Cathode for Thin-Film Lithium-Ion Batteries

Abstract

Thin-film lithium-ion batteries (LIBs) have attracted considerable attention for energy storage device application owing to their high specific energy compared to conventional LIBs. However, the significant breakthroughs of electrochemical performance for electrode materials, electrolyte, and electrode/electrolyte interface are still highly desirable. This chapter firstly gives an overview of cathode materials including lithium-containing cathode (e.g., LiCoO2, LiMn2O4, LiFePO4, LiNi1−x−yMnxCoyO2, LiNi0.5Mn1.5O4) and lithium-free cathode (e.g., vanadium oxides) for LIBs in terms of specific capacity, energy density, working voltage, cycling life, and safety. In the meanwhile, the existing drawbacks and limitations of various battery chemistries are also analyzed. Furthermore, some modification strategies for these cathode materials have also been discussed for improving electrochemical performance. Finally, the thin-film Li-ion battery applications of these cathode materials are summed up toward next-generation flexible and high-energy devices.

1. Overview of cathode materials

Since lithium-ion batteries (LIBs) gained the commercial success in 1991 [1], they have received considerable attention in various areas including consumer electronics, power tools, electric vehicles, and grid energy storage owing to their high energy density, low cost, long cycle life, and environmentally benignity. The energy density of LIBs in 1991 is only 80 Wh kg⁻¹ at a cell level and has been currently increased to 200–250 Wh kg⁻¹ (at a cell level). The main reason is attributed to the significant improvement of battery materials and battery technology. Especially for cathode materials, they dominate the whole cell performance compared to graphite anode.

Up to now, various cathode materials have been proposed and developed such as layered structure LiCoO₂, LiMnO₂, LiNiO₂, ternary LiNi_1−x−yMn_xCo_yO₂, spinel LiMn₂O₄ (LMO), spinel LiNi_0.5Mn_1.5O₄ (LNMO), and polyanion-based cathode including LiFePO₄ and Li₃V₂(PO₄)₃ [2], as shown in Figure 1. Typically, transition metals of such cathode materials undergo an oxidation process to higher oxidation state when Li⁺ ions are removed. While oxidation of the transition metals can retain charge neutrality in the compound, Li⁺ removal/insertion usually causes phase changes and structural strain. In this regard, cathode materials that are stable enough over a wide composition range must be employed. Thus, the structural stability of cathode materials is a key during Li⁺ extraction/insertion processes. As for discharge process, Li⁺ ions are reinserted into the cathode materials and electrons from anode reduce transition metal ions in the cathode to a lower valence. The rates of the two processes, as well as access of Li⁺ ions in the electrolyte to electrode surface, dominate the maximum discharge current. Therefore, cathode performance depends critically on the microstructure and morphology of electrode as well as the intrinsic electrochemical properties of the cathode materials. For instance, much effort has been currently focused on the development of nanosized electrode materials for shortening Li⁺/electron transport length and increasing contact areas of electrode/electrolyte interface. This chapter will mainly discuss cathode materials for LIBs and their potential applications for thin-film Li-ion microbatteries.

Figure 1.

Gravimetric and volumetric energy densities of various cathode materials at a material level. LFP, LiFePO₄; LMO, LiMn₂O₄; NMC-111, LiNi_1/3Mn_1/3Co_1/3O₂; NCA, LiNi_0.8Co_0.15Al_0.05O₂; NMC-811, LiNi_0.8Mn_0.1Co_0.1O₂; HV-spinel, LiNi_0.5Mn_1.5O₄.

2. Layered transition metal oxide cathode

Layered transition metal oxides are the most commonly used cathode materials for LIBs such as LiCoO₂, LiNiO₂, LiMnO₂, and LiTiS₂. Such cathode materials have typical layered structure, as illustrated in Figure 2. LiCoO₂ is relatively one of the most mature cathode materials, which will be discussed in the following section. However, the crystal structure of Li_xCoO₂ becomes unstable when Li⁺ extracts beyond x > 0.5, leading to oxygen release [3]. Considering the shortcomings and costly Co resource of LiCoO₂, layered ternary transition metal oxides LiNi_1−x−yMn_xCo_yO₂ (NMC) have been widely investigated and developed in recent years.

Figure 2.

Layered structure of LiMO₂ (M = Ni, Co, Mn), showing the lithium ions between the transition metal oxides.

