF 6 dissolved in fluoroethylene carbonate and

32 dimethylene carbonate. A porous polyethylene-based membrane was used as a separator. The above components were assembled into CR2032-type coin cells in a dry room. The typical loading of the active mass was 10 mg/cm 2. The cells were charged to 4.7 V for one cycle and then cycled between 2.5 and 4.6 V vs. Li/Li Results and discussion Material Characterization The XRD patterns of Li x Ni 0.25 Co 0.10 Mn 0.65 O (3.4+x)/2 samples (x = 1.6, 1.4, 1.2, 1.0, 0.8) heat-treated at 900ºC are shown in Figure 2.1. The oxygen stoichiometry was calculated to balance the total positive charge from the most stable oxidation states of Li +, Ni 2+, Co 3+, and Mn 4+.[37, 44, 45] In Figure 1 (b), peaks of the Li 2 MnO 3 (C2/m), LiMO 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Ni, Co, Mn) were detected. The designed and calculated structure compositions are listed in Table 1. Note that different types of the spinel phase can lead to a number of possible compositions, which will be explained later. The actual elemental compositions obtained by ICP-AES in Table 2.1 agree with the designed stoichiometry within experimental error. The Rietveld refinement was performed in order to clarify the phase contents and crystalline structures in the powder samples. Table 2.2 summarizes the lattice parameters of Li 2 MnO 3 (C2/m), LiMO 2 (R3 m), and LiM x Mn 2-x O 4 (Fd3 m). The phase contents were calculated from the main

33 and characterized peaks of each phase in Figure 2.1 (b). None of the samples could be assigned to a single phase structure; instead they were composites of two or three phases mentioned above. This composite nature led to the observed peak broadening, peak position shift and additional peaks. [37] Layered Li 2 MnO 3 (C2/m) phase can be distinguished in º range in Figure 2.1 (b), and is named Li 2 MnO 3 -like super-lattice peak. These peaks exist in all compositions studied here, but are particularly prominent in the low lithium samples (Li1.20, Li1.00 and Li0.80). The baseline composition Li1.40 consists of Li 2 MnO 3 (C2/m) and LiMO 2 (R3 m). The composite structures of layered and spinel phases are also identified in high-resolution TEM results, as shown in Figure 2.2 (a). Figure 2.2 (b-c) display the fast-fourier transformations (FFT) for regions marked 1 and 2 in (a), which exhibit layered and spinel structures, respectively. The row of spots indicated by the white arrow in Figure 2.2 (c) corresponds to the spinel phase, which is integrated with layered phase at the nanometer scale to form the composite structure. Figure 2.3 shows the phase contents calculated from refined HRPD. In Li1.40 the phase contents of Li 2 MnO 3 (C2/m) and LiMO 2 (R3 m) are 56.4% and 43.6% respectively. As mentioned earlier, the low lithium samples Li1.20, Li1.00 and Li0.80 each contained three different phase structures: Li 2 MnO 3 (C2/m), LiMO 2 (R3 m) and LiM 2 O 4 (Fd3 m). From Li1.20 to Li0.80, the phase content of (Fd3 m) increased[21] while that of (R3 m) decreased. The (C2/m) phase content reached maximum at Li1.20 and decreased monotonously to Li0.80 at lower lithium contents. The larger phase content of Li 2 MnO 3 phase might be exaggerated, though. This is

34 because Li 2 MnO 3 has a higher degree of crystallinity than the other phases due to the high temperature calcination. The high-lithium sample Li1.60 has increased (C2/m) and decreased (R3 m) phase contents, when compared to the baseline composition Li1.40. This is consistent with the tendency found in a previous study. [37] At the initial calcination stage, the lithium would be used preferentially to form the (C2/m) phase, and this phase is based on the formation of LiMn 6 clustering, which is thermodynamically stable during the heat treatment process. [46-48] The (Fd3 m) phase appears after (C2/m). [49] The phase content of (R3 m) phase finally gets smaller. Therefore the phase fraction of (R3 m) phase decreases with the lithium content, from 43.6% in Li1.40 to 13.7% in Li0.80, as shown in Figure 2.3. In Table 2.2, the lattice parameters in the a-axis and c-axis of (C2/m) show only small variations across all lithium contents, as it suggests there is little change in the transition metal composition. On the other hand, the a-axis and c-axis of (R3 m) are significantly decreased at the higher lithium content, which suggest the composition variation of the transition metals. Figure 2.4 (a-c) display the XANES spectra as the lithium content is varied. Generally, XANES peaks are shifted to a lower energy region as the number of valence electrons increases due to the core-hole screening effect, and the peak shapes are related to its interaction with neighboring elements. For the Ni K-edge, the peak shapes for all compositions except Li1.60 are almost identical. Only the peak of Li1.60 shifted to higher energy in Figure 2.4 (a), which means nickel is partially changed from Ni 2+ to Ni 3+ for the charge balance. The peak shapes for all

