Search

Small pouch cells containing Si/graphite electrodes were systematically charged to 100, 120, 140, 160, 180 and 250% state of charge. Characterization of the Si/graphite electrode after the overcharge showed some features similar to those found in a pure graphite electrode. Since the amount of silicon in the electrode was ∼15 wt%, this was not unexpected. The effect of silicon was seen in the composition of the SEI layer and the trends in two components of the SEI layer.

Cells based on NMC/graphite, containing poly(vinylidene difluoride) (PVDF) binders in the positive and negative electrodes, were systematically overcharged to 100, 120, 140, 160, 180, and 250% state-of-charge (SOC). At 250% SOC the cell vented. Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) of the anodes showed several state-of-overcharge-dependent trends. Starting at 120% SOC, dendrites appeared and increased in concentration as the SOC increased. Dendrite morphology appeared to be dependent on whether the active material was on the “dull” or “shiny” side of the copper collector. Significantly more delamination of the active material from the collector was seen on the “shiny” side of the collector particularly at 180 and 250% SOC. Transition metals were detected at 120% SOC and increased in concentration as the SOC increased. There was considerable spatial heterogeneity in the microstructures across each laminate with several regions displaying complex layered structures.

Cells based on nickel manganese cobalt oxide (NMC)/graphite electrodes, which contained polyvinylidene difluoride (PVDF) binders in the electrodes, were systematically charged to 100, 120, 140, 160, 180, and 250% state of charge (SOC). Characterization of the anodes by inductively-coupled-plasma mass spectrometry (ICP-MS), X-ray photoelectron spectroscopy (XPS), and high-performance liquid chromatography coupled with electrospray ionization mass spectrometry (HPLC-ESI-MS) showed several extent-of-overcharge-dependent trends. The concentrations (by wt) of nickel, manganese, and cobalt in the negative electrode increased with SOC, but the metals remained in the same ratio as that of the positive. Electrolyte reaction products, such as LiF:LiPO3, increased with overcharge, as expected. Three organic products were found by HPLC-ESI-MS. From an analysis of the mass spectra, two of these compounds seem to be organophosphates, which were formed by the reaction of polymerized electrolyte decomposition products and PF3 or O=PF3. Their concentration tended to reach a constant ratio. The third was seen at 250% SOC only.

Two types of small pouch cells based on LiFePO4/graphite, one containing a N-methyl pyrrolidinone (NMP)-soluble binder, poly(vinylidene difluoride), and the other an aqueous-soluble binder, styrene-butadiene rubber/carboxymethylcellulose were systematically charged to 100, 120, 140, 160, 180, and 250% state of charge (SOC). The cells were then discharged to 3.0 V at room temperature before being disassembled for postmortem analysis. Microstructural changes in the anode associated with increasing SOC were more pronounced in the aqueous processed cells in comparison to the NMP-processed cells. Dendrite formation was observed on the aqueous-processed anode at 120% SOC, while the NMP-processed anode surface does not show dendrites until 250% SOC. Overall, the aqueous-processed anode surfaces displayed more evidence of microstructural degradation as a function of increasing SOC. In the NMP-processed cells, four organic compounds on the anode surface were found to show a dependence on SOC, while only two displayed a similar dependency in the aqueous-processed cells. The nature of the binder changed the number and composition of the species found at the anode.

Cells based on NMC/graphite, containing styrene-butadiene rubber/carboxymethylcellulose binder in the anodes and pVdF latex/carboxymethylcellulose in the cathodes, were systematically overcharged to 100, 120, 140, 160, 180, 250% and 270% state-of-charge. The impact of the binder was characterized by elemental analysis, SEM, and HPLC. These results were compared to similar cells just using the poly (vinylidene difluoride) binder. Not only did the binder impact the rate of transition metal transport from the cathode to the anode, it also had a marked effect on the microstructure and composition of the materials on the anode surface.

Author(s): