No abstract

With lower potential than oxides, lithium transition-metal nitrides are promising anode materials for Li-ion battery. Two types of structure can be found in the family of lithium transition-metal nitrides: anti-fluorite type 3D structures such as Li7MnN4, Li3FeN2, Li7VN4; and Li3N-type layered structure with a general formula Li3-xyMxN (M = Co, Ni, Cu; y = oxidation state of metal). The synthesis and electrochemical properties of the Li3-xyNixN system have been recently investigated in our group and Li2.0Ni0.67N was identified as the most promising compound, with the highest Ni content and the highest capacity in the system (200 mAhg−1 in the 1.25 V - 0.02V potential window). Here we report the first kinetic study to characterize the charge transfer and lithium diffusion properties in Li2.0Ni0.67N system during charge-discharge cycle, by using Electrochemical Impedance Spectroscopy (EIS).Li2.0Ni0.67N exhibits excellent capacity retention upon cycling: almost 100% capacity retention at C/10 and more than 88% capacity retention at 1C over at least 100 cycles. Such remarkable cycling properties can be explained partially by its “zero-strain” fully reversible structural change upon cycles, with a maximum volume variation of 2% . The structural changes of Li2.0Ni0.67N during charge-discharge cycle involve a supercell in the hexagonal family, which leads to two different space groups depending on the depth of discharge. When the material is charged (Li1.67+xNi0.67N, 0≤ x ≤ 0.25), vacancies in the structure induce a symmetry lowering and a P-62m space group must be used to describe the structure. This supercell is 3 times larger than the P6/mmm unit cell adopted to describe the discharged material (Li1.67+xNi0.67N, 0.25 ≤ x ≤ 0.5).A kinetic study was carried out using electrochemical impedance spectroscopy during charge-discharge cycle. The simulation of equivalent electric circuit was carried out, which allows to deconvolute the different contributions in the impedance spectra. The calculated lithium diffusion coefficient, in the order of 10− 9 cm2S− 1, does not vary with the lithium content. Interestingly, the charge transfer resistance (Rtc) is found to depend on the depth of discharge. Indeed, a value close to 0 is observed during the first half-discharge (Li1.67+xNi0.67N, 0≤ x ≤ 0.25, space group P-62m), then Rtc gradually increases along the second half-discharge (Li1.67+xNi0.67N, 0.25 ≤ x ≤ 0.5, space group P6/mmm). Such behavior might be related to cell volume variation, which was found to be zero up to mid-discharge while increasing in the second part of discharge. A fully reversible behavior is observed for Rtc during the charge-discharge cycle. The overall impedance of the cell remains stable upon long cycling, indicating a good chemical stability of the electrochemical surface area and of SEI formed during the first reduction cycle.

No abstract