35 compositions except Li1.60 are almost identical. The Co K-edge spectra display more variation in peak positions and shapes. The shape change is due to rearrangement of the surrounding atoms induced by lithium content change. The peak shift confirms the oxidation state change in Co. When the lithium content was lower than the baseline Li1.40, the spectral peaks shifted to lower energy, which could be interpreted as partial reduction of Co 3+ to Co 2+. On the other hand, the Co spectrum change between Li1.40 and Li1.60 was negligible. Mn K-edge XANES spectra of the samples exhibited negligible changes with Li contents (Figure 2.4 (c)). In Figure 2.4 (d) the Ni K-edge XAFS shows that the spectral distribution in Li1.60 was increased because Ni 3+ induced the Jahn-Teller distortion. The detailed EXAFS data are shown in Table 2.3. Co K-edge EXAFS shows the Co coordination changed from the octahedral to the tetrahedral site as the lithium content was lowered. In lower lithium compositions Co 2+ might be formed in the tetrahedral site. Previously reported (Fd3 m) phase considered only Ni and Mn transition metals. [9, 26]Since Li 2 MnO 3 (C2/m) and LiMO 2 (R3 m) do not have stable site for divalent Co ions, it was believed that the divalent Co ions participate in forming the (Fd3 m) spinel phase. This phase might be the mixed spinel phase [Li +,Co 2+ ][Ni 2+,Co 3+,Mn 4+ ] 2 O 4,[35, 50] with Co 2+ and Co 3+ occupying the tetrahedral and octahedral sites, respectively. The spectra of Mn K-edge in XANES and EXAFS are all very similar, indicating that the oxidation state and local interactions of Mn were not significantly affected by the lithium content. In Figure 2.5, the XRD pattern of Li1.00 composition is

36 compared with those of three spinel structures: (1) LiNi 0.5 Mn 1.5 O 4, (2) LiCo 0.5 Mn 1.5 O 4, and (3) LiMn 2 O 4. The spinel (111) peak position in Li1.00 was located in a position close to LiNi 0.5 Mn 1.5 O 4 and LiCo 0.5 Mn 1.5 O 4 phases. This is also consistent with the spinel phase in Li1.00 being the mixed spinel phase (4) [Li +,Co 2+ ][Ni 2+,Co 3+,Mn 4+ ] 2 O 4. Figure 2.6 shows SEM images of the precursor Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 and all Li x Ni 0.25 Co 0.10 Mn 0.65 O (3.4+x)/2 samples heattreated at 900ºC. In the precursor, flake-shaped primary particles agglomerated to form secondary particles about 5um in size after coprecipitation. After heat treatment, the primary particles grew into spheres with faceted morphology. The smallest primary particles, approximately 200 nm in size, occurred in Li1.40 baseline composition. At higher lithium content (Li1.60), the size increased to 400 nm (Figure 5 (a)), which could be attributed to the Li 2 CO 3 phase which acts as a flux medium. [10] At lower lithium contents, the primary particles were larger and octahedral-shaped (Figure 5 (c)-(e)). The morphology and primary particle size change might be caused by an increase of the spinel (Fd3 m) phase content (Figure 3), as well as segregation of Co which is seen in the EPMA mapping results (Figure 6). Figure 2.7 shows the EPMA composition mapping data. EPMA is generally used to determine the compositional distribution of transition metals. The powders were molded in epoxy and Ar ion-milled to reveal the vertical section of particles. In the baseline composition Li1.40, Ni, Co, and Mn were distributed uniformly within the spherical secondary particles. At lower lithium contents (Li1.20 and Li1.00), separate Ni-rich and Co-rich

37 regions formed inside a single particle. Mn, on the other hand, was uniformly distributed in all 5 calcinated compositions. Different diffusional mobilities and composition differences of Ni, Co and Mn in Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 precursor might cause the segregation during the heat treatment at lower lithium levels. The segregated Co induced by lower lithium promotes particle growth. [51] Figure 2.8 shows the charge-discharge profiles at the first cycle, i.e. activation at 4.7 V. The cells delivered discharge capacities of 188, 272, 216, 146, and 91 mah g 1 with Li1.60, Li1.40, Li1.20, Li1.00, and Li0.80 cathode materials, respectively. Table 2.4 lists the rate capabilities of these compositions. The specific capacity vs. cycle number results are shown in Figure 2.9. The baseline sample Li1.40 shows the best performance, as expected from our previous studies, [20] because it had the smallest particle sized and the optimized amount and composition of (C2/m) and (R3 m) phases.[16] During charging all the compositions display a plateau around 4.5 V, which is known as the activation region of (C2/m) phase found at all lithium levels (Figure 2.8). In Li1.20, the plateau around 2.7 V during discharging is related to the electrochemical reaction of the spinel phase. This plateau remains visible in Li1.00 and Li0.80. The spinel plateau at 4.8 V during charging, shown in Figure 2.10, is also known as characteristic of high-voltage redox reaction of the spinel LiNi 0.5 Mn 1.5 O 4 phase Identification of Composite phases and Their Growth Mechanism Based on the structural, morphological and electrochemical results

38 above, the possible compositions of different phases are calculated in Table 2.5. The assumption is that during heat treatment lithium will be ordered preferentially with Mn in LiMn 6 clustering in the transition metal layers, and form the Li 2 MnO 3 phase. The remaining lithium will react with Ni, Co, and Mn to form the LiMO 2 and LiM x Mn 2-x O 4 phases. Li1.60 contains both (C2/m) and Ni-rich (R3 m) phases, the latter having been observed in XANES spectra with coexisting Ni 2+ and Ni 3+. At lower lithium levels, the spinel phase has been detected in XRD and electrochemical chargingdischarging results. The types of the spinel phase should be specified for each composition. A total of four types should be considered, as mentioned earlier. The calculated compositions are given in Table 1. At lower lithium levels, the most possible composite phase is the mixed spinel phase (4) [Li +,Co 2+ ][Ni 2+,Co 3+,Mn 4+ ] 2 O 4, indicated in bold and red color. Table 1 and Table S5 show the formation energy in the composite structure from ab initio calculations. The trend in these formation energies is consistent with the structural analysis results. For example, in Li1.00 the mixed spinel phase 0.4Li 2 MnO 3 0.2LiNi Co Mn O 2 0.4LiNi Co Mn O 4 has the lowest formation energy and therefore is the most stable structure. To match the phase content of 63.3% Li 2 MnO 3 in Li1.00 obtained from Rietveld refinement (Figure 2.3), the phase content of 50% Li 2 MnO 3 is also calculated. The LiMO 2 (R3 m) phase, however, cannot exist in this case; only Li 2 MnO 3 (C2/m) and LiM x Mn 2-x O 4 (Fd3 m) are present as 0.5Li 2 MnO 3 0.5LiNi 0.75 Co 0.30 Mn 0.95 O 4. Figure 2.11 schematically sketches the phase evolution during

