For the 0.4Li2TiO3·0.6LiTi0.5Ni0.5O2 sample, the study identified a redox mechanism consisting of the sequential Ni2+ → Ni3+/4+ and oxygen oxidation upon charging, but surprisingly, the subsequent sequential order on discharge Ni3+/4+ → Ni2+ and oxygen reduction.

First author: Biao Li 3

full text quick overview

reversible anionic oxidation Reduction reaction can create advanced high-energy-density cathode materials for lithium-ion batteries . The activation mechanism for these reactions is often related to the ligand-metal charge transfer (LMCT) process. This process has not yet been fully experimentally validated due to a lack of suitable model materials. This paper shows that the activation of anionic redox in cationic disordered rock salt Li1.17Ti0.58Ni0.25O2 involves a long-lived intermediate Ni3+/4+ species, which can be completely converted to Ni2+ during the relaxation process. Combining electrochemical analysis and spectroscopic techniques, this study quantitatively determined that the reduction of this Ni3+/4+ species undergoes a kinetic LMCT process (Ni3+/4+–O2− → Ni2+–On−). The findings provide experimental validation of previous theoretical hypotheses and help rationalize several properties related to the anion redox, such as cation-anion redox inversion and voltage hysteresis. This work also provides additional guidance for designing high-capacity electrodes by screening suitable cationic species to regulate LMCT.

Background Introduction

Lithium-ion batteries provide the best performance for the electrification and digitalization of today's society, but their energy density still needs to be improved to overcome shortcomings, such as "range anxiety" when driving electric vehicles . This requires a positive electrode (cathode) that can carry more lithium ions per unit mass. Now popular conversion methods rely on lithium-rich materials (Li1+xM1−xO2, 0 x 1, where M is a transition metal), which relies on cationic and anionic redox processes to provide high capacity. However, the extra capacity for anionic redox is associated with slow kinetics and voltage hysteresis, posing great challenges for practical applications. Therefore, it is necessary to further study how the anionic redox is activated to clarify and solve these problems.

In order to achieve anionic redox, thermodynamics band structure considerations assume that the anion p lone pair state (fixed below the Fermi level) is required. However, this situation fails in cases such as Li2TiO3 or Li2TiS3, where the anionic redox cannot be activated by the anionic p state (located on top of the Ti4+ empty d band). However, these electrochemically inert compounds can be activated by doping with cations such as Fe2+/Fe3+, Co2+ and Ti3+, or with anions such as Se2-. Both strategies use cationic redox simultaneously. A similar situation is encountered with most lithium-rich oxides/sulfides based on d0-M, so the question has been raised whether cationic redox is essential for activating anionic redox.

Researchers proposed this necessity about nine years ago, when it was envisaged that anionic redox occurs through a "reductive coupled " mechanism, i.e. initial peroxidation of M followed by reduction, while the anion is partially oxidized, such as Ru6+–O2− → Ru5+ –O− (in Li2Ru0.75Sn0.25O3). A similar two-step activation (Ir5.5+–O2− → Ir5.5+–O(2−n)−) was proposed in Li2Ir0.75Sn0.25O3, which is called ligand -metal charge transfer (LMCT).The well-known Li1.2Ni0.13Mn0.54Co0.13O2 (lithium-rich NMC) compound also has this mechanism, where a Mn7+ intermediate species is theoretically predicted and can be spontaneously reduced through an "O to Mn" charge transfer process. Regardless of the name, all explanations point to a LMCT-like process that has never been directly observed, possibly because of the short lifetime of these cationic intermediates. Although the "Fe4+" intermediate was recently captured in Li1.17Ti0.33Fe0.5O2 by Mössbauer spectroscopy, the transient excited Fe4+–O2- state could not be detected by diffraction and spectroscopic X-ray techniques. Therefore, LMCT remains an unverified hypothesis.

Graphic analysis

Figure 1. xLi2TiO3·(1 − x)LiTi0.5Ni0.5O2 (0 Structure and electrochemical properties of x 1). a, X-ray diffraction pattern. b, Galvanostatic charge-discharge curves (first two cycles) at a current density of 20 mA g-1 and a voltage range of 2.0-4.8 V. For Li2TiO3, a low current density of 5 mA g-1 was used. c, Rietveld refinement results of synchrotron X-ray diffraction at x = 0.4 (0.4LTO–0.6LTNO or Li1.17Ti0.58Ni0.25O2). d, [001] HAADF-STEM image of the pristine 0.4LTO–0.6LTNO sample, with the corresponding [001] electron diffraction pattern and crystal structure inset. The white arrows in the electron diffraction pattern indicate the diffuse intensity originating from short-range ordering.

Figure 2. The redox mechanism of 0.4LTO–0.6LTNO was studied by ex-situ XAS and DFT calculations. a, Cycle curve for 0.4LTO-0.6LTNO, marking specific points used for ex situ XAS studies. The points are numbered sequentially and correspond to the spectra in b and c. Annotated colored arrows highlight the redox processes inferred from the XAS results. b, Ni L-edge iPFY-XAS results. Due to spin-orbit splitting, the spectrum has two distinct regions corresponding to the L3 and L2 edges respectively; each region is further split into two peaks through the octahedral ligand field effect (Ni 2p → Ni 3d–t 2g and the electron transition of 3d–eg). c, O K-edge TFY-XAS results. The front edge peak (525–536 eV) corresponds to the electronic transition from O (1s) to the unoccupied O (2p)–Ni/Ti (3d) hybrid state. d, Calculated DOS during delithiation process for 0.4LTO–0.6LTNO within the DFT+U framework. The Fermi level is marked by a vertical dashed line. e, Charge transfer of nickel and oxygen determined from calculated Bard charges during delithiation of 0.4LTO-0.6LTNO.

