2-1 Features of Rechargeable Lithium-ion Batteries 2-1-1 Basic Concepts of Lithium-ion Batteries
The rechargeable lithium-ion battery has been considered as the most promising battery because of its high power/energy density, long cycling life, environmentally friendly and no memory effect. The energy densities per unit volume (Wh/l) and per unit weight (Wh/kg) of various rechargeable batteries are shown in Fig. 2-1 [6].
Lithium-ion batteries are appealing for these high power applications since they provide higher energy density than the other rechargeable battery systems, such as lead-acid, nickel-cadmium (Ni-Cd), and nickel-metal hydride (Ni-MH) batteries. The nominal voltage of a lithium-ion battery is around 3.7 V, which is much larger than that of conventional nickel batteries (1.2 V) and lead-acid batteries (2 V). The high voltage of the lithium-ion battery arises as it uses the chemistry of intercalation reactions of Li-ion with the cathode and anode.
Rechargeable lithium-ion batteries are generally composed of Li-containing transition metal oxide as the positive electrode and a carbon material as the negative electrode material. When the cell is constructed, it is in the discharge state. When charged, Li-ions move from the positive electrode through the electrolyte and electrons also move from the positive electrode to the negative electrode through the external circuit with the charger. As the potential of the positive electrode rises and that of the negative electrode descends during charging, the voltage of the cell becomes higher.
The cell is discharged by the connection of a load between the positive and negative electrodes. In this case, the Li-ions and electrons move in opposite directions to charging. Consequently, electrical energy is obtained. The principle charge/discharge
doi:10.6342/NTU201601970
4
for Li-ion batteries, such as C/LiCoO2 system, is illustrated in Fig. 2-2. The reaction mechanisms are shown as follows:
Cathode reaction: between graphene layers of carbon and the charge balance is maintained by a oxidation of the Co3+ ions to Co4+. During discharge, exactly the reverse process involving the extraction of lithium from the graphene layers and an reduction of Co4+ ions to Co3+
occurs. In other words, when the battery is charged, the Li ions in the LiCoO2 migrate via a separator to between the layers of carbon material that form the anode, and charging current is applied. When the battery is discharged, the Li ions in the carbon material migrate via a separator to the LiCoO2, and a discharging current occurs.
doi:10.6342/NTU201601970
5
Figure 2-1 Comparison of the different battery technologies in terms of volumetric and gravimetric energy density [6].
Figure 2-2 Schematic illustration of the charge/discharge process in a Li-ion cell consisting of Li insertion compounds as both anode and cathode.
doi:10.6342/NTU201601970
6
2-1-2 Developments of Lithium Batteries
Over the past decades, there has been enormous research in the development of Li-based batteries, especially for rechargeable Li-ion batteries. The motivation for using a battery technology based on metallic Li as anode relied on the fact that Li has low atomic mass (6.94 g/mol), high specific capacity (3862 mAh/g) and a high electrochemical reduction potential (-3.045V versus standard hydrogen electrode) [7].
The advantage in using Li metal was first demonstrated in the 1970s with the assembly of primary (for example, non-rechargeable) Li cells by Ikeda et.al [6]. In 1972, a first secondary rechargeable Li battery was proposed by Exxon [8]. The battery was composed of TiS2 as the positive electrode, Li metal as the negative electrode and Li perchlorate in dioxolane as the electrolyte. TiS2 was the best intercalation compound available at the time, having a very favorable layered-type structure. Though all these cathode materials have proven their competence for Li batteries, it soon encountered the shortcomings of a metallic Li/non-aqueous electrolyte combination. Uneven (dendritic) Li metal deposit was formed onto anode during repeated recharge cycles, which caused the electrode to swell and led to explosive hazard.
Because of the safety issues of using metallic Li, several alternative approaches were pursued in which either the electrolyte or the anode was modified. The first approach [9] involved substituting Li metal for a second insertion material. The concept was first demonstrated in the laboratory by Murphy et al. [10] and then by Lazzari et al. [11], in which LiWO2 was demonstrated to be able to work as a new anode without the above safety problem since the presence of Li was then in the ionic state, but not metallic. Since then, the name of Li battery was changed to Li-ion battery. In June 1991, Sony Corporation [12] marketed the all-time most successfully commercialized LIB (to date) of LiCoO2/C system. And just as the circumstances mentioned at the beginning,
doi:10.6342/NTU201601970
7
this small but long-time usable battery was quickly applied in those portable electronic devices immediately made the sale number of mobile phones skyrocket.