2.1 LiCoO₂

As one of the typical intercalation compound families, LiCoO₂ was first reported and studied by John B. Goodenough in 1980 for LIBs [4]. It has an a-NaFeO₂ structure containing only one kind of cation in the transition meta layer. LiCoO₂ is widely used as cathode materials for LIBs, especially for portable electronics. Although LiCoO₂ is the most widely used cathode materials for LIBs, the practical capacity available is only around 140 mAh g⁻¹, half of the theoretical capacity (ca. 280 mAh g⁻¹). It corresponds to the 50% Li removal from LiCoO₂ because further lithium extraction from LiCoO₂ usually induces the structural transition from hexagonal to monoclinic phase and Co dissolution at the highly oxidized state, thus leading to rapid capacity decay. To enhance the capacity retention of LiCoO₂ at the higher cutoff potential (4.4–4.5 V), surface modifications (e.g., surface coating) on the LiCoO₂ surface have been employed since such the strategies can effectively restrict the lattice constant change and the transition of hexagonal to monoclinic phase change upon cycling. The mechanism of such capacity improvement for LiCoO₂ was not widely identified. Dahn group found that the poor capacity retention of LiCoO₂ at the high delithiation potential (4.5 V) was attributed to the growth of impedance at the LiCoO₂ surface caused by side reactions [5]. As for LiCoO₂, although a higher reversible capacity by using a higher cutoff potential can be obtained, such a cathode still faces a thermal stability concern, resulting in a safety issue. Moreover, Co in the delithiated Li_1−xCoO₂ possesses high oxidizing power and easily induces the more side reactions with organic electrolytes. Thus, LiCoO₂ is mainly used as low-power devices such as consumer electronics.

2.2 LiNi_1−x−yMn_xCo_yO₂

Although LiCoO₂ exhibits the most early commercial success, the intrinsic structural stability upon delithiation and high raw material cost of LiCoO₂ have pushed the development of a new class cathode materials. As an alternative cathode, layered structured Li_xNiO₂ has been developed owing to its relatively low cost and high working potential [6]. However, pristine Li_xNiO₂ is not a promising cathode candidate because of its some drawbacks. For example, more than 200 mAh g⁻¹ can be obtained from stoichiometric LiNiO₂ at a relatively high voltage (4.5 V vs. Li/Li⁺). Moreover, a 20% capacity loss occurs in the first cycle owing to the structural modification when charged below x = 0.5. Besides electrochemical properties, stoichiometric LiNiO₂ with a Li/Ni ratio of 1:1 is still difficult to be synthesized because of the cation mixing of Ni²⁺ and Ni³⁺ owing to the similarity of ionic radii of Li⁺ (0.76 Å) and Ni²⁺ (0.69 Å) [7]. Moreover, Li_xNiO₂ is usually Li-deficient because it is difficult to prepare stoichiometric LiNiO₂. To address this issue, the synthesis condition usually needs oxygen instead of air atmosphere and a slight excess of LiNO₃.

Considering the advantages and shortcomings of LiNiO₂, metal-doped LiNiO₂, especially ternary layered cathode NMC and LiNi_0.8Co_0.15Al_0.05O₂ (NCA), has recently attracted considerable attention because it shows good structural stability, higher specific capacity up to 220 mAh g⁻¹, higher energy density, and relatively simple synthesis [8]. By introducing Mn and Co or Al and Co into LiNiO₂, the cation mixing of Ni²⁺/Ni³⁺ can be effectively restricted. In the meanwhile, LiNi_1−x−yMn_xCo_yO₂ can be synthesized in the air without the necessity to use pure oxygen atmosphere. As for NMC cathode, LiNi_1/3Mn_1/3Co_1/3O₂ (NMC-111) is currently the most common form of NMC cathode materials and is widely employed in LIBs [9]. To further enhance energy density and reduce cost, increasing Ni content and lowering Co content in NMC cathode are effective solutions such as LiNi_0.5Mn_0.3Co_0.2O₂ (NMC-532), LiNi_0.6Mn_0.2Co_0.2O₂ (NMC-622), and LiNi_0.8Mn_0.1Co_0.1O₂ (NMC-811), as shown in Figure 3.

Figure 3.

(a) The state-of-the-art NMC cathode materials, near-future NMC materials, and far-future cathode materials. (b) Schematic diagram of the positions of the various redox couples relative to the top of the oxygen: 2p band.

