Degradation and Safety Analysis of Large Format LiFePO4/Graphite Battery

Large format LiFePO 4 /Graphite (LFP/Gr) Lithium-ion battery (LIB) have been developed for electric vehicles (EV) and stationary use. In case of large surface electrodes, intrinsic two-phase reaction of LFP has a risk to induce inhomogeneous reaction in planar electrode surface. We applied >200 Ah class LFP/Gr LIBs and cycled them around 80-65 % SOH levels. Among them, we applied non-destructive differential voltage analysis (dV/dQ) to three cells (Cell1:SOH 80.9 %, Cell2:SOH 73.5 %, Cell3:SOH 68.4 %), and determined the degree of inhomogeneity of the electrodes by disassembly of the cell in the Ar-filled glove box. We could obtain dV/dQ peak corresponds to Gr 2 nd stage single phase by charge curve analysis between SOC 0 and 100 % at C/20. Based on the assumption that Gr capacity is twice of the Gr 2 nd stage peak capacity, Gr capacity was sufficiently larger than cell capacity. Therefore, cell capacity fade was not related to the capacity fade of Gr but caused by the shift of the cathode / anode operation window (SOW) due to loss of lithium inventory (LLI) and/or capacity fade of LFP. After dV/dQ analysis with capacity check, cells were disassembled at 2.5 V, and confirmed LFP and Gr capacity by reassembling coin cells with Li counter electrode. Reassembled LFP/Li cells was discharged to 2 V, and Gr/Li cells were deintercalated to 2.5 V. The subtracted value of these residual capacities, we could estimate SOW capacity. The obtained estimated cell capacity by LFP-SOW showed strong correlation with SOH of cells before disassembly (R=0.98). Therefore, we confirmed that capacity fade of the cell was due to increase in SOW and/or capacity fade of LFP. To determine the degree of inhomogeneity in electrode planar, we applied XRD to LFP. As LFP reaction proceeds typical two-phase reaction, we can indirectly estimate Li content by estimating ratio of LFP and FePO 4 (FP). The LFP/FP ratio was optimized by Rietveld refinement of the diffraction peaks in various sampling points of the electrode sheet. We confirmed several macroscale inhomogeneous Li distributions in planar LFP electrode. Li poor area was observed at the beginning and the final edge of the roll electrode. The correlation details between cell SOH and degree of inhomogeneity will be discussed in the presentation. In addition, we applied two kinds of abuse test, overcharge and forced internal short circuit test by laser spot heating to three cells with different SOH. In case of overcharge, cell was charged at 0.78 C from SOC 100 %. In case of laser heating, pulse laser with 0.5 sec supply / 0.1 sec rest at 2000 W applied to the cell. The difference of the obtained parameters, and correlation with SOH will be discussed in the presentation.