39 heat-treatment at high and low lithium levels. When the lithium is at the baseline (Li1.40) level or higher, the transition metals are distributed uniformly within the secondary particles during heat treatment. At lower lithium levels Co and Ni become segregated. We have already seen that cobalt plays an important role in the elemental segregation and morphology. Cobalt has been reported to reduce the amount and domain size of (C2/m) phase.[25] Li et al. reported that increased cobalt content led to the acceleration of grain growth.[51] Cobalt also favors the layered structure and suppresses the spinel phase according to Deng et al.[21] The cobalt-rich domains, such as those in Figure 2.7, might easily form the (R3 m) phase, suppressing the growth of (C2/m). However, the nickel-rich domains might favorably form (Fd3 m) and (C2/m) phases instead. The voltage decay and capacity loss in the batteries induced by phase transition might be reduced through the possibility of a spinel phase with mixed transition metals. 2.4 Conclusions Lithium-rich oxides Li x Ni 0.25 Co 0.10 Mn 0.65 O (3.4+x)/2 (x = 1.6, 1.4, 1.2, 1.0, 0.8) were synthesized by hydroxide co-precipitation method. The baseline x=1.4 composition showed the best electrochemical performance due to its small primary particle size and homogeneous distribution of nanosized phase domains. XRD, XANES and EXAFS studies indicate that the high-lithium compositions Li1.60 and Li1.40 consisted of two phases (C2/m) and (R3 m) phase which contains Ni 2+ and Ni 3+. At lower lithium levels (x<1.4) the composition contains three different phases: (C2/m), (R3 m) and

40 the spinel phase LiM x Mn 2-x O 4 (Fd3 m). EXAFS analysis shows that low lithium content forces Co to move from octahedral site to tetrahedral site, suggesting that the spinel phase might be composed of mixed transition metals as [Li +,Co 2+ ][Ni 2+,Co 3+,Mn 4+ ] 2 O 4. Low lithium condition is also shown to induce Co, Ni segregation and primary particle growth. As the spinel phase is known to affect the electrochemical performance of the composite, results presented here will help formulating the composition and synthesis protocol of cobalt-containing, lithium-rich oxides for optimal Liion batteries performance.

41 Table 2.1 Elemental analysis for the materials under study

42 Table 2.2 The results of the Rietveld refinement for the materials under study

43 Table 2.3 XAFS analysis at the Ni K edge (d), Co K edge (e), and Mn K edge (f) for Li 1.60, Li 1.40, Li 1.20, Li 1.00, and Li 0.80.

44 Table 2.4 Rate capability tests of (1) Li1.60, (2) Li1.40, (3) Li1.20, (4) Li1.00, (5) Li0.80 at different currents (C rates) in the potential range V in coin type cells. Li counter electrodes.

45 Table 2.5 Composition calculation for 5 compositions with different lithium levels. Labels (1)-(4) refer to the spinel phase types LiNi 0.5 Mn 1.5 O 4, LiCo 0.5 Mn 1.5 O 4, LiMn 2 O 4 and [Li +,Co 2+ ][Ni 2+,Co 3+,Mn 4+ ] 2 O 4 respectively.

46 Figure 2.1 (a) Panels (1)-(5) are XRD patterns of Li1.60, Li1.40, Li1.20, Li1.00 and Li0.80 respectively, after heat-treatment at 900 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

47 Figure 2.2 (a) High-resolution TEM image taken from the Li1.00 composition. (b) and (c) Local structural information from Fast-Fourier Transformations of regions 1 region 2, respectively. The extra row of spots indicated by the arrow in (c) show a spinel structure.

48 Figure 2.3 Rietveld refinement for the phase contents in Li1.60-Li0.80 compositions. The fractions are calculated with the composite notation like Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiM x Mn 2-x O 4 (M=Ni, Co, Mn).

49 Figure 2.4 XANES spectra of Li1.60-Li0.80 compositions at (a) the Ni K- edge, (b) Co K-edge, and (c) Mn K-edge. XAFS spectra at (d) the Ni K-edge, (e) Co K-edge, and (f) Mn K-edge.

50 Figure 2.5 (a) XRD patterns of LiMn 2 O 4 and LiNi 0.5 Mn 1.5 O 4 heat-treated at 900degC. (b) and (c) Selected XRD angular range highlighting the difference between LiMn 2 O 4 and LiNi 0.5 Mn 1.5 O 4.

51 Figure 2.6 FESEM images of (a0) Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 precursor and (a)- (e) Li1.60-Li0.80 compositions heat-treated at 900 C.