The second approach [13] involved with replacing the liquid electrolyte by a dry polymer electrolyte, which was called Li solid polymer electrolyte (Li-SPE) batteries.
But this technology required temperature up to 80 oC; therefore it was limited to large systems and not to portable devices. After that, Kelly et al [14] tried to develop a Li hybrid polymer electrolyte (Li-HPE) battery. However, Li-metal dendrites were still a safety issue to HPE systems. In 1996, Bellcore researchers introduced polymeric electrolyte in a liquid Li-ion system [15], which is called plastic Li-ion (PLiON) batteries. PLiON system can be manufactured in various size and shape (Fig. 2-3) [6].
Such a thin film battery technology that offers shape versatility, flexibility and lightness, has been developed commercially since 1999.
Until now, rechargeable Li-ion battery is the fastest growing battery technology in the world. With so much effort of researchers, various electrode materials have been investigated and developed. For the cathode materials, there are LiNiO2, LiMn2O4, Li(NixCoyMn1-x-y)O2, LiFePO4, and LiMnPO4, etc. And for the anode materials, there are still TiO2, Li4Ti5O12, Si-C composites, Sn/C/Co and the other metals or alloys.
Each material shows distinct advantages and drawbacks. The potential window versus capacity for positive and negative electrode materials is show in Fig. 2-4 [16]. In the following sections, some specific materials were selected to be particularly discussed.
doi:10.6342/NTU201601970
8
Figure 2-3 Schematic drawing showing the shape and components of various Li-ion battery configuration: a: cylindrical; b: coin; c: prismatic; d: thin and flat [6].
Figure 2-4 Voltage versus capacity for positive- and negative-electrode materials [16].
doi:10.6342/NTU201601970
9
2-2 Introduction to Cathode Materials for Li-ion Batteries
For advanced the energy density of LIB, various alternatives to LiCoO2 cathode have been investigated and developed, such as LiNiO2, Li(M, Mn)2O4, LiCoxNiyMn 1-x-yO2, Li1+x(M,Mn)1-xOy, LiFePO4, and LiMnPO4…etc. Each material shows distinct advantages and drawbacks. From a crystal structural point of view, we can categorize the cathode materials into three major groups: layered structure, spinel structure and olivine structure as shown in Fig.. 2-5.
Figure 2-5 Representative crystal structures of cathode materials for lithium-ion batteries: (a) layered; (b) spinel; (c) olivine structured. Black lines demarcate one unit cell in each structure [17].
2-2-1 Olivine Structure
LiMPO4 (M=Fe, Mn, Co) with olivine structure are the advanced Li insertion cathode materials. In 1997, Goodenough’s group firstly proposed lithium iron phosphate (LiFePO4) as a cathode material for secondary Li-ion batteries [18]. This compound has the orthorhombic unit cell with space group of Pmna, containing a nearly close-packed-hexagonal oxide-ion array. The unit cell parameters are a=1.0332(4) nm, b=0.6010(5) nm, and c=4.692(2) nm [19], LiFePO4 is basically composed of
edge-doi:10.6342/NTU201601970
10
shared LiO6 octahedra, corner-shared FeO6 octahedra, and PO4 tetrahedra, as shown in Fig. 2-5 (c). The P-O covalent bonds in PO43- not only strongly bind the oxygen from easily release with large thermal energy causing thermal runaway, but also lower the energy of Fe3+/Fe2+ redox couple and hence raise open-circuit voltage vs. Li, as in other cases of polyanions [20, 21]. For the former effect of P-O covalent bond, in fact, using polyanions was the initial concept to overcome the structural instability of layered iron oxide, LiFeO2, in which the Fe3+ ions tended to migrate from octahedral to tetrahedral sites. The open circuit voltage of LiFePO4 is ~3.4V. When charging and discharging at low rate, i.e., at nearly equilibrium condition, LiFePO4 exhibits a flat plateau at ~3.4V as well, implying that the reversible electrochemical process is via a two-phase reaction.