It is well known that Ni is beneficial for achieving high energy density but poor stability, especially at the charged state. Mn will induce the formation of a spinel structure to realize a low internal resistance but results in a low energy density. Compared to Ni and Mn, Co has the limited resources (Co reserve available is around 7.1 million tons) [10], high price, and high toxicity. Thus, NMC battery chemistry will gradually lessen the use of Co element and simultaneously employ high Ni and low Mn for realizing the best electrochemical performance by combining the respective merits of Ni, Mn, and Co. NMC-111, NMC-442, and NMC-532 are currently the state of the art of cathode materials. In the very recently, Ni-rich NMC cathodes including NMC-622 and NMC-811 have been widely investigated and will be adopted in the LIBs owing to their higher energy density and lower cost. This is very favorable for automotive applications. On the other hand, Mn³⁺ has the lowest octahedral site stabilization energy (OSSE) and tends to migrate, resulting in a layered-to-spinel transition upon lithiation/delithiation process. In contrast, Co³⁺ has the highest OSSE and exhibits excellent structural stability, but it suffers from poor chemical stability during lithium extraction over 50%, which is attributed to an overlap of the t_2g band of low-spin Co³⁺/⁴⁺ with the top of the 2p band of O²⁻ [11], as shown in Figure 3b. Compared to Co, Mn provides better chemical stability because Mn³⁺/⁴⁺ e_g band is further above the top of the 2p band of the O²⁻. Compared to Co and Mn, Ni shows a moderate structural and chemical stability considering that it has a higher OSSE than Mn³⁺, and the e_g band of low-spin Ni³⁺/⁴⁺ just touches the top of the O²⁻ 2p band. Thus, developing Co-less and Ni-rich NMC cathode will be very desirable.

The commercial success of NMC-111 facilitates the development of other NMC materials by combining the individual advantages of Ni, Mn, and Co in the layered crystal structure with varying concentration to realize high capacity and high energy density. As discussed previously, a high Ni content is favorable to achieve a high reversible capacity. On the other hand, the introduction of Mn can enhance the structural and thermal stability at the deeply delithiated state, while Co dopant in NMC can improve the layered ordering as well as rate performance and specific capacity owing to its redox activity. Sun et al. have systematically investigated the effect of chemical compositions for the ternary NMC materials on reversible capacity, cyclability, and thermal stability [12], as displayed in Figure 4. The NMC materials with higher Ni content show higher reversible capacities, but it shows a decreased capacity retention at both 25 and 55°C. In addition, the exothermic reaction for Ni-rich NMC usually occurs at a lower temperature with a larger heat generation upon charging to 4.3 V vs. Li/Li⁺, which is a good agreement with the phenomena of capacity retention decrease with increasing Ni content (Figure 4). Such findings confirm that Ni-rich NMC possesses a higher specific capacity but poor thermal stability and capacity retention upon cycling, as shown in Figure 4. To address this problem, a novel material design by tuning chemical composition in the core and shell has been proposed by Sun group [13]. The first generation is a core-shelled structure based on Ni-rich NMC-811 as core and Mn-rich LiNi_0.5Mn_0.5O₂ as shell (Figure 5a). The core can offer high capacity, while the shell guarantees the surface stability. Such structural design, however, also has some drawbacks. For example, boundary cracks between active particles often formed owing to different volume changes between the core and the shell structures upon cycling tend to induce the formation of boundary cracks, leading to major mechanical fracture of active particles and capacity decay. Thus, another core-shelled gradient-type and full concentration gradient-type NMC materials have been proposed and developed [14], as illustrated in Figure 5b. This unique design minimizes the propagation of boundary cracks and is beneficial to enhancing cycling stability.

Figure 4.

Discharge capacity as a function of cycle number for various LiNi_xMn_yCo_zO₂ (x = 1/3, 0.5, 0.6, 0.7, 0.8, and 0.85) cathodes at (a) 25°C and (b) 55°C. The current density is 100 mA g⁻¹ (0.5 C) in the voltage range of 3.0–4.3 V [12]. (c) The map of relationship between discharge capacity and thermal stability and capacity retention for LiNi_xMn_yCo_zO₂ (x = 1/3, 0.5, 0.6, 0.7, 0.8, and 0.85) cathode. (d) DSC curves of the above NMC cathode materials.

Figure 5.

(a) Core-shelled structure NMC: core: NMC-811 and shell: LiNi_0.5Mn_0.5O₂; (b) concentration gradient-based NMC cathode.

2.3 Layered NCA cathode

Similar to NMC-811 cathode, LiNi_0.8Co_0.15Al_0.05 (NCA) is also a Ni-rich cathode, showing a relatively high energy density and low cost. The introduction of element Al in NCA can minimize the detrimental phase transition and improve the thermal behavior of the cathode. Al is a favorable dopant in a specific capacity because of its lower atomic weight than the other transition metal elements. Thus, Al-doped LiNiO₂ shows an increased specific capacity. Moreover, the presence of Al³⁺ in the transition metal layer tends to reduce the a-axis but increase the c-axis parameters because of α-LiAlO₂ (a = 2.8 Å and c = 14.23 Å), which decrease the cationic disorder in the Li layer. It is similar to the effect of Co³⁺. However, a large amount of Al doping would lead to the reduction of specific capacity because Al is electrochemically inactive. On the other hand, Al³⁺ can effectively enhance the working voltage owing to the weakening of the Ni-O bond by the stronger Al-O bond through the inductive effect. Thus, a reasonable balance between reduced capacity and increased working voltage should be considered for designing high-performance Al-doped LiNiO₂. It is acknowledged that Al doping (5 atom%) and Co doping (15 atom%) is the optimized results [11]. In recent years, double metal doping has been widely accepted to improve the structural stability and electrochemical performance of layer transition metal oxides.