52 Figure 2.7 Cross-sectional EPMA mapping and images of (a) Li1.40, (b) Li1.20, and (c) Li1.00. A scale bar is located at the lower right corner.

53 Figure 2.8 a)-(e) Initial charge-discharge profiles measured upon galvanostatic cycle of Li1.60-Li0.80 compositions. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes.

54 Figure 2.9 Plots of specific capacity vs. cycle number of (1) Li1.60, (2) Li1.40, (3) Li1.20, (4) Li1.00, (5) Li0.80 at 1C rate (230mA/g) in the potential range V in coin type cells. Li counter electrodes.

55 Figure 2.10 Initial charge-discharge profiles measured upon galvanostatic cycle of (a) Li1.60, (b) Li1.40, (c) Li0.80 at 23mA/g (C/10) rate in the potential range of V in coin type cells, Li counter electrodes.

56 Figure 2.11 Schematic drawing showing the proposed phase evolution process for high and low lithium compositions.

57 Chapter 3. Spinel phase composite of lithiumrich oxide

58 3.1 Introduction Lithium-rich composite oxide has attracted research focus for a decade due to high discharge capacity, improved safety, and low cost. The high capacity above 250mAh/g come by sacrificing structural stability oxygen evolution, cation migration, and phase transformation.[4, 52, 53] Commercial exploitation of lithium-rich composite oxide has been challenged and delayed by degradation of the capacity and voltage during charge-discharge cycling.[54] A lot of research activities have been tried to improve the capacity retention and voltage decay in lithium rich oxide. Doping, surface treatment, and composite structure are extensively researched.[4, 55, 56] To prevent phase transition during cycling stable structure is needed.[54] Using another stable phase embedding has a possibility to improve performance. Candidates for stable embedding are spinel phase.[9, 32, 57, 58] To implement spinel phase in lithium-rich composite oxide multiple synthesis techniques are tried, but does not satisfactorily explain the phase evolution and electrochemical performance improvement.[9, 57] To improve the electrochemical performance spinel embedding composite structure is designed. Spinel types LiM 0.5 Mn 1.5 O 4 (M = Ni, Co, Mn) are varied and synthesized by controlling spinel composition and content. Spinel LiMn 2 O 4 is widely used for safe battery design.[59] Spinel LiNi 0.5 Mn 1.5 O 4 is researched for high voltage battery system.[60-63] Spinel LiCo 0.5 Mn 1.5 O 4 can operate high voltage but isn t researched widely.[64-67] Lithium-rich oxide with various spinel type is designed and

59 evaluated for electrochemical performances. 3.2 Experimental Procedure Synthesis The Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 precursor was prepared through a hydroxide co-precipitation process.[68] Proper amounts of NiSO 4 6H 2 O, CoSO 4 7H 2 O, and MnSO 4 H 2 O were stirred in deionized water to form a homogeneous solution. The solution was chelated using NH 4 OH and precipitated with NaOH. The co-precipitated Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 after drying was mixed with Li 2 CO 3, MnCO 3, Ni(OH) 2, and Co(OH) 2 to form the composite material in Figure Their specific formulae are as follows: Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiMn 2 O 4 Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiNi 0.5 Mn 1.5 O 4 Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiCo 0.5 Mn 1.5 O 4 The detailed compositions are as follows : (1-x)[0.4Li 2 MnO 3 0.6LiNi Co Mn O 2 ] xlim 0.5 Mn 1.5 O 4, (x = 0.00/0.02/0.05/0.10/0.20, M = Ni, Co, Mn) The mixed powders were then calcined at C for 10 h in flowing air or oxygen Characterization and evaluation The crystal structures of the powder samples were determined by XRD using Cu K. Morphology changes of the powders were determined using SEM (S- 4700N, Hitachi). The inductively coupled plasma technique (ICP-AES) was

60 used to determine the ratios of Li, Ni, Co, and Mn elements in each sample. In order to observe the transition metal distributions in the secondary particles, cross sections of the powder particles were prepared by Ar ion milling on a LN2 cooled stage and measured by an Electron Probe MicroAnalyzer (EPMA, JEOL JXA-8530F). The atomic-level structure and local phases were identified by diffraction and high-resolution TEM (FEI, Titan-cubed ). The electrodes were prepared by making a slurry of 92 wt% active material (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiM 0.5 Mn 1.5 O 4 (M = Ni, Co, Mn)) 4 wt% conductive Denka Black, and 4 wt% polyvinylidene difluoride (PVDF) binder in N-methyl-2-pyrrolidone (NMP) as a solvent. The slurry was coated using doctor-blade method onto Al foil as a current collector. The electrodes were then dried at 120 C in vacuum and pressed. Metallic Li was used as the anode. The electrolyte solution was 1.3 mol L -1 LiPF 6 dissolved in fluoroethylene carbonate and dimethylene carbonate. A porous polyethylene-based membrane was used as a separator. The above components were assembled into CR2032-type coin cells in a dry room. The typical loading of the active mass was 10 mg/cm 2. The cells were charged to 4.7 V for one cycle and then cycled between 2.5 and 4.6 V vs. Li/Li Computational Methods The first principle calculations were performed using the Vienna ab initio simulation package (VASP) [39, 40] with the Projector-Augmented- Wave (PAW) method. [41] The exchange correlation interactions were included with the generalized gradient approximation Perdew-Burke- Ernzerhof (GGA-PBE) functional, [42] and the plane wave cutoff energy

