Is oxygen redox as easy as 3d8, 3d8L, 3d8L2?

The 3d transition metal layered oxides derived from LiCoO2 are critical components of our lithium-ion battery technology due to their high energy density and stability. Simply put, the Co and Ni ions are expected to oxidize to formal 4+ configurations when all the lithium is extracted upon charging and reduce upon discharge. In real Li-ion batteries, the layered oxides are never fully delithiated during cycling because of degradation (oxygen loss) issues at high states of charge, i.e., above 2/3 of lithium extraction. As a result, the ionic approximation has proven “Goodenough” (pun intended), where the oxygen ions are considered inert O2- ions, and all the redox can be solely linked to electron counting the metal d orbitals (i.e. d8, d7, d6).



Lithium excess compounds have demonstrated capacities that cannot be accounted for by metal oxidation alone, indicating that oxygen is redox active. This prompts a reconsideration of “conventional” layered oxide charge compensation.[1] Often, oxygen redox has been reported in terms of O-O dimer formation and is considered distinct from conventional redox mechanisms in stoichiometric layered cathodes. By considering Ni-rich layered oxides as negative charge transfer insulators, I will demonstrate studies that confirm how oxygen participation in charge compensation naturally arises from ligand holes (i.e., L, charge transfer from O2p orbitals) without necessarily requiring O2 formation.[2,3] As a result, Ni2+/4+ redox is better reflected as d8 to d8L2. The implications are discussed in terms of gas evolution and design principles for next generation cathodes.

[1] Nature Reviews Materials 7 (2022), 522-540

[2] Joule 7 (2023), 1623-1640

[3] Joule (in press 2025) https://doi.org/10.1016/j.joule.2024.10.007