As discussed previously, the surface stability of Ni-rich cathode is a key issue for their electrochemical performance. The residual lithium oxide at the particle surface after annealing tends to absorb moisture and CO₂ in air during storage, leading to the formation of some unwanted products on the surface such as LiOH and Li₂CO₃. These products would react with electrolyte, forming surface interface. In addition, Ni⁴⁺ at the delithiated state would accelerate the decomposition of electrolyte, further resulting in the thickening of interfaces and sluggish Li⁺ diffusion kinetics. Considering these problems, surface protection of Ni-rich cathode materials is highly desirable such as various surface coatings including metal oxide, phosphates, carbon, fluorides, artificial SEI, and conducting polymers [11].

3. Spinel cathode

3.1 Spinel LiMn₂O₄

In contrast to layered LiCoO₂, spinel LiMn₂O₄ possesses more stable structure and higher voltage (4.0 V). As shown in Figure 6, LiMn₂O₄ has a cubic crystal structure with Li occupying the tetrahedral 8a sites and Mn occupying the octahedral 16d sites. The interstitial sites for Li, 16c, share faces with the occupied sites, forming a continuous three-dimensional diffusion network by the [Mn₂O₄] framework. Thackeray et al. have demonstrated that lithium can be reversibly intercalated and deintercalated for the range of 0 < x < 2.2 in Li_xMn₂O₄ in the cutoff range (2–4.5 V) with two plateaus at 4.0 and 3.0 V [15]. However, LiMn₂O₄ is usually cycled in the voltage range of 3.5–4.5 V because of the poor structural stability during cubic-tetragonal phase transition for 1 < x < 2 in the cutoff window (2–4.5 V). Such cubic-tetragonal phase transition is driven by the Jahn-Teller distortion of MnO₆ octahedron as Mn⁴⁺ is reduced to Mn³⁺, leading to a 6.5% increase of the unit cell volume. Thus, the specific capacity available for LiMn₂O₄ is theoretical 148 mAh g⁻¹. More importantly, LiMn₂O₄ just needs to use earth-abundant and low-cost Mn resources without using Ni and Co element, which is very beneficial for its large-scale application.

Figure 6.

Crystal structures of (a) spinel LiMn₂O₄ and (b) spinel LiNi_0.5Mn_1.5O₄.

Although spinel LiMn₂O₄ shows attractable merits for LIBs, it also suffers from some drawbacks, such as Jahn-Teller distortion and Mn dissolution. To inhibit the Jahn-Teller distortion, doping with Li or multivalent cations (e.g., Al, Mg) has been widely used because partial replacement of Mn can reduce the Mn³⁺ concentration and increase the average valence of Mn, thus suppressing the onset of Jahn-Teller effect. Although metal-doped LiMn₂O₄ shows the enhanced electrochemical properties, it inevitably reduces the theoretical capacity owing to inactive element introduction. In this regard, a small amount of metal substitution is desirable. Another problem for spinel LiMn₂O₄ is the Mn dissolution owing to the disproportionation reaction of Mn³⁺ by forming Mn⁴⁺ and soluble Mn²⁺ catalyzed by acidic species in the electrolyte. Moreover, Mn²⁺ can react with LiPF₆ salt to produce more acidic by-products. To mitigate or address such problems, the surface modification strategy has been extensively employed.

3.2 High-voltage spinel LiNi_0.5 Mn_1.5O₄

Besides 4 V-spinel, another Mn-based spinel LiNi_0.5Mn_1.5O₄ (LNMO) also shows considerable attention for LIBs in recent years. Owing to high working voltage (4.7 V vs. Li/Li⁺), such spinel shows the highest specific energy (650 wh kg⁻¹ at a material level) compared to other cathode materials (LCO, 540 Wh kg⁻¹; LFP, 500 Wh kg⁻¹; LMO, 500 Wh kg⁻¹). The interstitial space in the spinel framework offers 3D Li⁺ diffusion channels. Owing to its high working voltage, LNMO has become a very promising candidate for replacing spinel LiMn₂O₄ cathode. Moreover, relatively benign metal elements make such a material a top choice for next-generation high-power batteries. However, there are still some serious limitations to restrict its commercial applications. Ni is active with the two Ni²⁺/Ni³⁺ and Ni³⁺/Ni⁴⁺ couples at a 4.7 V, while Mn is inactive without capacity contribution. Although Ni²⁺/Ni³⁺ and Ni³⁺/Ni⁴⁺ electrochemical couples provide favorable high working voltage, they also bring some unwanted side reactions between active component and electrolyte, which gives rise to a considerable thickening of the protective solid-electrolyte interphase (SEI) layer passivating the material surface, thus leading to Li⁺ ion diffusion obstacle.