61 was set to 500 ev. The structure relaxations were carried out with a criteria of 10-4 ev for the total energy, and 0.02eV/Å for the forces on each atom. The effective on-site Hubbard Ueff corrections were 6.885, 5.95 and 5.0eV on the 3d electrons for Ni, Co and Mn atoms, respectively. [43] The supercells selected in this work contained between 8 and 28 unit cells, depending on the sample composition and structure Results and Discussion Spinel LiMn 2 O 4 composite The XRD patterns of spinel LiMn 2 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 900ºC are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Ni or Mn) were detected. The composite composition is layered-layered-spinel-spinel (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiMn 2 O 4 LiNi 0.5 Mn 1.5 O 4 ). At 10mol% spinel LiMn 2 O 4 embedding LiMn 2 O 4 (400) peak is discerned, but 20mol% spinel LiMn 2 O 4 embedding newly LiNi 0.5 Mn 1.5 O 4 (400) peak is emerged. Li 2 CO 3 and MnCO 3 are consumed to Li 2 MnO 3 and spinel phase is changed to lower Mn content spinel type - LiNi 0.5 Mn 1.5 O 4 in 20mol% composition. The intensity of Li 2 MnO 3 superlattice peak in 20-25deg is increased in 20mol% compared to designed composition. As spinel phase is increased, Li 2 MnO 3 content has to be decreased in composition design. The phase evolution mechanism is discussed in Session in more detail.

62 Figure shows SEM images of spinel LiMn 2 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 900ºC. After heat treatment, the primary particles grew into spheres with faceted morphology. The smallest primary particles, approximately 200 nm in size, occurred in spinel LiMn 2 O 4 0.0mol% composition. At higher spinel LiMn 2 O 4 content (20mol%), the size increased to 1um (Figure (e)), which could be attributed to the Li 2 CO 3 and MnCO 3 phase which acts as a flux medium to promote the grain growth. At higher spinel LiMn 2 O 4 contents, the primary particles were larger and octahedral-shaped (Figure (b)-(e)). Figure shows the charge-discharge profiles at the first cycle, i.e. activation at 4.7 V. The cells delivered discharge capacities of 278, 261, 257, 214 and 119 mah g 1 with spinel LiMn 2 O 4 = 0, 2, 5, 10, and 20mol% cathode materials, respectively. Table lists the rate capabilities of these compositions. The specific capacity vs. cycle number results are also shown in Table The higher spinel LiMn 2 O 4 content shows the lower capacity and cycle performance. The growth of primary particle size during high temperature condition(900degc) limits lithium ion diffusion during chargedischarge even though spinel embedding improves structure stability. During charging all the compositions display a plateau around 4.5V, which is known as the activation region of Li 2 MnO 3 (C2/m) phase found at all spinel LiMn 2 O 4 levels (Figure 3.1.4). In spinel LiMn 2 O 4 10mol%, the plateau around 2.7 V during discharging is related to the electrochemical reaction of the spinel phase. This plateau remains visible in spinel LiMn 2 O 4 10 and 20mol%. The XRD patterns of spinel LiMn 2 O 4 embedded samples (x = 0, 2,

63 5, 10, 20mol%) heat-treated at 750ºC are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Ni or Mn) were detected. The composite composition is layered-layered-spinel-spinel (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiMn 2 O 4 LiNi 0.5 Mn 1.5 O 4 ) the same as heattreated at 900ºC. The intensity of Li 2 MnO 3 superlattice peak in 22deg is increased as spinel contents are increased in designed compositions. Figure shows SEM images of spinel LiMn 2 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 750ºC At higher spinel LiMn 2 O 4 content (20mol%), the size is the same as lower spinel content, which could be attributed to low temperature for grain growth. Table lists the rate capabilities of these compositions. The specific capacity vs. cycle number results are also shown in Table The higher spinel LiMn 2 O 4 content (20mol%) shows the lower capacity (190mAh/g at 0.1C) and cycle performance (74.1% at 50 th ). At spinel LiMn 2 O 4 content (2mol%) the cycle performances are improved. Capacity retention after 50 th cycle is 90.6%, and voltage decay -55mV. The spinel embedded structure improved the structure stability during charge-discharge at small primary particle morphology.

64 Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, (4) 10, and (5) 20mol% heat-treated at 900degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes.

65 Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, (4) 10, and (5) 20mol% heat-treated at 750degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes.

66 Figure Synthesis procedure for spinel composite oxide

67 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

68 Figure FESEM images of spinel composites (a) 0, (b) 2, (c) 5, (d) 10, and (e) 20mol% heat-treated at 900 C.

69 Figure (a)-(d) Initial charge-discharge profiles measured upon galvanostatic cycle of spinel 0-20mol% compositions heat-treated at 900degC. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes.

