Breaking Free from Cobalt Reliance in Lithium-Ion Batteries

Summary

The exponential growth in demand for electric vehicles (EVs) necessitates increasing supplies of low-cost and high-performance lithium-ion batteries (LIBs). Naturally, the ramp-up in LIB production raises concerns over raw material availability, where constraints can generate severe price spikes and bring the momentum and optimism of the EV market to a halt. Particularly, the reliance of cobalt in the cathode is concerning owing to its high cost, scarcity, and centralized and volatile supply chain structure. However, compositions suitable for EV applications that demonstrate high energy density and lifetime are all reliant on cobalt to some degree. In this work, we assess the necessity and feasibility of developing and commercializing cobalt-free cathode materials for LIBs. Promising cobalt-free compositions and critical areas of research are highlighted, which provide new insight into the role and contribution of cobalt.

Graphical Abstract

Energy Storage; Energy Materials

Introduction

The global demand for lithium-ion batteries (LIBs) is no longer solely based on portable electronics but primarily driven by the electrification of the transportation industry. The increase in market share for battery and plug-in hybrid electric vehicles (EVs) over the next decade is expected to reach 20% with more than 150 million cumulative EV sales (Figure 1) (International Energy Agency, 2019). The exponential EV sales growth is driven by government incentives, such as subsidies, rebates, tax breaks, and implementation of charging stations. In addition, expanded EV driving ranges and lowered costs due to steady improvements in LIB technology are critical factors. As such, the utilization of LIBs has grown 30% per year from 2010, reaching 180 GWh in 2018 and is anticipated to reach 2,600 GWh by 2030 (World economic forum, 2019). Global expansion of LIB production capacity is projected to balloon from ∼290 GWh in 2018 to ∼1,700 GWh in 2028 (Figure 1) (Benchmark Mineral Intelligence, 2019). The EV-LIB market response is bidirectional—battery packs account for up to half the cost of EVs, and in order for EVs to compete with internal combustion engine (ICE) vehicles, key performance metrics must be improved including the cost-to-range ratio, safety, and power capability (Cano et al., 2018). This has galvanized efforts among research communities to develop higher-energy-density (Wh kg⁻¹) batteries at a lower cost.

Figure 1.

Historical and Projected Global Plug-in EV Sales and LIB Production Capacity

Projected EV sales based on a 21% CAGR (Benchmark Mineral Intelligence, 2019; International Energy Agency, 2019).

Cobalt Supply and Demand

LIB technology for consumer EVs is currently dominated by two main cathode compositions, layered LiNi_1-x-yCo_xAl_yO₂, x + y < 0.2 (NCA), and LiNi_xMn_yCo_zO₂, x + y + z = 1 (NMC) (Grand View Research, 2017). Owing to the degree of commercialization and production of these materials worldwide, it is expected that they will remain the primary cathode chemistries for LIBs over the next decade. However, large-scale LIB production raises concerns over resource availability. It is estimated that 50%–80% of the production cost of LIBs is associated with materials, with up to 50% of this ascribed to the cathode active material (AVICENNE Energy, 2017; Berckmans et al., 2017). As LIB manufacturing costs have substantially decreased owing to economies of scale, transitions toward larger cell formats, and maturity in battery pack production techniques, the material input cost will become more prevalent (Few et al., 2018; Nykvist and Nilsson, 2015). Constraints in the supply of raw materials can lead to severe price fluctuations and make LIBs unreasonable for large-scale applications, bringing the momentum and optimism of the EV market to a halt.

Among the raw resources required for LIB production, concerns have been raised over the supply chain of lithium and cobalt, which is closely linked with battery production. Although the exact quantity of recoverable global lithium reserves is difficult to determine, most projections concur that they are sufficient to meet long-term projected demands (up to 2100) (Gruber et al., 2011; Narins, 2017). However, there are concerns over the uneven geographic distribution of the reserves and whether production can meet high demands for lithium by 2050, which can cause price spikes (Speirs et al., 2014; Vikström et al., 2013). On the other hand, reports have indicated that deficits in the cobalt supply could occur as early as 2030 (Alves Dias et al., 2018). As seen in Figures 2A and 2B, cobalt is by far the most valuable metal used in LIBs. In 2010, ∼25% of all cobalt produced was used in secondary batteries (LIBs and minor quantity in Ni-MH batteries), which grew to 30% in 2017 and is expected to expand to 53% by 2025 (Azevedo et al., 2018). Moreover, cobalt continues to be an important component in catalysts, integrated circuits, semiconductors, magnetic recording devices, and various high-strength alloys. However, cobalt is scarce and expensive to process, as it is mostly derived from low-concentration by-products of nickel and copper mining. Cobalt is considered a critical resource as ∼60% of the worldwide mine production in 2018 originated from copper-cobalt ores in the Democratic Republic of the Congo (DRC), where geopolitical instability and unethical working conditions are well documented and can lead to halting of cobalt exports (Schulz et al., 2017; Tsurukawa et al., 2011). This was apparent in 1978 where civil conflict generated a drastic price spike known as the Cobalt Crisis (Figure 2A).

