6/7Li NMR study of the Li1-zNi1+zO2 phases

Effect of Cobalt Substitution on the Activation of the LiNiO2 Discharge Reaction

Cobalt (Co) has been introduced to most of the practical Ni-rich layered positive electrode materials owing to its ability to stabilize the layered structure and lessen cation-mixing. However, it has been unclear whether a highly ordered structure is essential or Co addition itself has some effects. In this study, we synthesized Co-substituted LiNiO2 (LNO) with and without the introduction of cation-mixing to investigate the detailed effects of Co on crystal/local structures and electrochemical properties. It was found that the charge-discharge reversibility of LNO was enhanced by Co substitution with an additional discharge capacity at around 3.5 V, showing that the reaction at the end of discharge was activated. This behavior was observed in Co-substituted LNO even if cation-mixing was largely introduced, implying the intrinsic effect of Co on reversibility. Solid-state NMR results showed that the local structure in LNO with cation-mixing significantly changed after charge-discharge, whereas that of Co-substituted LNO hardly changed even when cation-mixing was introduced, which seems to be responsible for better reversibility. Density functional theory calculation also supports the positive effect of Co on lithium transportation at the end of discharge.

Probing Jahn-Teller Distortions and Antisite Defects in LiNiO2 with 7Li NMR Spectroscopy and Density Functional Theory

The long- and local-range structure and electronic properties of the high-voltage lithium-ion cathode material for Li-ion batteries, LiNiO2, remain widely debated, as are the degradation phenomena at high states of delithiation, limiting the more widespread use of this material. In particular, the local structural environment and the role of Jahn-Teller distortions are unclear, as are the interplay of distortions and point defects and their influence on cycling behavior. Here, we use ex situ7Li NMR measurements in combination with density functional theory (DFT) calculations to examine Jahn-Teller distortions and antisite defects in LiNiO2. We calculate the 7Li Fermi contact shifts for the Jahn-Teller distorted and undistorted structures, the experimental 7Li room-temperature spectrum being ascribed to an appropriately weighted time average of the rapidly fluctuating structure comprising collinear, zigzag, and undistorted domains. The 7Li NMR spectra are sensitive to the nature and distribution of antisite defects, and in combination with DFT calculations of different configurations, we show that the 7Li resonance at approximately -87 ppm is characteristic of a subset of Li-Ni antisite defects, and more specifically, a Li+ ion in the Ni layer that does not have an associated Ni ion in the Li layer in its 2nd cation coordination shell. Via ex situ7Li MAS NMR, X-ray diffraction, and electrochemical experiments, we identify the 7Li spectral signatures of the different crystallographic phases on delithiation. The results imply fast Li-ion dynamics in the monoclinic phase and indicate that the hexagonal H3 phase near the end of charge is largely devoid of Li.

Further Improving Coulombic Efficiency and Discharge Capacity in LiNiO2 Material by Activating Sluggish ∼3.5 V Discharge Reaction

The electrochemical activity of LiNiO₂ at the initial cycle and factors affecting its activity were understood. Even though LiNiO₂ can achieve almost theoretical charge capacity, it cannot deliver the theoretical discharge capacity that would result in low 1st Coulombic efficiency (CE). For different upper cut-off voltages at 4.3 and 4.1 V, the 1st CE barely increases. Given that the H2-H3 phase transition occurs at ∼4.2 V, the low 1st CE is not caused by this phase transition but is a result of the additional 3.5 V discharge reaction, which is kinetically limited and thereby not activated even at a reasonable current density. We found out that the several phase transitions during charge/discharge in LiNiO₂ barely affect the 3.5 V reaction. Under galvanostatic intermittent titration technique (GITT) conditions, LiNiO₂ can achieve ∼250 mAh/g of discharge capacity and 100% CE even with the 4.3 V cut-off voltage by fully activating the 3.5 V reaction. Using neutron diffraction and ⁶Li nuclear magnetic resonance (NMR) measurements, the sluggish kinetics of the 3.5 V reaction can be ascribed to difficult insertion of Li at the end of the discharge because this reaction can be accompanied by the rearrangement of cations or local structure change in the structure. To achieve high discharge capacity in LiNiO₂ with the 4.3 V cut-off voltage, this 3.5 V sluggish reaction should be improved. The finding and understanding underlying the mechanism of the electrochemical activity will stimulate further research on high-capacity Ni-rich layered materials for high-performance Li-ion batteries.

