Fully Exploited Oxygen Redox Reaction by the Inter‐Diffused Cations in Co‐Free Li‐Rich Materials for High Performance Li‐Ion Batteries

Abstract To meet the growing demand for global electrical energy storage, high‐energy‐density electrode materials are required for Li‐ion batteries. To overcome the limit of the theoretical energy density in conventional electrode materials based solely on the transition metal redox reaction, the oxygen redox reaction in electrode materials has become an essential component because it can further increase the energy density by providing additional available electrons. However, the increase in the contribution of the oxygen redox reaction in a material is still limited due to the lack of understanding its controlled parameters. Here, it is first proposed that Li‐transition metals (TMs) inter‐diffusion between the phases in Li‐rich materials can be a key parameter for controlling the oxygen redox reaction in Li‐rich materials. The resulting Li‐rich materials can achieve fully exploited oxygen redox reaction and thereby can deliver the highest reversible capacity leading to the highest energy density, ≈1100 Wh kg−1 among Co‐free Li‐rich materials. The strategy of controlling Li/transition metals (TMs) inter‐diffusion between the phases in Li‐rich materials will provide feasible way for further achieving high‐energy‐density electrode materials via enhancing the oxygen redox reaction for high‐performance Li‐ion batteries.


I. Preparation of materials and their properties
The No-PBM + Q sample that has the quenching process but does not have any PBM process shows the separation of peaks of R phase (such as LiNi 0.5 Mn 0.5 O 2 ) and M phase (Li 2 MnO 3 ) in XRD patterns in Figure S1a. This indicates less interaction between the two phases like the NIC sample (Figure 1a), and thereby shows much poor electrochemical activity than the IC sample that have both the quenching and PBM process in Figure S1b. The No-PBM + Q sample can show poor oxygen redox activity partly due to insufficient interaction between Li 2 MnO 3 and LiNi 0.5 Mn 0.5 O 2 at high temperature even after a quenching process. The result indicates that quenching process by itself is not sufficient for achieving desired Li/TMs interdiffusion between the two phases. Intimate mixing process via additional high energy Planetary Ball Milling (PBM) process accompanied with the quenching process is necessary for obtaining desired Li/TMs inter-diffusion in a composite material.
In addition, the PBM-only sample that has only PBM process without any heating and cooling processes in Figure S1c also shows the separation of peaks of R phase and M phase like the NIC sample ( Figure 1a). This indicates that the PBM-only sample has less interaction between the two phases, especially Li/TM inter-diffusion between two phases than the IC sample. Even if the PBM-only sample has much smaller particle size ( Figure S1d) than the IC sample (Figure 1d), the PBM-only sample shows much poor electrochemical activity, especially oxygen redox activity above 4.5V, than the IC sample ( Figure S1e). This indicates that the particle size is not limiting factor for the oxygen redox activity and the Li/TM interdiffusion between the two phases is not achieved by only PBM process without heating process, which does not have any further slow cooling and quenching process. Thus, both the quenching process and PBM process are needed to achieve certain Li/TMs inter-diffusion between the phases in a composite in Li-rich materials. It should be emphasized that the quenching process accompanied with the PBM process is one of ways to achieve certain Li/TMs inter-diffusion between the phases in a composite.
-Rietveld Refinement details for the pristine materials Rietveld refinements were performed using a two-step strategy. [1] 1. To determine the amount of Ni that had diffused from LNMO phase to Li 2 MnO 3 phase, the occupancy of each 3a and 3b site in LNMO phase was constrained regardless of atom species (Li or Ni). Refinement was applied to the first step to determine 1.1 The amount of deficient Ni compared to total 0.5 mol Ni in LNMO, which will diffuse to Li 2 MnO 3 phase, and 1.2 The amount of Ni/Li exchange between Li layers and TM layers in the R-3m space group, in which only one crystallographic site exists for each layer.
2. The occupancy sites of diffused Ni from the LNMO phase into the Li 2 MnO 3 phase and the vacancy sites of Li that can be caused by Li diffusion from the Li 2 MnO 3 phase into the LNMO phase sites were refined concurrently.