70 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

71 Figure FESEM images of spinel composites (a) 0, (b) 2, (c) 5, (d) 10, and (e) 20mol% heat-treated at 750 C..

72 3.3.2 Spinel LiNi 0.5 Mn 1.5 O 4 composite The XRD patterns of spinel LiNi 0.5 Mn 1.5 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 900ºC in air are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Ni) are detected. The rocksalt NiO is also detected. The composite composition is layeredlayered-spinel-rocksalt (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiNi 0.5 Mn 1.5 O 4 NiO). The XRD patterns of spinel LiNi 0.5 Mn 1.5 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 750ºC in air are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Ni) are detected. The rocksalt NiO is detected. The composite composition is layered-layeredspinel-rocksalt (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiNi 0.5 Mn 1.5 O 4 NiO). The XRD patterns of spinel LiNi 0.5 Mn 1.5 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 900ºC in oxygen are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Ni) are detected. The rocksalt NiO is detected. The rocksalt NiO phase heat-treated at 900degC in oxygen is decreased compared with air atmosphere. The composite composition is layered-layered-spinel-rocksalt (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiNi 0.5 Mn 1.5 O 4 NiO). The XRD patterns of spinel LiNi 0.5 Mn 1.5 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 750ºC in oxygen are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Ni) are

73 detected. The rocksalt NiO is detected. The rocksalt NiO phase heat-treated at 750degC in oxygen is the same as air atmosphere. The composite composition is layered-layered-spinel-rocksalt (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiNi 0.5 Mn 1.5 O 4 NiO). Figure shows the charge-discharge profiles at the first cycle, i.e. activation at 4.7 V. The cells delivered discharge capacities of 278, 266, 262, 232 and 169 mah g 1 with spinel LiNi 0.5 Mn 1.5 O 4 0, 2, 5, 10, and 20mol% cathode materials, respectively. Table lists the rate capabilities of these compositions. The specific capacity vs. cycle number results are also shown in Table The higher spinel LiNi 0.5 Mn 1.5 O 4 content shows the lower capacity and cycle performance. The growth of primary particle size during high temperature condition(900degc) limits lithium ion diffusion during charge-discharge even though spinel embedding improves structure stability. During charging all the compositions display a plateau around 4.5 V, which is known as the activation region of (C2/m) phase found at all spinel LiNi 0.5 Mn 1.5 O 4 levels (Figure 3.2.5). In spinel LiNi 0.5 Mn 1.5 O 4 20mol%, the plateau around 2.7 V during discharging is related to the electrochemical reaction of the spinel phase. This plateau is only visible in spinel LiNi 0.5 Mn 1.5 O 4 20mol%. The specific capacity vs. cycle number results are also shown in Table The higher spinel LiNi 0.5 Mn 1.5 O 4 content (20mol%) shows the lower capacity (169mAh/g at 0.1C) and cycle performance (87.5% at 50th). At spinel LiNi 0.5 Mn 1.5 O 4 content (5mol%) the voltage decay is improved. Capacity retention after 50th cycle is 92.7%, and voltage decay -55mV. The spinel embedded structure improved the structure stability during charge-

74 discharge.

75 Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, (4) 10, and (5) 20mol% heat-treated at 900degC in air at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes.

76 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C in air. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

77 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C in air. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

78 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C in oxygen. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

79 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C in oxygen. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

80 Figure (a)-(d) Initial charge-discharge profiles measured upon galvanostatic cycle of spinel 0-20mol% compositions heat-treated at 900degC in air. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes.

81 3.3.3 Spinel LiCo 0.5 Mn 1.5 O 4 composite The XRD patterns of spinel LiCo 0.5 Mn 1.5 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 900ºC are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Co) were detected. Layered Li 2 MnO 3 (C2/m) phase can be distinguished in º range in Figure 3.3.1, and is named Li 2 MnO 3 -like super-lattice peak. These peaks exist in all compositions studied here. The composite composition is layeredlayered-spinel (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiCo 0.5 Mn 1.5 O 4 ). The Rietveld refinement with XRD data was performed in order to clarify the phase contents and crystalline structures in the powder samples. Table summarizes the lattice parameters and crystal sizes of Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m), and LiM x Mn 2-x O 4 (Fd3 m) (M = Co). The crystal sizes are increased as spinel LiCo 0.5 Mn 1.5 O 4 content is increased. The phase contents were calculated from the main and characterized peaks of each phase in Figure None of the samples could be assigned to a single phase structure; instead they were composites of two or three phases. As spinel LiCo 0.5 Mn 1.5 O 4 content is increased, the phase content of Li(Ni,Co,Mn)O 2 (R3 m) is decreased and Li 2 MnO 3 (C2/m) is slightly increased. The larger phase content of Li 2 MnO 3 phase might be exaggerated, though. This is because Li 2 MnO 3 has a higher degree of crystallinity than the other phases due to the high temperature calcination.[69] Figure shows SEM images of spinel LiCo 0.5 Mn 1.5 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 900ºC. After heat treatment, the primary particles grew into spheres with faceted

82 morphology. The smallest primary particles, approximately 200 nm in size, occurred in spinel LiCo 0.5 Mn 1.5 O 4 0.0mol% composition. At higher spinel LiCo 0.5 Mn 1.5 O 4 content (20mol%), the size increased to 1um (Figure (c)), which could be attributed to the Co(OH) 2 and MnCO 3 phase which act as a flux medium. At higher spinel LiCo 0.5 Mn 1.5 O 4 contents, the primary particles were larger and octahedral-shaped (Figure (c)). Figure shows the EPMA composition mapping data. EPMA is generally used to determine the compositional distribution of transition metals. The powders were molded in epoxy and Ar ion-milled to reveal the vertical section of particles. At high spinel LiCo 0.5 Mn 1.5 O 4 content (20mol%), separate Ni-rich and Co-rich regions formed inside a single particle. Mn, on the other hand, was uniformly distributed. Different diffusional mobilities and composition differences of Ni, Co and Mn in Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 precursor might cause the segregation during the heat treatment at high spinel LiCo 0.5 Mn 1.5 O 4 level.[34, 70-72] The segregated Co induced by spinel LiCo 0.5 Mn 1.5 O 4 (raw material - Co(OH) 2 and MnCO 3 ) promotes particle growth. Co and Ni segregations were also visible from the HAADF image and EDS mapping of Co and Ni, as shown in Figure The composite structures of layered and spinel phases are also identified in high-resolution TEM results, as shown in Figure (d). The row of spots indicated by the white arrow in Figure (d) corresponds to the spinel phase, which is integrated with layered phase at the nanometer scale to form the composite structure. Figure shows Mn XPS spectra for spinel LiNi 0.5 Mn 1.5 O 4 and