High redox couple of Ni³⁺/Ni⁴⁺ easily incurs the decomposition of conventional carbonate-based electrolyte. It is difficult for pure LNMO phase to be synthesized because of mixed ordered and disordered LNMO phase owing to oxygen deficiency caused by high-temperature annealing. In addition, the Li_xNi_1−xO₂ impurity probably forms as the second phase. Compared to ordered LNMO phase with P4₃32 space group, disordered phase is relatively easily synthesized. The disordered spinel shows a better electrochemical performance than ordered spinel because the disordered one has an increased Li⁺ ion diffusion coefficient. However, the presence of a small amount Mn³⁺ in the disorder phase will cause a non-negligible capacity fading during cycling, especially at elevated temperature. In this respect, introducing metal doping is an effective strategy to increase the average valence of Mn in the spinel. Mn³⁺ concentration in the spinel can be also decreased by reannealing treatment, which can also reduce oxygen defect and impurity amount, resulting in better electrochemical performance. In recent years, various strategies have been widely used to enhance electrochemical performance of LNMO, such as doping and surface coating. Metal doping can provide better chemical and structural stability and higher electrical conductivity. For example, Ti-doped LNMO improved the disordering of the transition metals and lowered the symmetry from simple cubic structure (P4₃32) to face-centered spinel (Fd3m) [16]. Importantly, Ti-doped LNMO spinel shows higher working voltage, faster Li⁺ ion diffusion, and better rate capability than undoped spinel. Fe doping in the high-voltage spinel can stabilize the solid during repeated cycling, leading to better cycling performance [17]. Thus, double doping by using Ti and Fe is an effective strategy to synergistically enhance the structural stability of LNMO. Additionally, other dopants including Cr, Ru, Zr, Al, and Mg can show better rate capability and cycling performance. Besides metal doping, anion dopant can also enhance electrochemical performance of high-voltage spinel. For example, the F-doped LNMO can restrict the formation of NiO impurity and change the lattice parameters and bonding energy [18], leading to fine structure and significantly enhanced cycling performance.

On the other hand, similar to layered NMC, the interface stability of LNMO between electrode and electrolyte is also a key for achieving long-term cycling performance. It has been demonstrated that better stability can be obtained by coating more stable surface layer, such as metal (Zn, Au, Ag), metal oxides (ZnO, SnO₂, ZrO₂), and metal phosphates (Li₃PO₄). For instance, ZnO-coated LNMO can effectively suppress Mn dissolution and increase the stability. More importantly, HF content in the electrolyte can be restricted because ZnO protection layer acts as a scavenger of fluoride anions from HF generated from the decomposition of LiPF₆ salt in the electrolyte by transforming HF to ZnF₂ [19, 20]. Li₃PO₄ protects the surface of LNMO and functions as a solid-electrolyte interface between active component and solid polymer electrolyte (SPE) to prevent the SPE degradation [21]. Such findings show that surface coating can not only enhance structural stability of LNMO but also restrict Mn dissolution, thus leading to improved cycling stability.

4. Polyanion-based cathode

4.1 Olive LiFePO₄

Polyanion cathode materials have received considerable attention in LIBs owing to their stable crystal structure and long-term cycling performance. Since John B. Goodenough proposed LiFePO₄ as cathode materials for LIBs in 1997, it has received much investigation including academic and industrial activities. As shown in Figure 7, LiFePO₄ possesses an olive structure with space group Pnma. The O atoms are located in a slightly distorted, hexagonally close-packed arrangement. The P atoms occupy tetrahedral sites, while Li and Fe atoms occupy octahedral 4c and 4a sites, respectively. The LiO₆ octahedra from edge-sharing chain along the b-axis. Each FeO₆ octahedron is linked with four FeO₆ octahedra by common corners in the b–c plane, forming zigzag planes. One FeO₆ octahedron has common edges with two LiO₆ octahedra. PO₄ groups share one edge with one FeO₆ octahedron and two edges with LiO₆ octahedra.

Figure 7.

The crystal structure of olivine LiFePO₄ in projection along [001].