Refinement results obtained using this strategy were used as a starting crystal structure for neutron powder-diffraction measurements, which can provide more detail information of the structure such as Li occupancy. Even if synchrotron x-ray diffraction refined with this strategy, it is hard to get lot of information than neutron powder diffraction because X-Ray interacts with electron gave a positive intensity but the heavy atom (such as Ni, Mn not Li) dominates the intensity due to its high number of electron unlike neutron. [2] The amount of Ni that had diffused into Li 2 MnO 3 phase and the degree of Li/Ni cation disordering for residual Ni in LNMO phase after Li-Ni inter-diffusion between the two layered phases were also refined from neutron diffraction pattern. In the LNMO (LiNi 0.5 Mn 0.5 O 2 ) phase, the IC sample showed not large difference of Li/Ni cation-disordering (~ 0.050 mol) to that of the NIC sample (~ 0.057 mol), which is well-known behavior in the LNMO phase during synthesis. [3] In contrast, in the IC sample, ~ 0.14 mol of Ni had diffused into the Li 2 MnO 3 phase from the LNMO phase with increasing cation disordering (Ni occupied the Li layer), whereas in the NIC sample there was no Ni inter-diffusion into the Li 2 MnO 3 phase without creating any cation disordering in Li 2 MnO 3 phase; these observations indicate that only the LNMO phase in the NIC sample has the cation disordering. In the IC sample, the location of the excess Li in the LNMO phase cannot be specified using these refinements, but   Table   S2, 3).  layers. The additional broad resonance at 800 -900 ppm, which is related to Li in the tetrahedral sites (interstitial site) of the spinel structure at ca. 800 -900 ppm. [6] In those environments, Li is surrounded by both Mn and Ni, in agreement with the lower hyperfine shifts measured compared to their M phase counterparts. [6a] The larger linewidth for Li local environments in R phase than in M phase can be ascribed to a distribution of Li local environments that corresponds to Li with different nearby Ni/Mn ratios and excess Li from octahedral site to tetrahedral site (interstitial site) (Table S4).  Cooled) and ZFC (Zero Field Cooled) curves, which is a direct measurement of the extent of AFM ordering. [8] In Figure S6, the IC sample shows much lower ordering transition temperature (T N ) and lower strength of FC and ZFC curve bifurcation (i.e., a narrower bifurcation) than the NIC sample. These results indicate that the magnetic ordering from  The metal-oxygen and metal-metal distances can be estimated from EXAFS spectra to understand the changes in the oxidation state of each transition metal and in their local environment. [9] Figure S7 shows metal-first oxygen neighbor distances and metal-first metal neighbor distances obtained from the refinement of k 3 -weighted Fourier transforms of EXAFS spectra with the absolute error of ±0.01 (Table S5)  -DFT calculations of formation energies of the IC samples that have the Li-Ni interdiffusion between the two layered phases DFT total energy was computed using the Vienna ab initio Simulation Package (VASP) [12] and the PBE functional in the GGA [13] and PAW method. [14] The on-site interaction is corrected by the Hubbard U parameters of 6.0 eV for the Ni and 3.9 eV for the Mn 3d orbitals. [15] Spin-polarized calculations were performed with a plane-wave energy cutoff = 520 eV and the k-points grid density > 0.01 k-points/Å 3 . The lattice parameters and atomic positions were optimized until the energies converged to < 10 5 eV/atom and forces and Li 2−2y Ni y MnO 3 phases separately using the layered and flower structures of LNMO, [16] and Li 2 MnO 3 in the C2/m space group. We used supercells that contained 8 formula units

II. Anion redox reaction in both R phase and M phase of the IC sample with full Li
extraction.

Ernzerhof (HSE) hybrid functional
In the Li-Ni inter-diffused system, the Li 2−2x Ni x MnO 3 region gains an additional redox center upon insertion of Ni 2+ , because in theory it can be oxidized up to Ni 4+ , with consequent increase in the contribution of this phase to the net capacity of the system. In contrast, the functional [19] and applied it to the most stable structures predicted in Figure 2c.  other Li-rich materials exhibiting the highest first discharge capacity and average potential. [21] All electrochemical tests of IC sample and NIC sample were performed at room temperature and the operating voltage window was from 2.5 V to 4.7 V. Figure S15. Voltage profiles at C/3 charge and discharge rate during cycles for the NIC and IC samples. Electrochemical tests were performed at room temperature and the operating voltage window was from 2.5 V to 4.7 V. Average potential is defined as the potential obtained at half discharge capacity or at half energy density during discharge.    (7) Atom occupancies (Li     (7) 6.0(12)