Figure 3. In situ X-ray diffraction pattern of 0.4LTO–0.6LTNO. a, First two electrochemical cycles between 4.8 and 2 V. b, Changes in a lattice parameters extracted from in situ X-ray diffraction patterns. c, In situ X-ray diffraction patterns of the first two cycles. d, function relationship between lattice parameter changes and lithium content . A large hysteresis is highlighted between changes in lattice parameters for charge and discharge. e, In situ X-ray diffraction pattern collected on the electrode; the electrode was rapidly recovered from the liquid cell charged to 4.8 V, followed by washing and drying. Arrows indicate the direction of peak shift. f, In situ X-ray diffraction on electrodes recovered from ASSB to eliminate self-discharge effects due to liquid electrolyte. The top colored bar (black arrow) indicates significant peak shift.

Figure 4. Quantification of the LMCT process by electrochemical titration. a, Discharge curves of samples relaxed at different temperatures for 24 hours. The sample was recovered from a batch of powder charged to 4.5 V and then heated in aliquots. The discharge curves of samples dried without heating were used for benchmark testing.Inset: Histogram showing the capacity of the nickel redox plateau (4.5-2.5 V) and oxygen redox plateau (2.5-1.85 V) extracted from the discharge curve. b, The corresponding dQ/dV curve shows the capacity rearrangement between nickel and oxygen redox. Note that the oxygen reduction peaks are slightly aligned with the same position, allowing for easy observation of peak intensity changes. Arrows indicate the direction of peak evolution. c, d, Discharge curve (c) of the sample relaxed at 140°C for different times and the corresponding dQ/dV curve (d). The arrow in d indicates the direction of peak evolution. Sample preparation and cell assembly were performed in the same manner as in a.

Figure 5. Characterization of LMCT by HAXPES (with hν = 10 keV photon energy), and its structural analysis. a, b, Ni 2p (a) and O 1s (b) spectra of samples charged to 4.5 V and relaxed under heating at 120 °C for different times. c, Ni 2p spectrum of the original 0.4LTO–0.6LTNO reference sample showing pure Ni2+ signal. d, O 1s spectrum of a Li-rich NMC reference sample charged to 4.8 V. Deconvolution of the O 1s spectrum gives three distinct species: (1) lattice O2− at ~529.5 eV (grey); (2) oxidized lattice oxygen species On− (0 n 2), ca. 530.5 eV (red); (3) Surface oxygen-containing species, such as carbonate (white), Sat., satellite peaks related to Ni 2p peak (a) and lattice O2- peak (b). e,f, [001] HAADF-STEM images of the relaxation of 0.4LTO–0.6LTNO before (e) and (f) after (24 h at 140 °C) charging to 4.5 V, and their corresponding [001] electrons Diffraction pattern.

Figure 6. Explain redox inversion and voltage hysteresis. a,b, Second cycle of 0.4LTO–0.6LTNO compound with different cutoff voltages (a) and corresponding dQ/dV curves (b). The Ni-O redox inversion is highlighted in b. c, d, Second cycle of Li-rich NMC compound with different cutoff voltages (c) and corresponding dQ/dV curve (d). Like the 0.4LTO–0.6LTNO compound, the cation–anion redox inversion behavior is highlighted in d. e, Discharge curves of samples relaxed to different states. Samples were recovered from the same batch of powder charged to 4.5 V and then heated at 140 °C for different times (from 0 s to 24 h). Cyclic discharge curves after charging to 4.5 V are used for comparison. Inset: OCV change as a function of rest time before discharge. The orange arrow indicates that the lower the amount of nickel intermediate (inferred from the nickel reduction plateau), the lower the OCV. f, Ratio of OCV to nickel intermediate fitted by least squares method.

Summary and Outlook

Based on the above results, this paper shows that in the cationic disordered xLi2TiO3·(1−x)LiTi0.5Ni0.5O2 system , for direct experimental validation of LMCT during anionic redox activation. For the 0.4Li2TiO3·0.6LiTi0.5Ni0.5O2 (Li1.17Ti0.58Ni0.25O2) sample, the study identified a redox mechanism consisting of sequential Ni2+ → Ni3+/4+ and oxygen oxidation upon charging, but surprisingly What is interesting is that the sequential Ni3+/4+ → Ni2+ and oxygen reduction will occur during subsequent discharge. This Ni3+/4+ is a kinetically activated long-lived intermediate rather than a thermodynamically stable substance as demonstrated experimentally and theoretically. Through in situ powder X-ray diffraction, electrochemical analysis, hard Highly dependent on temperature and time.This cationic intermediate was finally rationally explained to be responsible for some anionic redox electrode properties, including cation-anion redox inversion, voltage and structural hysteresis.