Figure 2.

Metal Supply for LIB Components

(A) Historical inflation-adjusted commodity price of cobalt.

(B) Commodity price of other common metals used in LIBs.

(C) Trade flow of raw and processed cobalt. Image adapted with permission from Olivetti et al. (2017).

Moreover, China has dominance over the cobalt supply chain as the world largest producer, supplier, and consumer (Figure 2C). China has heavily invested and acquired foreign cobalt mining operations primarily in the DRC since 2000, which has reduced their net import reliance of raw cobalt from 97% to 68% (Gulley et al., 2019). Considering that China itself may experience cobalt supply deficits by 2030 unless efficient recycling targets are achieved (Zeng and Li, 2015), China's cobalt production will likely be prioritized for domestic battery manufacturers. Continued global reliance on cobalt may lead to competition for raw materials and ensuing conflicts similar to the rare earth metals trade dispute in 2010, which will urge investigations into cobalt-free energy storage technologies (Overland, 2019).

Current Cobalt-free Commercial Cathodes

The issue with cobalt resource scarcity was acknowledged early on by LIB pioneers, motivating research into abundant and sustainable cathode chemistries. Goodenough's research group first reported the olivine LiFePO₄ (LFP) cathode in 1997, and the path toward commercialization was paved after the development of a carbon-coated nanoparticle morphology to address its poor intrinsic electronic and ionic conductivity (Li et al., 2018b). LFP is attractive owing to its high thermal stability associated with the covalent phosphate moieties, excellent cycle life, flat charge/discharge profile, and high electrochemical stability over ∼100% depth of discharge (DOD). However, it has achieved little market traction in Western markets owing to its low energy density and nominal voltage (∼3.3 V versus graphite anode), which directly affects the driving range of EVs. However, LFP technology generates significant interest in China owing to its cobalt-free composition and has been developed and adopted by major EV manufacturers, such as the BYD Company and the Wanxiang Group Corporation. Most electric buses (∼99% of the global stock concentrated in China) utilize LFP batteries (International Energy Agency, 2019). Thus, many speculate that LFP will play an important role in public transportation and stationary energy storage, where safety and stability are more critical than the energy density (Zeng et al., 2019). An area of improvement of LFP is lowering the production cost associated with the complex synthesis methods to favor the cost/performance ratio. The electrochemical performance of LFP is highly sensitive to the preparation method, requiring rigorous control of morphology, particle size distribution, coating homogeneity, and reagent purity while ensuring that Fe²⁺ is not oxidized to Fe³⁺ for performance consistencies (Jugović and Uskoković, 2009; Yuan et al., 2011).

Similarly, spinel LiMn₂O₄ (LMO) is a commercially relevant cobalt-free cathode that was first reported by Goodenough's group in 1983 (Thackeray et al., 1983). The host structure enables 3D solid-state diffusion of Li⁺, resulting in high LIB rate performances. However, its main setback is the low practical capacity and cycle stability caused by the presence of Mn³⁺. The electronic configuration of Mn³⁺ (t_2g³-e_g¹) induces Jahn-Teller distortion, which can cause lattice changes from the cubic to tetragonal phase and constrict Li⁺ diffusion (Yamada, 1996). The distortion is more pronounced when discharging at high rates as Li⁺ is more concentrated at the surface of the LMO particles, which pronounces the distortion, resulting in particle cracking and exfoliation (Li et al., 2009). Moreover, Mn³⁺ generates soluble Mn²⁺ based on the disproportionation reaction: 2Mn³⁺(solid) → Mn⁴⁺(solid) + Mn²⁺(solution). Suppression of Jahn-Teller distortion has commonly been addressed by partially substituting Mn with other cations to reduce the amount Mn³⁺(Capsoni et al., 2003; Kim and Lee, 2007; Li et al., 2009), while various coatings have been explored to suppress the dissolution of Mn (Guan et al., 2011). Ultimately, the poor reliability of LMO limits it toward niche applications.