The typical charge and discharge curves show flat-plateau feature, and this feature of flat plateau would remain even at considerable high C-rate as shown in Fig. 2-6, though the voltage is lower at higher C-rate due to high polarization [22]. Padhi et al. further demonstrated that the other phase could be FePO4. They used chemical de-lithiation method to obtain FePO4 and showed that it had similar discharge curve as LiFePO4
discharged [18]. This assumption was then confirmed by Anderssen et al. by using in-situ X-ray diffraction and Mössbauer spectroscopy technique [23]. This reversible electrochemical reaction can thus be described as the following equations:
During charge:
4 4
4
Li xe ( 1 x ) LiFePO xFePO
LiFePO
(2-4)During discharge:
4 4
4
xLi xe xLiFePO ( 1 x ) FePO
FePO
(2-5)doi:10.6342/NTU201601970
11
The densities of two end compounds have a difference of 2.59 % since the volume of FePO4 decreases by 6.81% after Li ions diffuse out from LiFePO4. This small volume change along cycling is one of the reasons that LiFePO4 being considered as a cathode material with high stability. According to the data of Delacourt’s group [24], the electronic conductivity of LiFePO4 is as low as 2.39×10-9 S/cm at 25 oC, while the ones of the currently used LiCoO2 and LiMn2O4 are ~10-3 and ~10-5S/cm, respectively.
Following the success in the development of LiFePO4 cathode, there have been encouraging efforts in investigating the LiMnPO4 and LiCoPO4 as cathode materials.
The former has the moderate redox potentials of 4.1V, while the latter has the highest one of 4.8V among olivine cathodes [25]. However, LiCoPO4 cathode suffers from stability problems caused from itself and non-aqueous electrolytes that are reflected in a relatively low specific capacity <120 mAh/g (versus 165 mAh/g theoretical) and capacity fading during prolonged cycling.
Figure 2-6 Discharge curves of olivine LiFePO4 at high rates [22].
doi:10.6342/NTU201601970
12
2-2-2 Spinel Structure
Another promising cathode material is LiMn2O4 that forms a spinel structure ( Fd3m ), where manganese occupies the octahedral sites and lithium predominantly occupies the tetrahedral sites. In spinel structure, the lithium ion migration pathways are a 3-dimensional network of channels rather than planes, as in the α-NaFeO2 structure. It can deliver a high energy density comparable to LiCoO2
because of its high voltage and high specific capacity which is only 10% less than that of LiCoO2. In addition to energy density, spinel LiMn2O4 also has friendly properties, such as low cost, low toxicity and safety in overcharge [26, 27]. The characteristic potential signatures for LiMn2O4 exhibit two plateaus at 4.1 V and 4.05 V (Fig. 2-7 (a) [28]). Further discharged to about 2.8 V, another plateau corresponding to a cubic-tetragonal phase transition (Fig. 2-7 (b),(c)), LiMn2O4→ Li2Mn2O4 (valence state:
+3.5+3, in Fig. 2-7.)
For the spinel lithium manganite, several reasons have been proposed for the capacity loss of Li//LixMn2O4 cells as follows [27, 29, 30] :
(i) The disproportional reaction according to 2Mn(III)Mn(IV) + Mn(II) , resulting in the dissolution of manganese metal. The dissolution of Mn could also result in higher fading rate of cycle life of LiMn2O4 cathode. Its own in that a coating of MnO2 coming from the second reaction shown above on LiMn2O4 particles was observed as well.
Surface coating with transition metal oxides has been shown as an effective way to solve the dissolution problem [27].
(ii) The instability of the delithiated spinel structure by oxygen loss in organic electrolyte solvents in the end of the charge.
(iii) Jahn-Teller distortion at the end of discharge. A phase transition from cubic to tetragonal can cause a large and anisotropic volume change that results in the severe
doi:10.6342/NTU201601970
13
damage to spinel LiMn2O4 cathode. It is due to the incompatibility between oxygen arrays in cubic and tetragonal structure, which can decrease the electrical contact between particle surfaces.
In addition, Substitution of cations into the Mn sites in LiMn2O4 structure has been proven to be another effective way to improve the drawbacks as mentioned. Among these doped spinel cathode materials, Ni-substituted LiNi0.5Mn1.5O4 spinel has been regarded as special and drawn more attention because of its high operational voltage (~4.7 V vs. Li+/Li0), good cycling and rate capabilities at room temperature. It shows great potential as one of the next generation cathode materials with higher energy density. However, it is need to overcome that the fact that spinel LiNi0.5Mn1.5O4 exhibits a significant capacity loss at an elevated temperature, which is a critical environment for HEV/EV applications [31].
