Li-ion batteries Materials

Transition metal oxide

LiCoO2 (LCO) introduced by Goodenough  is the first and the most commercially successful form of layered transition metal oxide cathodes. It was originally commercialized by SONY, and this material is still used in the majority of commercial Li-ion batteries. The Co and Li, located in octahedral sites occupy alternating layers and form a hexagonal symmetry (Fig. 4a). LCO is a very attractive cathode material because of its relatively high theoretical specific capacity of 274 mAh g−1, high theoretical volumetric capacity of 1363 mAh cm−3, low self-discharge, high discharge voltage, and good cycling performance .

The major limitations are high cost, low thermal stability, and fast capacity fade at high current rates or during deep cycling. LCO cathodes are expensive because of the high cost of Co (Fig. 1). Low thermal stability refers to exothermic release of oxygen when a lithium metal oxide cathode is heated above a certain point, resulting in a runaway reaction in which the cell can burst into flames . Thermal runaway is a major concern in the application of Li-ion batteries, resulting, for example, in the grounding of all Boeing 787 airplanes in 2013 . While this issue is general to transition metal oxide intercalation cathodes, LCO has the lowest thermal stability of any commercial cathode material . Although thermal stability is also largely dependent on non-materials factors such as cell design and cell size, LCO typically experiences thermal runaway past ∼200°C due to an exothermic reaction between the released oxygen and organic materials. Deep cycling (delithiation above 4.2 V, meaning approximately 50% or more Li extraction) induces lattice distortion from hexagonal to monoclinic symmetry and this change deteriorates cycling performance . Many different types of metals (Mn, Al, Fe, Cr)  were studied as dopants/partial substitutes for Co and demonstrated promising but limited performance. Coatings of various metal oxides (Al2O3, B2O3, TiO2, ZrO2,  were more effective in enhancing LCO stability and performance characteristics even during deep cycling, because mechanically and chemically stable oxide material could reduce structural change of LCO and side reactions with electrolyte.

LiNiO2 (LNO) has same crystal structure with LiCoO2 and a similar theoretical specific capacity of 275 mAh g−1. Its relatively high energy density and lower cost compared to Co based materials are the main research driving forces. However, pure LNO cathodes are not favorable because the Ni2+ ions have a tendency to substitute Li+ sites during synthesis and delithiation, blocking the Li diffusion pathways . LNO is also even more thermally unstable than LCO because Ni3+ is more readily reduced than Co3+ [60]. Partial substitution of Ni with Co was found to be an effective way to reduce cationic disorder . Insufficient thermal stability at high state-of-charge (SOC) can be improved via Mg doping , and adding a small amount of Al can improve both thermal stability and electrochemical performance .

As a result, the LiNi0.8Co0.15Al0.05O2 (NCA) cathode has found relatively widespread commercial use, for example, in Panasonic batteries for Tesla EVs. NCA has high usable discharge capacity (∼200 mAh g−1) and long storage calendar life compared to conventional Co-based oxide cathode. However it was reported that capacity fade may be severe at elevated temperature (40–70°C) due to solid electrolyte interface (SEI) growth and micro-crack growth at grain boundaries .

LiMnO2 (LMO) can also be promising because Mn is much cheaper and less toxic compare to Co or Ni. Anhydrous and stoichiometric layered LMO was prepared almost two decades ago , improving on a previous aqueous methods which induced impurities, different stoichiometries, poor crystallinity, and undesirable structure change during cycling . However, the cycling performance of LMO was still not satisfactory (i) because the layered structure has a tendency to change into spinel structure during Li ion extraction  and (ii) because Mn leaches out of LMO during cycling . Mn dissolution occurs when Mn3+ ions undergo a disproportionation reaction to form Mn2+ and Mn4+, and this process is observed for all cathodes containing Mn. Mn2+ is thought to be soluble in the electrolyte, and destabilize the anode SEI. Indeed, Mn concentration in the electrolyte and anode SEI has been observed to increase with aging for Mn containing cathodes . Also, the anode impedance is seen to increase with Mn dissolution on carbon anodes , but not LTO (which has a negligible SEI). Stabilization of LMO via cationic doping was conducted both experimentally and theoretically , but even so, the poor cycle stability of LMO (especially at elevated temperatures) has hindered widespread commercialization.