Layered Mixed-type Cathodes

On the other hand, layered mixed-type transition metal oxide cathodes are the most suitable to reach the high-energy requirements for EVs. Among the NMC-type compositions, LiNi_1/3Mn_1/3Co_1/3O₂ (NMC111) is the most established and known for its stability and safety (Belharouak et al., 2003; Noh et al., 2013). In this solid solution, Mn⁴⁺ is electrochemically inactive and assists with thermal and electrochemical stability, whereas Co³⁺ contributes to electronic conductivity and suppresses cation mixing between Ni²⁺ and Li⁺ (Myung et al., 2017; Zeng et al., 2018). Capacity is primarily dependent on the Ni²⁺/Ni⁴⁺ redox couple with contribution from Co³⁺/Co⁴⁺ at higher voltages. Motivated by the higher capacities (at 4.3 V versus Li/Li⁺ upper cut-off) and lower cobalt content, there is ongoing commercial development to substitute NMC111 (∼160 mAh g⁻¹, 20.4 wt % Co) with LiNi_0.6Mn_0.2Co_0.2O₂ (NMC622, ∼180 mAh g⁻¹, 12.2 wt % Co) and LiNi_0.8Mn_0.1Co_0.1O₂ (NMC811, ∼200 mAh g⁻¹, 6.1 wt % Co) (Myung et al., 2017). The increase in nickel content raises the capacity at the expense of cycle and thermal stability (Noh et al., 2013). With higher nickel content, the degree of (de)lithiation increases, resulting in anisotropic lattice volume changes that contribute to particle cracking (Li et al., 2019b). Nickel-rich cathodes also suffer from increased moisture sensitivity, Li⁺/Ni²⁺ cation mixing, electrolyte side reactions (solvent oxidation from Ni⁴⁺), and phase transitions toward spinel and rock-salt structures along with oxygen evolution at elevated temperatures and voltages (Myung et al., 2017). A promising approach to address these issues is to synthesize NMC as micron-scale single crystal particles. Compared with polycrystalline NMC that comprises 10- to 15-μm-size agglomerates of nanoparticles, the single-crystal morphology has a minimal number of grain boundaries and thus can mitigate particle cracking and side reactions with the electrolyte (Kim, 2012; Li et al., 2017). A more popular approach in the literature involves synthesizing core-shell NMC, where the core is Ni-rich and enhances capacity, whereas the shell is Mn-rich and can alleviate volume expansion and protect the core from electrolyte reactions (Sun et al., 2006, 2009, 2010, 2012). Similarly, LiNi_0.8Co_0.15Al_0.05O₂ (NCA) is a “low-cobalt” (9.2 wt % Co) cathode composition known to display a similar energy density to NMC811, although with better cycle retention but poorer thermal stability (Xia et al., 2018).

These developments in high-energy cathodes with lower cobalt content raises the question as to whether they are enough to sustain long-term and large-scale LIB applications. Olivetti et al. projected cobalt demand assuming that the LIBs implemented in EVs comprise 50% NMC111, 35% NMC622, and 15% NMC811 (Olivetti et al., 2017). Their conservative projected demand was based on a 36% compound annual growth rate (CAGR) of EV sales and assuming an average battery pack size of 75 kWh, whereas the aggressive projection assumed 10 million EV sales (10% of all vehicles) in 2025. These scenarios correspond to 136 and 336 kt of cobalt demand. On the other hand, the projected expansions in cobalt supply was 180 and 290 kt for the conservative and aggressive projections, respectively. Similarly, the International Energy Agency projected a cobalt demand of 170 ± 60 kt in 2030 based on currently announced policy ambitions, assuming that the LIB chemistry then comprises 10% NCA, 40% NMC622, and 50% NMC811 (International Energy Agency, 2019). These findings indicate that, even while implementing low-cobalt cathodes in EVs, cobalt supply strains could occur in the near future if aggressive EV targets are met. It is also worth mentioning that high-nickel LIBs will need to be replaced more frequently owing to their lower cycle life. Furthermore, this projection does not account for sudden constraints in supply due to political issues and other risk factors as discussed earlier.

LIB metal recycling can help address the resource constraints. However, owing to the continuous expansions in battery production (Figure 1), production is expected to substantially outweigh the amount of end-of-life batteries entering the waste stream over the foreseeable future. Furthermore, LIB recycling activities are concentrated in China, whereas infrastructure elsewhere consists of select private facilities focused on recovering high-value cobalt (Wang et al., 2014). Although the recycling of end-of-life LiCoO₂ batteries in portable electronics is highly feasible and profitable, there are significant technical challenges and decreased financial motivation in recycling mixed-metal LIBs in EVs that will dominate the waste stream over the next decades (Or et al., 2020). Furthermore, the dismantling of cells from EV battery packs is currently not economically feasible. Although the concept of repurposing end-of-life battery packs (∼80% capacity retained) toward stationary energy storage has been proposed, it is not well developed (Ahmadi et al., 2014). Worldwide realization of LIB recycling will require legislation and political pressure, likely in the form of economic incentives (e.g., refundable deposits with LIB purchases), public education, landfill disposal regulations, and defined responsibilities on the collection and disposal of LIBs for consumers, retailers, and EV and battery manufacturers. Taken together, it is evident that cobalt-free cathodes are required for sustainable long-term applications of LIBs.