Figure 2-7 (a) Profile of discharge curve in spinel lithium manganite and schematic structure of (b) cubic spinel of LiMn2O4 and (c) tetragonal spinel of Li2Mn2O4 [28].
doi:10.6342/NTU201601970
14
2-2-3 Layered Structure
Layered Lithium cobalt oxide (LiCoO2)
LiCoO2 was first proposed as a cathode material by Goodenough et al. in 1980 [32]. Even now, The LiCoO2 cathode material is most commonly used in lithium ion batteries. LiCoO2 has α-NaFeO2 crystal structure (space group: R m3 ) with oxygen in a cubic closed-packed arrangement (Fig. 2-8 [33]). With the CoO6 octahedrons formed as planes, LiCoO2 shows a layered structure which enables Li ions diffuse in the lattice.
Theoretically, LiCoO2 has a capacity of 274 mAh/g when fully de-lithiated, but the total de-lithiation would go through an unfavorable phase change, hexagonal-monoclinic transition, which may cause 50% volume expansion along c-axis [34, 35].
This phase change occurs around 4.3V. Hence, in practical, LiCoO2-based Li-ion battery would be prevented from charging to higher than 4.3V, corresponding to Li0.5CoO2, and shows only half of theoretical capacity, 140 mAh/g. In this range, LiCoO2 exhibits great stability and low self-discharge capacity.
In addition, LiCoO2 is its high cost because Co element is less available, and thereby limits its applications on electric vehicles. The other one is the toxicity over environments. And the final one is the Co dissolution which is induced by HF when using LiPF6 as electrolyte or poly-vinylidene-fluoride (PVdF) as binder, particularly at high temperatures or with moisture. Some researchers have tried to coat metal oxides on LiCoO2 particles to reduce the contact between HF and LiCoO2. And it was revealed that metal oxide coating can not only effectively enhance the cycle life, but also improve the reversible capacity to 170 mAh/g by the oxide coating suppressing the volumetric expansion [36-38].
doi:10.6342/NTU201601970
15
Figure 2-8 The layered structure of LiCoO2 with R m3 symmetry [33].
Layered lithium nickel oxide (LiNiO2)
LiNiO2 has the same structure with LiCoO2 (Fig. 2-8 [33]), only the Ni atoms occupy the Co sites. It was first reported as a cathode material for lithium ion batteries early in 1979. Its theoretical capacity is also very close to LiCoO2, but it can deliver higher practical capacity over 180 mAh/g. However, LiNiO2 is not widely fabricated for Li-ion batteries because of some annoying problems. First, LiNiO2 has the high irreversible capacity at the first cycle. Second, it was discovered that dramatic capacity
doi:10.6342/NTU201601970
16
fading would occur after tens of cycles which might be due to the migration of Ni ions from original layers to Li layers. The presence of Ni ions in Li layers may induce anisotropic volumetric lattice change along a- and c-axis resulting in cracks and reducing the diffusion coefficient of Li ions [39]. Third, the most serious problem is its stability at high temperatures. LiNiO2 become unstable at high temperature and is easy to release oxygen at charged state, which would react with electrolyte as well as lithiated carbon anode and lead to thermal runaway [40, 41]. The structure limitations of LiNiO2 can be partly overcome by replacement of Ni. LiMxN1-x, such as Co, Mn, Mg, Al, and Cr was proposed and thus result in some promising multi-cation oxides.
Layered lithium manganese oxide (LiMnO2)
LiMnO2 has also been extensively studied for rechargeable lithium batteries because of its high theoretical capacity 285 mAh/g based on a Mn3+/Mn4+ couple, about twice that of LiMn2O4, and low cost compared to LiCoO2. LiMnO2 is not thermodynamically stable as the layered structure, but as an orthorhombic phase o-LiMnO2. It is well known that LiMnO2 transforms irreversibly to a material with spinel-like structure during electrochemical cycling [42-45]. However, the mesostable LiMnO2 could be stabilized in the layer structure by a number of possible substituent elements. For example, substituting a small fraction (< 5%) of the Mn with Al [46] and with Cr [47]. Another phase of interest in this system is the monoclinic polymorph of LiMnO2, which has the cation ordering of the α-NaFeO2 structure, where Li ions are located in octahedral sites between MnO6 sheets. Unlike LiCoO2, which is rhombohedral ( R m3 ), LiMnO2 is monoclinic (C2/m) because the coordination polyhedron around the Mn3+ ions is distorted from a regular octahedron due to the Jahn-Teller effect.
