Prospects for lithium-ion batteries and beyond—a 2030 vision

This is an area with massive ongoing global fundamental and applied research effort. A strong focus is on mitigating degradation, to increase longevity (and indirectly cost), and because degradation becomes more severe as the voltages are increased, and, for example, more Ni and Si are added to the cathode and anode, respectively. It is also hoped that learning from these studies can be generalised and applied to the next generation of battery chemistries. These studies are aided by the impressive development of new experimental and theoretical tools and methodologies, including operando measurements that can study batteries that are closer to the practical device, with improved temporal and spatial resolution and increased sensitivity. In the case of NMR spectroscopy, one area that the authors focus on, dynamic nuclear polarisation (DNP) methods, involving the transfer of magnetisation from unpaired electrons to nuclear spins, has been used to enhance the signal of the SEI, or more recently to examine the Li metal–SEI interface1. Moving forward, the DNP method is likely to play an increasingly important role in examining the buried interfaces ubiquitous in batteries. We now discuss some specific challenges in more detail.

Cathodes

Figure 1 summarises current and future strategies to increase cell lifetime in batteries involving high-nickel layered cathode materials. As these positive electrode materials are pushed to ever-higher voltages and nickel contents, increased rates of electrolyte oxidation and surface rock-salt layer (RSL) growth become increasingly problematic for maintaining practical cell lifetimes, RSL formation generally leading to impedance rise2,3. RSL formation and the concomitant loss of oxygen have been proposed to be the primary driver of electrolyte oxidation at high voltages, rather than Faradaic currents—affecting materials from LiNiO2 through to LCO4,5. Yet many fundamental questions remain. What chemical factors determine the rate of oxygen diffusion and RSL growth? Why (and when) is singlet oxygen observed and how does it form? Are electrolyte components oxidised at the electrode surface or in the solution? Higher nickel content is also associated with larger anisotropic volume changes during cycling—representing a source of intra- and inter-granular cracking—and ‘fatigued’ phases with lower practical capacity.

Core-shell and gradient materials utilise more stable compositions (often lower Ni-content) near the electrode surface to minimise electrode-electrolyte reactivity and a nickel-rich core stoichiometry to increase energy density. Electrolyte additives are compounds added to the electrolyte solution on the order of a few weight per cent to improve cell lifetime and safety, for example by reacting with the electrode surface to form a protective ‘barrier’ layer. Surface coatings (applied via a variety of methods) on the electrode material can improve cycling stability and lifetime by scavenging corrosive HF, physical blockage of electrolyte components from reaching the electrode surface, slowing RSL growth by blocking oxygen loss from the active material, and via other chemical reactions with the electrolyte components. Heat treatments of surface-coated materials can be used to prepare surface-doped materials with improved chemical stability and that inhibit the growth of surface rock-salt layers. One trend in particle morphology research is to increase primary particle sizes (i.e., transition from polycrystalline to ‘single crystal’ materials), while future prospects include the synthesis of finely tuned particle shapes and sizes. (TEM of RSL adapted from Lin et al.14).

Full size image

While the timeline to establish answers is uncertain, these and other basic questions will almost certainly be increasingly studied and debated in the coming years. New understanding will allow for more strategic development of methods to mitigate degradation pathways (Fig. 1). Core-shell particles could be prepared with optimised gradients of different transitional metal and s/p-block metals, and layer thicknesses with stable surfaces and higher energy density cores—following on from a number of pioneering studies6; surface coating stoichiometries and doping elements could be chosen to lower the rate of oxygen loss and RSL formation; surface-modifying electrolyte additives could be designed to inhibit singlet oxygen evolution and to slow electrolyte oxidation. The development of detailed micromechanical models will guide particle morphology optimisation—size and shape—for various materials and applications. However, all of these possible advances hinge on the ability of the field to connect fundamental concepts with the complex multi-process behaviour of modern LIBs and ultimately to demonstrate that this leads to longer lifetime. For this, increased fundamental understanding, obtained via careful experimental and theoretical studies, is required.

Anodes

An ‘obvious’ win involves replacing graphite with either silicon or silicon oxide, due to their fivefold–tenfold higher energy densities. However, this is not straightforward: SiOx causes considerable first cycle irreversibly capacity loss associated with the formation of inorganics such as Li2O and Li4SiO47. A stable SEI does not form on silicon, in part because of the large volume expansion that is a direct consequence of its large capacity. While its first cycle irreversible capacity loss is lower, it is currently difficult to achieve high enough coulombic efficiencies for applications needing >300–500 cycles. Many current commercial cells include small amounts of SiOx (2–10%) into graphite anodes, providing modest capacity gains. Polymer and graphene (carbon) coatings (and mesostructures/shells) coupled with different electrolyte additives have all been proposed to increase coulombic efficiencies and enable the use of higher Si contents. Alternatively, limiting the range over which the silicon is lithiated minimises the volume expansion, leading to a more stable SEI. Graphite–Si composites bring with them other challenges including the mechanical grinding of graphite caused by the Si expansion/contraction. Calendaring graphite to increase its practical volumetric energy density will result in more mechanical grinding. While Si will play a role in future battery technologies, a question remains as to the extent and the degree to which the longevity of cells and safety will win out over increased energy density. The answers will vary across sectors, Si mostly likely playing a larger role in batteries where lifetime and safety are less critical.

Electrolytes and other cell components

To increase the volume fraction occupied by active electrode materials—again reducing cost—current collectors and polymer separators have become much thinner over the years. Higher loadings can also be achieved by increasing the active layer thicknesses, decreasing the binder fraction, and decreasing the porosity. All of these require increased electrolyte (ionic) transport to maintain rate capability, an area of active research already for fast-charging battery technologies8. The transport properties and molecular-scale structures of new solution chemistries (e.g., new solvent systems, highly concentrated salts) are becoming increasingly understood9,10. Basic studies—both experiments and calculations—of the physicochemical properties of new electrolyte compositions are expected to continue leading to new materials and insight into their properties. Beyond this, the structure and stability of the SEI in various solutions and conditions (temperature, voltage) must be better characterised. Such insights will feed development of optimised additive/coatings for enabling alternative electrolytes, while maintaining cell lifetimes. Intensive benchmarking and lifetime analysis of these systems remains a present and future need. Finally, their cost and safety of handling will need to be proven before wide or large-scale adoption is possible, the latter representing an important but underrepresented area of study.