Cobalt-free Layered Cathodes

Improving nickel-rich layered LiNi_1-xM_xO₂ (M = Mn, Al, and/or Co)-type materials is one of the most promising approaches to developing cathodes for LIBs with reduced reliance on cobalt and has attracted a considerable amount of research effort over the past several years (Aishova et al., 2020; Cheng et al., 2019; Cormier et al., 2019; Li et al., 2019a; Xu et al., 2017; Zhang et al., 2019). Nickel-rich cathodes such as NMC811 and NCA are a step forward to developing completely cobalt-free LiNiO₂-based materials (LNO) capable of delivering high energy density in line with performance targets set for future EVs. However, thermal instability (Dahn et al., 1994; Guilmard et al., 2003a, 2003b) and structural decomposition leading to poor electrochemical performance (Bianchini et al., 2020; Croy et al., 2019) are primary barriers inhibiting further applications of cobalt-free layered cathodes over existing materials in modern LIBs. Intermixing of Ni²⁺/Li⁺ cations during cycling is one of the main degradation pathways, causing localized formations of an inactive rock-salt phase that worsen during subsequent cycles. As seen in NMC and NCA-type compositions, substitution of nickel for cobalt is thought to be an effective method of stabilizing Ni²⁺ ions within the transition metal layer, thereby reducing the degree of cation mixing (Croy et al., 2019). However, recent work has called into question whether cobalt is necessary or if a similar effect could be achieved using other metal substituents (e.g. Al, Mn, or Mg) (Li et al., 2019a). Capacity degradation in layered transition metal oxides also occurs through the formation of microcracks in secondary particles owing to uneven volumetric expansion within the lattice (Makimura et al., 2012). The propagation of microcracks in secondary particles increases the available surface area for the formation of a passivated solid-electrolyte interface (SEI), causing an increased consumption of lithium that leads to irreversible capacity reduction (Watanabe et al., 2014). Additionally, intergranular cracking causes a loss of contact between the grains of secondary particles, giving rise to an increased impedance in the battery. The deleterious phase transition from hexagonal 2 (H2) → hexagonal 3 (H3) with low reversibility is well known to occur at high degrees of delithiation within Ni-rich layered cathodes and is thought to be a major contributor to particle cracking due to sharp volumetric contractions of the unit-cell (c-axis) (Li et al., 2015, 2018a; Weber et al., 2017). High-temperature operation (≥50°C) of the material can enhance the rate performance for cobalt-free compositions; however, operation above 30°C can accelerate structural degradation and thus should not be relied upon (Aishova et al., 2020; Ma et al., 2018; Sun et al., 2015). These issues need to be resolved in order to drive nickel-rich and cobalt-free layered oxide cathodes toward commercialization in next-generation EVs.

Recent work from Li et al. has cast uncertainty in previously theorized contributions of cobalt substitution in nickel-rich layered cathodes by investigating various compositions of LiNi_1-xM_xO₂ (M = Al, Co, Mg, or Mn; x = 0.05 or 0.1) (Li et al., 2019a). Similar amounts of Ni²⁺ in the Li⁺ layer (Ni_Li) were found for both LiNi_0.90Co_0.05Al_0.05O₂ and LNO substituted with 5 mol% Al, Mg, or Co (Ni_Li ∼1%), indicating that Co does not provide a significant reduction of cation mixing compared with Al or Mg (Figure 3C). The presence of Mg caused a preferential positioning of Li atoms in vertex-sharing sites within the layers above and below the substituent. This caused two Li atoms to become “locked in” for every one Mg atom, reducing the interlayer repulsion that is associated with detrimental phase changes in the cathode. However, this also reduced the reversible charge capacity, as two Li atoms became inactive for every Mg atom in the structure. Subsequently, LiNi_0.95Mg_0.05O₂ delivered an initial charge capacity of 205 mAh g⁻¹ (∼10% loss) between 3.0 and 4.3 V versus Li⁺/Li at 10 mA g⁻¹, compared with 219 mAh g⁻¹ and 223 mAh g⁻¹ for LiNi_0.95Mn_0.05O₂ and LiNi_0.95Al_0.05O₂, respectively. This report further indicated that there was no noticeable improvement from LiNi_0.95Co_0.05O₂ in structural stability throughout cycling, as minimal variation in unit-cell volume as a function of Li content (x in Li_1-xMO₂) was observed in all substituted samples tested (Figures 3A and 3B). The initial discharge capacity (IDC) and cycle retention of LiNi_0.95Al_0.05O₂ was similar to traditional NCA, with ∼95% of the initial capacity retained after 50 cycles at 10 mA g⁻¹, whereas LiNi_0.95Mg_0.05O₂ displayed the best cycle retention of 97% after 50 cycles albeit with lower IDC, implying that cobalt is not required for Ni-rich materials to attain acceptable long-term capacity retention. However, the rate performance of these compositions was not assessed. In addition, cobalt substitution did not improve thermal stability over conventional LNO with self-heating rates (SHRs) > 20°C min⁻¹ at 160°C, whereas Al and Mg demonstrated noticeable benefit by keeping SHR <1°C min⁻¹ over the range of 120°C–240°C (Figure 3E). This report challenged the necessity of cobalt in layered oxides and highlighted the comparable performance in cathodes with other metal dopants.