83 LiCo 0.5 Mn 1.5 O 4 (20mol%) embedded samples. The major peak centered at 642.2eV corresponds to Mn 4+ and the minor one at 641.0eV corresponds to Mn 3+ ions in the composite oxide. XPS result shows that spinel LiCo 0.5 Mn 1.5 O 4 embedded oxide increases Mn 3+ compared to spinel LiCo 0.5 Mn 1.5 O 4 embedding. Table lists the rate capabilities of these compositions. The specific capacity vs. cycle number results are also shown in Table The higher spinel LiCo 0.5 Mn 1.5 O 4 content shows the lower capacity and cycle performance. The growth of primary particle size during high temperature condition(900degc) limits lithium ion diffusion during charge-discharge even though spinel embedding improves structure stability. The initial coulombic efficiencies are improved by embedding spinel LiCo 0.5 Mn 1.5 O 4. The XRD patterns of spinel LiCo 0.5 Mn 1.5 O 4 embedded samples (x = 0, 2, 5, 10, 20mol%) heat-treated at 750ºC are shown in Figure In Figure (a), peaks of the Li 2 MnO 3 (C2/m), Li(Ni,Co,Mn)O 2 (R3 m) phases and the spinel phase LiM x Mn 2-x O 4 (Fd3 m) (M = Co) were detected. The composite composition is layered-layered-spinel (Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiCo 0.5 Mn 1.5 O 4 ). Figure shows SEM images of spinel LiCo 0.5 Mn 1.5 O 4 embedded samples (x = 0, 5, 20mol%) heat-treated at 750ºC. At higher spinel LiCo 0.5 Mn 1.5 O 4 content (20mol%), the size is the same as lower spinel content, which could be attributed to low temperature for grain growth. The smallest primary particles, approximately 200 nm in size, occurred in spinel LiCo 0.5 Mn 1.5 O 4 0.0mol% composition. Figure shows the EPMA composition mapping data heat-

84 treated at 750degC. At high spinel LiCo 0.5 Mn 1.5 O 4 content (20mol%), separate Ni-rich and Co-rich regions formed inside a single particle. Table lists the rate capabilities of these compositions. The specific capacity vs. cycle number results are also shown in Table The higher spinel LiCo 0.5 Mn 1.5 O 4 content (10mol%) shows the capacity (251mAh/g at 0.1C), cycle performance (90.9% at 50th) and voltage decay (- 64mV). By embedding spinel LiCo 0.5 Mn 1.5 O 4 the cycle performance and voltage decay are improved. The spinel embedded structure improved the structure stability during charge-discharge at small primary particle morphology. Figure shows the charge-discharge profiles at the first cycle, i.e. activation at 4.7 V. The cells delivered discharge capacities of 270, 270, 262, and 251 mah g 1 with spinel LiCo 0.5 Mn 1.5 O 4 0, 2, 5, and 10 mol% cathode materials, respectively. In spinel LiCo 0.5 Mn 1.5 O 4 10mol%, the plateau around 2.7 V during discharging is related to the electrochemical reaction of the spinel phase. This plateau is only visible in spinel LiCo 0.5 Mn 1.5 O 4 10mol%. Full cell cycle performance is evaluated using cylindrical cell with graphite anode in Figure Spinel LiCo 0.5 Mn 1.5 O 4 embedded composite shows better discharge capacity retention and voltage decay. The phase stability using spinel LiCo 0.5 Mn 1.5 O 4 embedded structure improves the cycling performance.

85 Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, and (4) 10mol% heat-treated at 900degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes.

86 Table The results of the Rietveld refinement for the materials under study

87 Table Electrochemical performance tests of spinel (1) 0, (2) 2, (3) 5, and (4) 10mol% heat-treated at 750degC at different currents (C rates) in the potential range /4.7V in coin type cells. Li counter electrodes.

88 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 900 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

89 Figure Rietveld refinement for the phase contents in (1- x)[0.4li 2 MnO 3 0.6Li(Ni,Co,Mn)O 2 ] xlico 0.5 Mn 1.5 O 4 compositions. The fractions are calculated with the composite notation like Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiM x Mn 2-x O 4 (M=Ni, Co, Mn).

90 Figure FESEM images of spinel composites (a) 0, (b) 5, and (c) 20mol% heat-treated at 900 C.

91 Figure Cross-sectional EPMA mapping and images of spinel 20mol% composite heat-treated at 900degC. A scale bar is located at the right corner.

92 Figure (a)-(c) HAADF image and EDS mapping taken from spinel 20mol% heat-treated at 900degC. (d) High-resolution TEM image taken from spinel 20mol% heat-treated at 900degC.

93 Figure XPS data taken from spinel 20mol% heat-treated at 900degC.

94 Figure (a) Panels (1)-(5) are XRD patterns of spinel 0, 2, 5, 10, and 20mol% respectively, after heat-treatment at 750 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

95 Figure FESEM images of spinel composites (a) 0, (b) 5, and (c) 20mol% heat-treated at 750 C.

96 Figure Cross-sectional EPMA mapping and images of spinel 20mol% composite heat-treated at 750degC. A scale bar is located at the right corner.