Compared to layered LiCoO₂, LiFePO₄ has excellent structural stability and thermal stability owing to the strong covalent P-O bonds in the (PO₄)³⁻ polyanionic clusters. The host framework can make host stable enough upon repeating cycling without evident structural changes and temperature limitations and no loss of oxygen. These advantages offer safe enough application for LIBs. However, long-term air exposure at a moderate temperature will give rise to surface oxidation of LiFePO₄/C nanocomposites with the formation of a LiFePO₄(OH) phase. Despite these drawbacks, the advantages of LFP including cycle stability, safety, environmental friendliness, and low cost make this material a very promising cathode candidate in LIBs. Currently, it has been successfully commercialized by many battery manufactures such as A123 and BYD.

Although LiFePO₄ possesses many merits, it still has some drawbacks such as low intrinsic electric conductivity (~ 10⁻⁹ S cm⁻¹) and low packing density. During past decades, much effort has been focused on addressing these problems by optimizing synthesis method, surface modification, reducing particle size, and metal doping [22]. Among them, surface modification has been widely accepted as an effective approach for enhancing the electronic conductivity of LiFePO₄. Carbon coating is the most popular surface modification strategy for LiFePO₄. It can not only improve the electric conductivity of active particle but also act as a reducing agent to prevent the formation of Fe³⁺ impurity phase, as illustrated in Figure 8. In the meanwhile, the dense protection layer can effectively isolate the LiFePO₄ particle to inhibit the undesirable particle growth and self-aggregation. Nevertheless, a large amount of carbon reduces the volumetric energy density of the electrode. Various carbon sources have been used to generate carbon layer such as sucrose, starch, resorcinol-formaldehyde gel, sugar, carbon black, polypropylene, hydroxyethylcellulose, and so forth [22]. Therefore, carbon coating should be thin enough on the active particle surface. Wang et al. prepared uniform carbon-coated LiFePO₄ nanoparticle by an in situ polymerization restriction method for the synthesis of C/LiFePO₄ nanocomposite with a semi-graphitic carbon shell with a thickness of 1–2 nm [23], as shown in Figure 9. Charge/discharge tests show that the so-formed nanocomposites exhibit superior rate capability and cycling stability (Figure 9). In addition to the effect of carbon content in the nanocomposite, the resultant carbon by different organic precursors shows different electronic conductivities, indicating that sp²-coordinated carbon exhibits higher electronic conductivity than sp³-coordinated or amorphous carbonaceous materials. Besides carbon coating, other surface modifications including metallic conducting layer and conducting polymer film are also investigated to enhance the electrochemical performance of LiFePO₄. Besides LiFePO₄, other polyanion-based cathode materials have also been developed such as Li₃V₂(PO₄)₃, LiFeSO₄F, LiFeSiO₄, and so on.

Figure 8.

Schematic diagram illustrating Li⁺ ion and electron transport pathways for active nanoparticle.

Figure 9.

(a) The ideal structure of carbon-coated LiFePO₄ nanoparticles by a complete and uniform coating, (b) HRTEM image of carbon-coated LiFePO₄ single nanoparticles, (c) cycling performance of carbon-coated LiFePO₄ at a current density of 0.1 A g⁻¹ between 2.0 and 4.3 V, and (d) charge/discharge profiles at different cycle numbers.

5. Li-free cathode

Besides the above Li-containing cathode, there are some Li-free cathode materials that have been studied for LIBs. The most common one is vanadium oxide. Owing to the multiple valence characteristic of vanadium (+2, +3, +4, +5), vanadium pentoxide (V₂O₅) possesses versatile redox-dependent properties and exhibits a great potential in the rechargeable batteries. V₂O₅ has a typical layered structure, which is very beneficial to Li⁺-ion intercalation/de-intercalation. In recent years, V₂O₅ has been widely studied as cathode materials for LIBs. V₂O₅ is a typical intercalation compound with a layer structure with an orthorhombic structure belonging to P_mmn space group. Figure 10 shows the crystal structure of orthorhombic V₂O₅, which consists of edge- and corner-sharing layered square pyramids, forming V₂O₅ nanosheets linked together by a weak V-O interaction orthogonal to the c-direction. In a single-layer slab, there are three crystallographically distinct oxygen centers (O1, O2, and O3). The single coordinated terminal/apical vanadyl oxygen atom, O(1), has a relatively short V-O bond length of ca. 1.54 Å; the bridging O2 oxygen connects with two adjacent V atoms by corner-sharing VO₅ square pyramids. The corresponding V-O bond length is ca. 1.77 Å. The O3 with triply coordinated chaining O atoms connects three V atoms by edge-sharing VO₅ square pyramids, and the three corresponding V-O bond length are 1.88, 1.88, and 2.02 Å. The orthorhombic V₂O₅ can be described as highly distorted VO₆ octahedral building blocks to form the layered anisotropic structure. The long sixth V-O bond underlines the structural anisotropy of such material and the ability to insert guest species in perovskite-like cavities. During lithium insertion into layered V₂O₅, it will induce the formation of vanadium bronzes as follows:

Figure 10.