Figure 3.

Unit Cell Dimensions

(A and B) (A) a-axis and (B) c-axis as a function of substituent content.

(C) Percentage of Ni found within the Li layer (Ni_Li) as a function of substituent content.

(D) Irreversible capacity [mAh g⁻¹] as a function of substituent content.

(E) Self-heating rate (SHR) for de-lithiated LiNiO₂ and LiNi_0.95M_0.05O₂ (M = Al, Co, Mg, Mn) charged to 230 mAh g⁻¹ as a function of temperature between 120°C and 250°C. Images adapted with permission from Li et al. (2019a).

The impact of various substituents on both Co-free Ni- and Mn-rich compounds has been studied over the last several years, where cation doping with Mg²⁺, Al³⁺, Fe³⁺, Na⁺, etc. has enabled advancements of the overall electrochemical performance and safety of these compounds (Cheng et al., 2019; Dong et al., 2013; Guilmard et al., 2003c; Mohan and Kalaignan, 2013; Wang et al., 2013). In addition to Li et al., critical progress has been made understanding the relationship among the cathode dopant, thermal stability, and reactivity with the electrolyte. Cormier et al. confirmed that thermal stability is improved in substituted LiNi_xM_1-xO₂ (M = Mn, Mg, Al, or Co; x = 0 or 0.05) (Cormier et al., 2019). The substituent can enable the cathode to maintain a greater Li content at the end of charge (∼4.4 V versus Li⁺/Li), which reduces the propensity for oxygen evolution and stabilizes the structure. The stabilized oxides display substantially less aggressive reactions with the electrolyte, causing the SHR to occur at much higher temperatures (220°C) than LiNi_0.95Co_0.05O₂/LNO (160°C).

In another study, Aishova et al. investigated the electrochemical performance of LiNi_0.9Mn_0.1O₂ (NM90), a nickel-rich derivative of the most widely studied cobalt-free composition LiNi_0.5Mn_0.5O₂ (Aishova et al., 2020). The authors confirmed that NM90 exhibits a higher degree of cation mixing (Ni_Li = 3.35%) than both cobalt-containing compounds, LiNi_0.9Mn_0.05Co_0.05O₂ (Ni_Li = 1.77%) and LiNi_0.9Co_0.1O₂ (Ni_Li = 0.67%). No difference in the IDC (236 mAh g⁻¹) was observed among the samples from 2.7–4.4 V versus Li⁺/Li at 0.1 C, and capacity retention improved as a function of Mn content, with NM90 retaining 93% after 100 cycles. Inactive Mn⁴⁺ ions are theorized to stabilize the structure upon delithiation, evidenced in an upshift of the redox potential associated with the destructive H2 → H3 phase transition. Broadening of the dQ/dV peak was also observed, indicating that the lattice volumetric change was spread over a wider voltage range, allowing relief of non-uniform strain in the lattice of NM90. Additionally, the fracture strength of secondary particles in materials with higher Mn content was increased, with 25% higher strength in NM90 (175 MPa) than LiNi_0.9Mn_0.05Co_0.05O₂ (140 MPa) and nearly double that of LiNi_0.9Co_0.1O₂ (95 MPa). Both of these mechanisms demonstrate an important role in the cycle stability of the Mn-substituted LiNiO₂.

Further improvements to the cyclability of cobalt-free layered oxides were reported by de Bosse et al. who observed that preparation of a Li-rich layered oxide material with O2-type oxide packing could suppress the H2→H3 phase transition by preventing the previously mentioned antisite exchange (de Boisse et al., 2018). A unique synthesis methodology was employed to achieve this structure, involving the ionic exchange of Na⁺ in P2-Na_0.83[Li_0.19Mn_0.73Ni_0.08]O₂ for Li⁺ in molten salt, yielding Li-rich O2-Li_1.19-yMn_0.73Ni_0.08O₂. The resulting O2-type LMNO delivered an IDC of 240 mAh g⁻¹ from 2.0–4.8 V versus Li⁺/Li at 0.05 C. However, the work also indicated an excellent capacity retention compared with the O3-type analogs, retaining >90% of the initial capacity after 50 cycles. This work suggests that structural adjustments through modifying synthesis procedures could be a key strategy to developing next-generation cobalt-free cathode materials.