Continuous research efforts on developing cathode material less expensive than LCO resulted in the formulation of the Li(Ni0.5Mn0.5)O2 (NMO) cathode. NMO could be an attractive material because it can maintain similar energy density to LCO, while reducing cost by using lower cost transition metals. The presence of Ni allows higher Li extraction capacity to be achieved. However, cation mixing can cause low Li diffusivity and may result in unappealing rate capability . Recent ab initio computational modeling predicted that low valence transition metal cations (Ni2+) provides high-rate pathways and low strain, which are the crucial factors to achieve high rate capability in layered cathodes. NMO recently synthesized by ion exchange method showed a very low concentration of defects in NMO and capacity as high as ∼180 mAh g−1 even at a very high rate of 6 C .

Polyanion compounds

In exploring new cathode materials, researchers have developed a new class of compounds called polyanions. Large (XO4)3− (X = S, P, Si, As, Mo, W) polyanions occupy lattice positions and increase cathode redox potential while also stabilizing its structure . LiFePO4 (LFP) is the representative material for the olivine structure, known for its thermal stability and high power capability. In LFP, Li+ and Fe2+ occupy octahedral sites, while P is located in tetrahedral sites in a slightly distorted hexagonal close-packed (HCP) oxygen array (Fig. 4c). The major weaknesses of the LiFePO4 cathode include its relatively low average potential (Fig. 4e, Table 1) and low electrical and ionic conductivity. Intensive research over the last decade resulted in significant improvements in both performance and mechanistic understanding of LFP. Reduction in particle size in combination with carbon coating and cationic doping  were found to be effective in increasing the rate performance. It is noteworthy that good electrochemical performance can also be achieved without carbon coating if particles are uniformly nano-sized and conductive nanocarbons are used within the cathodes . Virus-templated amorphous anhydrous FP/CNT composite, for example, demonstrated promising results . It was reported that the facile redox reaction in non-conducting LFP could be due to a curved one-dimensional lithium diffusion path through the [0 1 0] direction . In general, however, the low density of nanostructured LFP electrodes and their low average potential limit the energy density of LFP cells. Recently, a novel non-olivine allaudite LFP was reported and showed fundamentally different electrochemical behavior from that of olivine LFP .

Other olivine structures include LiMnPO4 (LMP) which offers ∼0.4 V higher average voltage compared to olivine LFP (Table 1), leading to higher specific energy, but at the expense of lower conductivity . LiCoPO4, LiNi0.5Co0.5PO4, and LiMn0.33Fe0.33Co0.33PO4 (LCP, NCP, MFCP) have also been developed and shown promising results, but further improvements in power, stability and energy density are required . Novel Li3V2(PO4)3 (LVP) exhibited relatively high operating voltage (4.0 V) and good capacity (197 mAh/g) . Quite remarkably, LVP/C nanocomposite exhibited 95% theoretical capacity at a high rate of 5 C in spite of the low electronic conductivity of LVP (similar with LFP).

LiFeSO4F (LFSF) is yet another interesting cathode material because of its high cell voltage and reasonable specific capacity (151 mAh g−1) . Fortunately LiFeSO4F has better ionic/electronic conductivity hence it does not desperately need carbon coating and/or nanoparticles. LiFeSO4F can also be economical since it can be prepared with abundant resources. LiFeSO4F is composed of two slightly distorted Fe2+O4F2 oxyfluoride octahedra connected by F vertices in the trans position, forming chains along the c-axis, and the Li+ are located along the (1 0 0), (0 1 0), and (1 0 1) directions (Fig. 4d). Tavorite-structured cathode materials were evaluated via simulation and reported that the fluorosulfate and fluorophosphate families of materials are the most promising, and the oxysulfate family is the least . The tavorite structured materials with 1D diffusion channels were suggested to exhibit low activation energies, allowing charge and discharge of Fe(SO4)F and V(PO4)F at very high rates, comparable to those observed in small olivine Fe(PO4) particles. The vanadium-containing material, LiVPO4F, cycles well, and has high voltage and capacity  but raise concerns about toxicity and environmental impact. Interestingly, Li+ can be intercalated at ∼1.8 V hence this material is able to be used in both anode (Li1+xVPO4 where x = 0–1) and cathode (Li1−xVPO4 where x = 0–1). For further detailed information on synthesis method, chemical properties and mechanism, much more specialized reviews are available elsewhere .