doi:10.6342/NTU201601970
17
Layered lithium multi-cation oxide (LiNixCoyMn1-x-yO2)
LiNixCoyMn1-x-yO2 is the cathode material which successfully integrates the advantages of LiCoO2, LiNiO2 and LiMnO2, and has garnered the attention of both fundamental scientists and applied researchers. LiNixCoyMn1-x-yO2 has the same structure as LiCoO2 and LiNiO2 in -NaFeO2 with a cubic close-packed arrangement of the oxide ions. The doped Co and Mn ions homogeneously substitute the Ni sites. Of course, some exceptions, like doping into Li sites, are also possible. To identify the doping sites of cations, the ratio of “a”, the distance between metal ions, to “c”, the distance between the slabs of interlayer (also shown in Fig. 2-8), was considered an important index. For ideal cubic close-packed lattice, the ratio of c/3a should be 1.633.
With higher content of transition metal cations in Li sites, the ratio would be closer to this value. And it was found that the ratio of c/3a of LiNiO2 was 1.639 [48] while the ratio of LiNi1/3Co1/3Mn1/3O2 was 1.657 [49].
Flashing back to the development of mixed-transition material. Delmas [39, 50] et al. developed that the homogeneous Co doping in LiNiO2 to form LiNi1-xCoxO2. The effects of Co doping include higher structural stability from suppressing Ni migration to Li sites and higher thermal stability from less losing oxygen [51]. In 1992, Dahn et al. [52] suggested doping Mn, instead of Co, into Ni sites forming LiNi1-yMnyO and found the optimum performance on capacity and cycle life when y=0.5. Furthermore, oxides with three-transition-metal mixing were proposed one from another from 1999 [53]. Because these new materials showed the attractive properties of higher capacity, higher operating voltage, and longer cycle life compared with LiCoO2 and LiNiO2, many related studies were soon published. Ohzuku et al. [54] introduced an oxide of LiNi1/3Co1/3Mn1/3O2 synthesized at 1000 oC and characterized a capacity of 150 mAh/g between 2.5–4.2 V. Its reversible capacity was measured to be 160 mAh/g in the
cut-doi:10.6342/NTU201601970
18
off range of 2.5–4.4 V (Fig. 2-9 [55]). Additionally, It was then further revealed that the capacity could be raised to more than 200 mAh/g when charging to 4.6 V and the cycle life was still good [49]. Electronic structure studies have shown that it consists of Ni2+, Mn4+ and Co3+, and reversible capacity involves the oxidation of Ni2+ to Ni4+ with a two-electron transfer during the initial stage of Co3+ to Co4+ in the later stage [56, 57].
Thus the higher capacity of layered LiNi1/3Co1/3Mn1/3O2 could be due to the improved chemical stability associated Ni2+/3+ and the Ni3+/4+ redox couple compared to Co3+/4+
redox couple
Figure 2-9 Voltage vs. capacity profiles for the cell, Li/ Li(Ni1/3Co1/3Mn1/3)O2 in (a) 2.5－4.4 V. (b)2.5－4.7 V (c) Cycling performance of the cells [55].
doi:10.6342/NTU201601970
19
Layered Lithium-rich Cathode Materials (LrMOs)
The layered lithium-rich Mn-transition metal oxides (LrMOs), generally represented as Li1+y(Mn, M)1-yO2 (M = Mn, Ni, Co), are of interest as a new generation of cathode materials for high energy density of lithium ion batteries. In 2001, the series of compounds Li[Li1/3-2x/3NixMn2/3-x/3]O2 (0≦x≦1/2) with the absence of expensive and toxic Co element were first reported by Lu and Dahn et al [58]. On the other hand, Thackeray et al. applied the ‘‘composite oxide’’ concept, by which the Li-rich Mn-based layered oxide can be regarded as a ‘‘composite oxide’’ consisting of LiMO2 (M=
Mn, Ni, Co) and Li2MnO3 components [59, 60]. This is an alternative approach to stabilize the structural of delithated Li1-xMO2 and improve the electrochemical
Mn, Ni, Co) and Li2MnO3 components [59, 60]. This is an alternative approach to stabilize the structural of delithated Li1-xMO2 and improve the electrochemical