Surface treatments and coatings have also been investigated as a solution to improve the stability and rate performance of Co-free cathodes. In general, their ability to act as a protective barrier between electrolyte and active material, and a scavenger to reduce the acidity of non-aqueous electrolytes can substantially restrict metal ion dissolution during cycling (Chong et al., 2016; Li et al., 2012; Wu et al., 2017; Zhang et al., 2012). One of the most promising studies recently was presented by Deng et al. who designed a modified F- B-rich SEI on LiNiO₂ with the addition of a small amount of difluoro (oxalato) borate (LiDFOB) in the electrolyte (Deng et al., 2019). The SEI formed as a robust and compact layer that mitigated the dissolution of Ni as well as the irreversible transformation to the unwanted NiO rock salt phase by protecting the surface from reacting with the by-products generated from the oxidation of electrolyte at high voltage. This material showed a high IDC of 216 mAh g⁻¹ from 2.0–4.4 V versus Li⁺/Li with an exceptional cycling performance of 94% after 100 cycles and 80% retention after 400 cycles. In comparison, the LNO cycled without the LiDFOB additive retained only 18% of the initial capacity after 100 cycles.

The design and engineering of spatially partitioned nanostructures for cobalt-free cathodes is an interesting new strategy that has seen some research activity in the past few years (Li et al., 2016b; Pan et al., 2018; Xu et al., 2019). Similar to the work conducted in Ni-rich NMC, the use of core-shell and gradient core-shell structures for cobalt-free materials was preliminarily investigated by Zhang et al. (2019). The authors prepared a LiNiO₂ core with a LiNi_1-xM_xO₂ shell (M = Al, Mg, Mn), finding that the core-shell materials delivered a high IDC of 194 mAh g⁻¹ (LiNi_0.83Mg_0.17O₂ shell) and 230 mAh g⁻¹ (LiNi_0.83Al_0.17O₂ LiNi_0.83Mn_0.17O₂ shell) from 3.0–4.35 V versus Li⁺/Li at 0.2 C. The LNO-LiNi_0.83Mg_0.17O₂ and LNO-LiNi_0.83Mn_0.17O₂ core-shell electrodes retained 94% and 92% of their initial capacity, respectively, after 55 cycles, whereas the LNO-LiNi_0.83Al_0.17O₂ electrode retained 93% of its initial capacity after 55 cycles and 69% after 400 cycles at 0.2 C. Despite the high capacities, it was found that the metal substituents diffused from the shell to the core structure during sintering, suggesting that the benefit of the core-shell structure was not fully realized. Nevertheless, the report demonstrates the promising nature of this approach and suggests room for improvement through a more optimized synthesis pathway. Wang et al. prepared Li-rich Li_1.2Mn_0.6Ni_0.2O₂ with a porous nanoflake network through a resorcinol-formaldehyde assisted sol-gel method (Wang et al., 2017). The specific surface area of the nanoflake material was 6.9 m² g⁻¹, which is a notable improvement over spherical analogs at 1.5 m² g⁻¹ (Li et al., 2016a). The high surface area facilitates the diffusion of lithium leading to an elevated rate performance, with discharge capacities of 273 mAh g⁻¹ at 0.1 C and 196 mAh g⁻¹ at 2 C from 2.0–4.8 V versus Li⁺/Li. A potential concern with this material would be the electrochemical stability due to the high surface area available for SEI formation and Li consumption; however, a stable capacity was reported with 93% retention after 150 cycles at 2 C indicating that the stability of the material was not negatively impacted by the modified morphology.

Cation Disordered Rock Salt Cathodes

Disordered rock salts are an emerging class of high-energy-density cathode materials. Disordered materials have been largely overlooked as they are associated with an electrochemically inactive phase formed within conventional layered materials during cycling. However, recent reports indicate that disordered materials can offer higher capacity than other modern LIB cathodes and suppress common failure mechanisms observed in layered oxide materials. Conventional layered materials exhibit distinct Li layers separated by the transition metal (TM) sublattice, whereas disordered materials have Li/TM intermixed in random fashion at octahedral sites within the same cubic close packed (ccp) lattice (Figures 4A and 4B). Additionally, the cation disordered lattice does not suffer from structural degradation caused by cation mixing that is observed in layered structures (Lee et al., 2014).

Figure 4.