97 Figure (a)-(c) Initial charge-discharge profiles measured upon galvanostatic cycle of spinel 0-10mol% compositions heat-treated at 750degC. The data are recorded at 23 ma/g (C/10) rate in the potential range of V in coin-type cells with Li counter electrodes.

98 Figure Full cell cycle life (a) discharge capacity retention (b) voltage decay in type full cells with graphite anode.

99 3.4. Conclusions Spinel embedded lithium rich oxides are synthesized and structural phases are analyzed. By adding Mn, Ni, Co sources in precursor Ni 0.25 Co 0.10 Mn 0.65 (OH) 2 the spinel embedded composite oxides are prepared. In spinel LiMn 2 O 4 embedded composition Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiMn 2 O 4 LiNi 0.5 Mn 1.5 O 4 phase is formed. At 750degC heat-treatment spinel LiMn 2 O 4 2mol% the capacity retention and voltage decay are improved. In spinel LiNi 0.5 Mn 1.5 O 4 embedded composition Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiNi 0.5 Mn 1.5 O 4 NiO phase is formed. Even using oxygen atmosphere heat-treatment the rocksalt NiO phase is maintained. At 750degC heat-treatment spinel LiNi 0.5 Mn 1.5 O 4 2mol% the capacity retention and voltage decay are improved. In spinel LiCo 0.5 Mn 1.5 O 4 embedded composition ternary Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiCo 0.5 Mn 1.5 O 4 phase is formed. Using EPMA Ni and Co segregation is analyzed. At 750degC heat-treatment spinel LiCo 0.5 Mn 1.5 O 4 2mol% the capacity retention and voltage decay are improved. The phase evolutions are summarized in Table As MnCO 3 is added to evolve LiMn 2 O 4 embedded composite LiNi 0.5 Mn 1.5 O 4 phase is also formed. Some of the added MnCO 3 is consumed to make Li 2 MnO 3 and the spinel phase is formed into LiNi 0.5 Mn 1.5 O 4 in high LiMn 2 O 4 embedded composition. As Ni(OH) 2 is used for LiNi 0.5 Mn 1.5 O 4 embedded composite rocksalt NiO phase is formed. Even in oxygen atmosphere rocksalt NiO phase is inevitable phase. Using Co(OH) 2 precursor the spinel

100 LiCo 0.5 Mn 1.5 O 4 embedded composite is homogeneously formed. The phases which are evolved in different spinel composition design are compared in XRD data in Figure Each spinel phase can be distinguished in spinel peak position spinel (311) and spine (400) peaks. The formation energies of spinel embedded composite are calculated using ab initio method in Figure In the baseline Li 2 MnO 3 Li(Ni,Co,Mn)O 2 composite the spinel LiCo 0.5 Mn 1.5 O 4 embedding decreases formation energy in (1) 0.8[0.4Li 2 MnO 3 0.6LiNi Co Mn O 2 ] 0.2LiCo 0.5 Mn 1.5 O 4. The phase stability of spinel LiCo 0.5 Mn 1.5 O 4 lowers the formation energy. Mixed spinel LiCo 0.25 Ni 0.25 Mn 1.5 O 4 phase shows similar formation energy as LiCo 0.5 Mn 1.5 O 4. As Li 2 MnO 3 is increased from 40% to 45%, the formation energy decreases to eV/f.u.. Li 2 MnO 3 and mixed spinel LiCo 0.25 Ni 0.25 Mn 1.5 O 4 embedding decreases the formation energy more. The phase evolution is analyzed using high temperature XRD in the spinel LiCo 0.5 Mn 1.5 O4 embedding in in Figure Layered Li(Ni,Co,Mn)O 2 phase is emerged at 400degC. Layered Li 2 MnO 3 -like superlattice peak is detected at 550degC. Spinel LiCo 0.5 Mn 1.5 O 4 peak is emerged at 650degC. The spinel LiCo 0.5 Mn 1.5 O 4 is stabilized after layered Li(Ni,Co,Mn)O 2 and Li 2 MnO 3 are formed. Lithium is going to be accommodated by the formation of Li 2 MnO 3 -like domains, and then layered Li(Ni,Co,Mn)O 2 and spinel LiCo 0.5 Mn 1.5 O 4 domains.[46] Figure schematically sketches the phase evolution during heat-treatment of spinel embedded composite. When the raw materials for spinel phase (MnCO 3, Ni(OH) 2, and Co(OH) 2 ) are added, cobalt and nickel

101 become segregated. In the segregated region the ternary Li 2 MnO 3 Li(Ni,Co,Mn)O 2 LiM 0.5 Mn 1.5 O 4 (M=Ni, Co, Mn) phase is formed. In lithium-rich composite oxide by adding spinel forming sources the different types of spinel phase embedded composites are synthesized. Controlling the spinel composition and content the electrochemical performance is improved.

102 Table Summary of spinel composites.

103 Figure (a) Panels (1)-(4) are XRD patterns of spinel 0mol%, 20mol% LiMn 2 O 4, 20mol% LiNi 0.5 Mn 1.5 O 4 and 20mol% LiCo 0.5 Mn 1.5 O 4 respectively, after heat-treatment at 900 C. Selected angular ranges highlight the (b) layered and (c) spinel peaks.

104 Figure Formation energy of spinel embedded composite.