Crystal structure of orthorhombicV₂O₅, view from (a) ac plane, (b) ab plane, and (c) the coordination environment around a single vanadium atom. The V and O atoms are represented in yellow and blue spheres.

$${\mathrm{V}}_2{\mathrm{O}}_5+{\mathrm{xLi}}^{+}+{\mathrm{xe}}^{-}={\mathrm{Li}}_x{\mathrm{V}}_2{\mathrm{O}}_5$$ E1

Besides crystallized V₂O₅, hydrated vanadium pentoxide (V₂O₅ nH₂O) also shows promising lithium storage properties, such as V₂O₅ nH₂O xerogels, V₂O₅ nH₂O aerogels, and V₂O₅ nH₂O glasses [24]. These low crystalline or amorphous materials exhibit attractive merits owing to their unique nanostructures (high active surface area, small particle size, high diffusion coefficients, low-volume change upon repeated Li insertion/extraction processes). An effective approach to increase the intercalation capacity of V₂O₅ is to adjust the interlayer structure and the interaction forces between the adjacent layers. Usually, when the interlayer distance between V₂O₅ layered structure increases, the lithium insertion capacity will increase. However, owing to the absence of lithium in the V₂O₅, Li metal or Li-containing anode must be used for LIBs.

6. Thin-film battery applications of cathode materials

Among various battery chemistries, thin-film batteries are considered as the most competitive power sources owing to their high volumetric energy density, gravimetric specific energy, superior power capability, and good flexibility, as shown in Figure 11. Each cell is constructed by cathode, anode, and electrolyte, which is the same to conventional LIBs. To achieve high specific energy of thin-film battery, two requirements must be met by electrode materials: (i) high specific charge density (in Ah kg⁻¹ and Ah L⁻¹) and (ii) high cell working voltage that requires high cathode redox potential and low anode redox potential. In addition, electrode reactions should be highly reversible on both cathode and anode for guaranteeing good cycling performance. Thus, developing advanced battery chemistry is a key for high-performance thin-film batteries. This section only focuses on cathode materials for their thin-film microbattery application.

Figure 11.

(a) Comparison of the gravimetric and volumetric energy density of various chemical batteries including Pb-acid, Ni-MH, Li-ion batteries, and thin-film Li-ion batteries. (b) Cross-sectional schematic diagram of thin-film Li-ion battery.

For practical cells, carbon-based anode and metal oxide cathode have been employed since they can offer maximum specific energy, sufficient power density, and long-term cycle life. Currently, several cathode materials (LiCoO₂, LiNiO₂, LiMn₂O₄, LNMO, and V₂O₅) are relatively popular. These electrodes are relatively stable in the air. In the case of V₂O₅, lithium metal laminate is attached to the carbon anode. After filling in the electrolyte, the metallic Li reacts with carbon and generates Li_xC₆, which is a lithium source in the cell.

LiCoO₂ is one of the most common used cathode materials for thin-film LIBs. Kuwata et al. prepared thin-film LIBs by using LiCoO₂, amorphous SnO, and amorphous Li₂O-V₂O₅-SiO₂ as cathode, anode, and solid electrolyte, respectively [25]. As illustrated in Figure 12, the battery has an area of about 0.23 cm², and the thickness is around 2 μm. All films were grown by PLD. The thin-film battery is fabricated on the quartz glass substrate, and metal Pt was deposited by PLD onto the substrate as a current collector. The solid electrolyte Li_3.4V_0.6Si_0.4O₄ was prepared by solid-state reaction from Li₂CO₃, SiO₂, and V₂O₅ and then was pressed into a disk with a diameter of 13 mm and annealed at 1100°C for the targets. The LiCoO₂ powder was pressed and sintered in air (900°C) as cathode, and the SnO powder was pressed into the disk without annealing. As displayed in Figure 12, the thicknesses of SnO, LVSO, and LiCoO₂ films are estimated to be 150, 1100, and 400 nm, respectively. The total thickness of the battery is around 2 μm. The interface between the LiCoO₂, SnO, and LVSO electrolyte remained smooth before and after cycling.

Figure 12.

(a) Schematic diagram (cross-sectional view) of thin-film LIBs, (b) SEM micrographs showing cross-sectional view before cycling, (c) after 100 cycles, (d) charge/discharge curves of the thin-film LIBs in the second cycle at a current of 44 mA/cm², and (e) the specific capacity as a function of cycle number.