Physical and Electrochemical Comparison of Layered Transition Metal Oxide and Cation Disordered Rock-salt Cathode Materials

(A) Scanning transmission electron microscopy (STEM) images along [010] zone axis for left: layered Li_1.211Mo_0.467Cr_0.3O₂ before cycling and right: fully disordered Li_1.211Mo_0.467Cr_0.3O₂ after 10 cycles. Images adapted with permission from Lee et al. (2014).

(B) Schematic of the general crystal lattice for left: conventional layered transition metal oxides; right: cation disordered transition metal oxides.

(C) Electrochemical performance (V versus Capacity) of cation disordered rock salt electrodes with various metal substituents (i.e., Ni, Mn, Fe, Co) at room temperature between 1.0 and 4.8 V versus Li⁺/Li at 10 mA g⁻¹ rate. Image adapted with permission from Yabuuchi et al. (2015).

Li⁺ diffusion in disordered rock-salts proceeds primarily through channels with no face-sharing TM ions (0-TM), whereas the typical pathway found in layered materials with one face-sharing TM ions (1-TM) is nearly inactive in disordered materials. The 0-TM channels in disordered materials have a low diffusion barrier, indicating facile movement of Li⁺ within the atomic structure; however, the infrequency of these channels in the disordered structure prevents the formation of a continuous percolating network and lowers the probability of macroscopic diffusion of Li⁺. Heterogeneity of the material also implies that localized cation ordering influences the Li⁺ conduction pathways, further contributing to disruptions of the continuous percolation network (Kan et al., 2018). Thus, disordered materials suffer from poor rate capability owing to these sluggish Li⁺ diffusion pathways and the presence of high-valence transition metals that cannot be oxidized and thus do not contribute to charge compensation (e.g., Nb⁵⁺).

Disordered rock salts suffer structural degradation primarily through the irreversible release of surface-level oxygen at high voltages (>4.3 V versus Li⁺/Li). The O²⁻/O₂²⁻ anions are used for charge compensation of Li⁺, enabling enrichment of Li content in the cathode. Utilization of the oxygen redox for charge compensation is critical toward achieving the high reversible capacities in disordered rock salts, which can be observed when charging above 4.3 V versus Li⁺/Li (Figure 4C). However, the release of oxygen as a reactive radical species (O⁻/O₂²⁻) initiates formation of a resistive interfacial layer through electrolyte decomposition, and structural degradation is further induced by vacancy defects from oxygen release (Cambaz et al., 2019; Luo et al., 2016; Yan et al., 2019). Currently, long-term capacity retention has not been reported for these materials, but the high theoretical energy densities have galvanized considerable research effort to alleviate these failure mechanisms.

Lee et al. investigated the partial substitution of O²⁻ for F⁻ in a Li_1.3Mn_0.4Nb_0.3O₂ disordered rock salt to facilitate the use of lower valence Mn²⁺ in the structure for higher theoretical capacity and reduced oxygen loss (Lee et al., 2018). The unsubstituted material employs the Mn³⁺/Mn⁴⁺ redox, leading to low theoretical capacity; however, F⁻ substitution reduced the valence of Mn in the structure allowing for use of the Mn²⁺/Mn⁴⁺ redox couple (270 mAh g⁻¹). As a result, the reliance on oxygen redox reactions to achieve capacities >300 mAh g⁻¹ is mitigated. The diminished use of the anionic redox minimized oxygen release, providing higher cycling stability to the cathode and lower polarization (Lee et al., 2017). Surface modification of disordered rock salts through minor doping with heavy elements (e.g., Os, Sb, Ru, Ir, Ta) was reported by Shin et al. to improve oxygen retention at the surface (Shin et al., 2018). Larger-sized dopants tend to segregate at the surface, and cations exhibiting stronger hybridization with neighboring oxygen atoms improved oxygen retention. Slight improvements to electrochemical stability were observed, with a significant reduction in irreversible oxygen loss in doped materials. In general, mitigating the loss of oxygen is seen as the key to achieving higher cycle stability in disordered rock salt materials. Despite the promising energy metrics and initial efforts to improve electrochemical stability, rate performance continues to hinder deployment of disordered rock salts in modern applications. Future research on this topic is critical to resolving this issue and pushing the material toward commercialization in next-generation LIBs.

