Computational Study for Insight into Li_xNiO₂ Interlayer Collapse and Pillar Dopant to Counter Oxygen Charge-Transfer at High Depth of Charge (x < 0.25)

As demand for high power and energy capacity has continued to rise, increasing nickel content in Li[Ni_1-y-zCo_yMn_z]O₂ (NCM) or Li[Ni_1-y-zCo_yAl_z]O₂ (NCA) cathode materials has been the subject of extensive study to improve the specific capacity of Li-ion batteries. Commercialized (Ni composition ~60%) NCM/NCA can deliver specific capacity near 170 mAh/g while Li_xNiO₂ (LNO) has been shown to offer over 200 mAh/g, but coincidently suffers from instability, and anisotropic volume strain due to Ni(III)-O₆ Jahn-Teller (JT) distorted shearing and especially lattice collapse during charging. The resulting crack formation in primary particles and pulverization of secondary particles increases surface area exposed to the liquid electrolyte and accelerates phase decomposition due to oxygen evolution and side reactions, with capacity fading increasingly well studied and characterized, including the known chemical instability of Ni—O system yields oxygen redox and evolution at LNO surfaces near 4.2 V. In potential solid-state battery applications, the mechanical nature of large anisotropic strain will incur broken contact with solid electrolyte and shorten lifespan. As the most critical lattice collapse occurs in charging beyond 75% de-lithiation (x < 0.25), mitigation by implementing cutoff voltage near 4.2 V (vs Li anode) has commonly been implemented, consequently restricting practical energy density. Detailed analysis of electronic charge and charge density distribution via first-principles calculations reveals oxygen-nickel charge transfer during charge/discharge is well correlated to oxygen-cation coordination. While charging, each Ni layer shrinks monotonically throughout de-lithiation, from 2.17 Å to 1.9 Å, and the Li interslab thickness first increases, from 2.55 Å to 2.82 Å at 75% de-lithiation, then decreases to 2.48 Å. The former is consistent with the perspective of Ni(III) [3d⁷: t_2g⁶e_g¹] oxidation to smaller Ni(IV) [3d⁶: t_2g⁶e_g⁰], collaborated by respective magnetic moments, a/b lattice constant decrease, and Ni-O bond length changes associated with the loss of JT distortion. The interslab behavior, however, requires recognition of Ni-O bonding character of e_g states, where modeling the interslab expansion and collapse in finer detail confirms approximately linear behavior in each regime, notably collapsing even with low but finite Li (e.g. 20-8%) in each interlayer (no empty layers or stacking order transitions). By Bader charge analysis, it is found that Ni exhibits less redox activity directly; rather oxygen charge is more significantly depleted due to charge transfer as e_g bonding states are de-occupied by electrons during charge, indicated in projected density of states. Smaller and substantially charge-depleted oxygens appear around 25% Li content and were all identified as oxygens lacking any Li nearest neighbors. Intuitively, as concentrations progress below 25% (> 4.2 V), the number of such oxygen rapidly increases with a corresponding dwindling of higher charge state oxygens local to the few remaining Li. The interslab (Li-layer) thickness decrease follows from these smaller oxygen ins and significantly lessened repulsion across the Li layer in local close-packed, O3 stacking sequence. To mitigate the collapse, high valent cations as dopants at Li-site are investigated in similar detail with this mechanism in mind. Potential cations were chosen on a basis of compatibility with Li-site octahedra, with radii similar to Li⁺, and less comparable to Ni(II), Ni(III), or Ni(IV). Models of LNO doped as Li_x-yM_yNiO₂ are found to donate charge to all six neighboring oxygen and the respective NiO₆ octahedra throughout de-lithiation (decreasing x, 1 > x > y). The charge-compensated oxygens retain stronger local coulomb interaction near the cation dopant and provide a pillaring effect to reduce lattice collapse even to 100% depth of discharge. Select dopants indicate low anisotropy and volumetric strain, e.g. ∆V/V ~ 2% over the range of 0-90% capacity. The presented role of cation dopants reinforces the promise of "zero-strain" LNO cathode material, with simultaneous improvement to thermal and mechanical stability, encouragingly reported by experimental works, and both greater capacity and power by unlocking deeper charging.