The thin-film battery was tested at a cutoff window of 3.0–0.7 V. As shown in Figure 12, the discharge capacity in the second cycle is about 9.5 mAh cm⁻², corresponding to 10–20% of LiCoO₂ utilization. The discharge potential gradually decreases in the range from 2.7 to 1.5 V owing to the amorphous nature of the SnO film. Such a thin-film battery shows a rapid capacity decay in the first cycle, which is probably attributed to the irreversible formation of Sn-Li alloy and amorphous Li₂O. Despite this, such battery was still cycled for 100 cycles with good Coulombic efficiency (Figure 12).

As discussed previously, various cathode materials (LiCoO₂, LiMn₂O₄, LiFePO₄, V₂O₅) have been studied for thin-film LIBs. However, they show limited specific energies. To further improve the energy density of the Li-ion thin-film microbattery, an effective solution is using high-potential cathode material. It is well known that LNMO is a very promising high-voltage cathode material owing to its high working potential at 4.7 V vs. Li/Li⁺. Nowadays, LNMO thin films have been deposited by various techniques such as sol-gel, electrostatic spray deposition, pulsed laser deposition (PLD), and RF magnetron sputtering deposition. Letiche et al. prepared LNMO thin film on functional current collector by a radiofrequency magnetron sputtering method [26]. The authors first solved the technological issue owing to PtSi phase which originates from the interdiffusion between silicon wafer and Cr/Pt current collector to form PtSi phase under annealing treatment. By doping the Cr layer with a dense and pinhole-free Al₂O₃ thin film deposited by ALD acting as a barrier diffusion between the Si substrate and the LNMO layer, the synthesis process has been validated. In addition, the authors demonstrated that the deposition pressure plays a key role on the Ni-Mn cation ordering in spinel-like structure (P4₃32 ordered vs. Fd3m disordered spinel) and consequently on the electrochemical performance. It indicates that LNMO thin film deposited at 10⁻² mbar and annealed at 700°C exhibits a normalized capacity of 65 μAh cm⁻² μm⁻¹ with good capacity retention during cycling. The Coulombic efficiency is highly dependent on the film thickness of high-voltage thin-film electrodes.

As shown in Figure 13, the TEM image obtained by the focused ion beam (FIB) technique shows the different stacked layers on Si wafer where the thickness is close to 85 nm for Al₂O₃, 30 nm for Pt current collector, and 480 nm for LNMO thin-film electrode. Moreover, LNMO layer exhibits a polycrystalline nature with a different crystallized orientation (domains of ~ 20 nm), except for a thickness of 26.5 nm along the Pt layer where a (111) preferred orientation is seen, as displayed in Figure 13. The LNMO thin film annealed at 700°C delivers the best performance.

Figure 13.

(a) The as-prepared LNMO thin film, (b) TEM analysis of the cross section Si/Al₂O₃/Pt/LNMO/Ti/SiO₂ prepared by focus ion beam technique, (c) HRTEM image of the LNMO layer with a focus close to the Pt current collector in order to check the Pt/LNMO interface, (d) discharge curve of LNMO thin film at different annealing temperatures, and (e) evolution of the surface capacity vs. cycle number as a function of the thin-film thickness.

7. Conclusions

This chapter mainly discusses the development of currently the state-of-art cathode materials for LIBs and analyzes their potential applications in thin-film Li-ion microbatteries. For Li-containing cathode, layered structured LiCoO₂, NMC, and NCA, spinel LiMn₂O₄, and polyanion-based LiFePO₄ are currently the most popular for LIBs. Among them, LiCoO₂ is mainly used in the field of consumer electronics, while LiMn₂O₄, LiFePO₄, NMC, and NCA can be employed as power battery cathode materials for EVs. In particular, Ni-rich NCA and NMC-811 have been currently considered as the most attractive cathode candidate for long-range EVs due to their high specific energy compared to Ni-less NMC. In this respect, Ni-rich NMC cathode shows a sizable potential for high-energy thin-film LIBs in the future, but they are seldom investigated in thin-film Li-ion battery. Besides developing high-capacity cathode materials for realizing high-energy thin-film LIBs, exploiting high-potential cathode materials is also an effective strategy. Spinel LNMO is a very promising cathode candidate owing to its high redox potential (4.7 V vs. Li/Li⁺), which has been widely investigated as cathode for thin-film LIBs. Additionally, Li-free cathode materials including V₂O₅, V₆O₁₃, MoS₂, TiS₂, and MnO₂ show attractive advantages for thin-film LIBs. Finally, the commercial success of thin-film Li-ion batteries will largely depend on the significant breakthrough of electrode materials including cathode, anode, and solid electrolyte in terms of energy density, weight, size, flexibility, and electrode/electrolyte interface.