Conversion Cathodes

Finally, conversion electrodes are an interesting next-generation technology that demonstrates ultra-high gravimetric capacity owing to their reaction with 2–3 stoichiometric equivalents of Li⁺ while comprising low-cost materials (Figure 5A). However, they possess low theoretical potentials (often <3.5 V versus Li⁺/Li). The most studied and promising material is the sulfur cathode (S/Li₂S), which has already been implemented in lightweight batteries for niche aerospace applications. Metal fluorides (MF₂, M = Fe, Mn, Cu, Co, or Ni) are also promising as they demonstrate higher theoretical potentials and can potentially compete with the energy density NMC/NCA-type cathodes when paired with a graphite anode (Wu and Yushin, 2017). However, severe drawbacks must be overcome, particularly the discrepancy between the theoretical and practical potential, poor intrinsic conductivity and diffusion kinetics, and low reversibility. One notable cause of poor reversibility is the high solubility of the cathode materials, resulting in dissolution and subsequent redeposition at the anode or cathode, which can throttle capacity and negatively influence the SEI makeup. In addition, the lithiation/delithiation cycles are associated with significant volume changes, resulting in particle fracturing, particle disconnection, and electrode swelling. There have been many studies published to mitigate these issues, particularly for the sulfur cathode, where promising performance benchmarks with respect to cycle life (<20% degradation over 1,000 cycles) and rate capability (>80% capacity retention at 1 C) have been achieved (Kang et al., 2016). However, many techniques are complex and costly, and have poor scalability. Engineering innovations are also required to optimize cell fabrication for new technologies. Furthermore, to achieve notable expansions in energy density, the cathode must be paired with a high-capacity anode, namely, silicon or lithium metal. These anode materials present similar challenges with respect to stability and safety. Taken together, it is difficult to predict when conversion cathodes will become commercially viable.

Figure 5.

Comparison of Selected Promising Co-free Cathode Compositions in the Literature

Commercial cathodes (LCO, NMC, and NCA) also included for comparison.

(A) Specific energy (Wh kg⁻¹) relative to material cost (US$ kWh⁻¹). Cost does not account for production, manufacturing, or synthesis approach. Material cost obtained from the metal commodity price in June 2020.

(B) Comparison of demonstrated cycle performance. Data used in calculations shown in Table S1.

Conclusion and Perspectives

Concerns over cobalt scarcity has been understood as an issue for several decades; however, the use of modern “low-cobalt” materials (e.g., NMC622, NMC811, NCA) is not a sufficient mitigation strategy to this issue, as supply strains could still arise as early as 2025 if aggressive EV sales targets are met. LIB recycling is a direct solution to cobalt recovery, but technical complications with battery pack disassembly and the mixed-metal components constrict its use. Currently, only China is positioned for LIB recycling and global implementation requires economic incentives, improved public education, and cooperation with LIB manufacturers. A more effective and lasting solution for the sustainable future of LIBs is the development of cobalt-free cathode materials.

Layered transition metal oxides based on LiNiO₂ have attracted significant research efforts for their high energy density. Inherent issues associated with the poor cycle stability (Figure 5B) of nickel-rich and cobalt-free layered cathodes include the volume expansion and particle cracking owing to the higher Li⁺ utilization (i.e., capacity), Ni²⁺/Li⁺ cation mixing, increased electrolyte oxidation, and lower thermal stability and rate performances. Partial substitution of Ni with Mn, Al, and Mg helps improve the thermal stability and cycle performance through the suppression of phase changes and particle cracking. Doping LiNiO₂ with Mn does not reduce the IDC, although it increases the degree of Ni²⁺/Li⁺ cation mixing. However, Al and Mg can help suppress Ni²⁺/Li⁺ mixing at the expense of the IDC. For optimal performance, the cation doping strategy can be coupled with morphology control (e.g., single crystal particles to suppress cracking), coatings that improve interparticle conductivity or provide a protective barrier against electrolyte side reactions and suppress oxygen evolution, and adjustments to the oxide packing structure to suppress phase changes. Research in this direction provides the most practical and short-term path toward the commercialization of cobalt-free cathodes. On the other hand, cation disordered rock salt cathodes are an emerging material class with high energy density (up to ∼1,000 Wh kg⁻¹). Future research for this material should be focused on improving the rate performance (e.g., through surface treatments) and minimizing oxygen loss (e.g., doping, core-shell structures) to extend cycle life. With increasing effort, the development of cobalt-free cathode materials can address the cobalt scarcity issue and sustain large-scale deployments of LIBs toward EVs and stationary energy storage.

Limitations of the Study

In Figure 5, data were obtained from referenced literature sources in half-cell configuration (versus Li/Li⁺ anode)—performance may not be representative of updated industry benchmarks. The material cost calculation was based on the metal commodity price from June 2020 and does not account for production, manufacturing, or synthesis approach.

Resource Availability

Lead Contact

Further information and requests should be directed to and will be fulfilled by the Lead Contact, Zhongwei Chen (zhwchen@uwaterloo.ca).

Materials Availability

This study did not generate any new materials.

Data and Code Availability

Any data utilized in this study can be found in the main manuscript and Supplemental Information.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.