Direct Observation of Lattice Aluminum Environments in Li Ion Cathodes LiNi1-y-zCoyAlzO2 and Al-Doped LiNixMnyCozO2 via (27)Al MAS NMR Spectroscopy

Direct observations of local lattice aluminum environments have been a major challenge for aluminum-bearing Li ion battery materials, such as LiNi1-y-zCoyAlzO2 (NCA) and aluminum-doped LiNixMnyCozO2 (NMC). (27)Al magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectroscopy is the only structural probe currently available that can qualitatively and quantitatively characterize lattice and nonlattice (i.e., surface, coatings, segregation, secondary phase etc.) aluminum coordination and provide information that helps discern its effect in the lattice. In the present study, we use NMR to gain new insights into transition metal (TM)-O-Al coordination and evolution of lattice aluminum sites upon cycling. With the aid of first-principles DFT calculations, we show direct evidence of lattice Al sites, nonpreferential Ni/Co-O-Al ordering in NCA, and the lack of bulk lattice aluminum in aluminum-"doped" NMC. Aluminum coordination of the paramagnetic (lattice) and diamagnetic (nonlattice) nature is investigated for Al-doped NMC and NCA. For the latter, the evolution of the lattice site(s) upon cycling is also studied. A clear reordering of lattice aluminum environments due to nickel migration is observed in NCA upon extended cycling.

NMR study of the LiMnPO(4)·OH and MPO(4)·H(2)O (M = Mn, V) homeotypic phases and DFT calculations

Following our previous work on the tavorite-like LiFePO(4)·OH and FePO(4)·H(2)O phases, we report here the magnetic and NMR characterizations of analogous LiMnPO(4)·OH, MnPO(4)·H(2)O and VPO(4)·H(2)O phases together with the DFT calculations of the NMR shifts. The first two compounds exhibit Curie-Weiss type magnetic behavior with Curie constants close to the theoretical ones for HS Mn(3+), while the vanadium compound is very close to a pure Curie-type behavior. (7)Li, (31)P and (1)H MAS NMR spectra are reported for the three compounds, and show strong Fermi-contact shifts for the first two nuclei, while the sign and magnitude of the (1)H shifts are very different for the three phases. DFT calculations (FLAPW in GGA+U approximation) using the WIEN2k code and the experimental susceptibilities are shown to reproduce closely the experimental data. This situation is compared to the case of the homologous and isostructural Fe compounds, which exhibit much more complex magnetic behaviors.

On Hahn-echo measurements of short 29Si T2 times in some silica-based materials

The Hahn-echo 29Si NMR experiments performed for determinations of T2 times in static supermicroporous materials SiO2-Al2O3-MnO and SiO2-MnO have shown the large loss in the echo intensity, observed at shortest echo delays. The relaxation patterns "intensity-tau", where the echo-intensity initially increases with increasing tau values and then it "normally" reduces due to spin-spin relaxation, are formally treated and discussed.

Solid-state NMR spectra of paramagnetic silica-based materials: observation of 29Si and 27Al nuclei in the first coordination spheres of manganese ions

Some silica-based solids, prepared by the sol/gel method in the presence of high Mn2+ concentrations, have been characterized by the 29Si, 27Al MAS NMR spectra and 29Si T1 measurements. The single-pulse 29Si and 27Al MAS NMR spectra have shown broad spinning sideband patterns that are interpreted in terms of anisotropic bulky magnetic susceptibility (BMS) and dipole-field effects. In the absence of paramagnetic isotropic shifts, the 29Si and 27Al nuclei observed in the single-pulse NMR spectra have been assigned to nuclei remote from paramagnetic centers. It has been demonstrated that the 29Si and 27Al nuclei, which are in the vicinity of the manganese ions, can be detected by the Hahn-echo MAS NMR experiments at different carrier frequencies.

Coupled Ion/Electron Hopping in LixNiO2: A 7Li NMR Study

Deintercalated "Li(x)NiO2" materials (x = 0.25, 0.33, 0.50, 0.58, and 0.65) were obtained using the electrochemical route from the Li0.985Ni1.015O2 and Li0.993Ni1.007O2 compounds. Refinements of X-ray diffraction data using the Rietveld method show a good agreement with the phase diagram of the Li(x)NiO2 system studied earlier in this laboratory. Electronic conductivity measurements show a thermally activated electron-hopping process for the deintercalated Li0.5NiO2 phase. In the Li(x)NiO2 materials investigated (x = 0.25, 0.33, 0.50, and 0.58), 7Li NMR shows mobility effects leading to an exchanged signal at room temperature. A clear tendency for Li to be surrounded mainly by Ni3+ ions with the 180 degree configuration is observed, particularly, for strongly deintercalated materials with smaller Li+ and Ni3+ contents, even upon heating, when this mobility becomes very fast in the NMR time scale. This suggests that Li/vacancy hopping does occur on the NMR time scale but that Ni3+/Ni4+ hopping does not occur independently. The position of Li seems to govern the oxidation state of the Ni around it at any time; the electrons follow the Li ions to satisfy local electroneutrality and minimal energy configuration. The observed NMR shifts are compatible with the Li/vacancy and Ni3+/Ni4+ ordering patterns calculated by Arroyo y de Dompablo et al. for x = 0.25 and x = 0.50, but not for x = 0.33 and x = 0.58.