Fluorine and chlorine compounds

Sulfur and lithium sulfide

Sulfur has an extremely high theoretical capacity at 1675 mAh g−1, while also being low cost and abundant in the Earth's crust. However, S based cathodes suffer from low potential vs. Li/Li+, low electrical conductivity, dissolution of intermediate reaction products (polysulfides) in electrolyte, and (in the case of pure S) very low vaporization temperature, which induces S loss while drying the electrodes under vacuum. Sulfur also suffers from ∼80% volume change , which may destroy the electrical contact in standard carbon composite electrodes . To mitigate the effects of both dissolution and volume expansion, S can be encapsulated in a hollow structure with excess internal void space. Polyvinyl pyrrolidone polymer , carbon , and TiO2  capsules have been impregnated with sulfur by using infiltration and chemical precipitation. When tested in half cells in thin electrode configurations, these composites show cycle life sometimes approaching 1000 cycles.

To avoid the negative effects of expansion, prevent S evaporation during drying, and form full cells with Li free (and thus safer) anodes, electrodes have also been fabricated in the form of Li2S . Li2S is not easily infiltrated into a host as with S because it has a much higher melting point. However, the high solubility of Li2S in various environmentally friendly solvents (such as ethanol) can be utilized to form various Li2S based nanocomposites such as, for example, Li2S nanoparticles embedded within a conductive carbon . Because the fully lithiated Li2S does not expand any further, no void spaces are necessary. In fact, carbon-coated Li2S showed no change in morphology after 400 charge/discharge cycles .

Electrolyte modification is a popular method for mitigating polysulfide dissolution (Fig. 3f). LiNO3  and P2S5  additives were used to form good SEI on the surface of Li metal to prevent the reduction and consequent precipitation of polysulfides. Lithium polysulfides can also be added to temporarily decrease cathode dissolution . Multiple papers also utilized higher molarity electrolytes, which also greatly reduces polysulfide solubility . Finally, solid state electrolytes can also prevent polysulfide dissolution, and at the same time, enhance cell safety by avoiding Li dendrite short circuiting .

Selenium and tellurium

Recently, Se and Te have attracted attention due to their higher electronic conductivities than S and high theoretical volumetric capacities of 1630 mAh cm−3 and 1280 mAh cm−3, respectively, in the fully lithiated state. Due to the higher electronic conductivity, Se and Te often show higher utilization of active materials and higher rate capability than S. Similar to S, the Se-based cathodes suffer from the dissolution of high-order polyselenides , resulting in fast capacity loss, poor cycle performance and low coulombic efficiency. So far, the dissolution of polytelluride has not been reported. As seen in Table 2, elemental Se and Te also suffer from large volume change. Fortunately, Se and Te are also similar to S in that they have low melting points. Both materials have been infiltrated into various porous carbon hosts , and dispersed or wrapped in conductive matrices to improve their performance. However, Te is far too expensive for practical use. Moreover, Se and Te are of similar abundance as Ag and Au (Fig. 1), and are very unlikely to be used in mass production.

Iodine

The lithium-iodine primary battery uses LiI as a solid electrolyte (10−9 S cm−1), resulting in low self-discharge rate and high energy density, and is an important power source for implantable cardiac pacemaker applications. The cathodic I is first reduced into the tri-iodide ion (I3−) and then into the iodide ion (I−) during discharge . For use in most other applications, this chemistry is problematic, however, because of its low power capability. Furthermore, in standard organic electrolytes, iodine, triodide, and lithium iodide are all soluble . Due to the high solubility of LiI in organic solvents, iodine ions have been considered for use in lithium-flow batteries instead. Recently, active iodine was infiltrated into the pores of porous carbon due to low melting point of I (113°C). The as-produced iodine–conductive carbon black composite showed a high discharge voltage plateau, good cycle performance, and high rate capability, which is attributed to the enhanced electronic conductivity and suppressed active material dissolution .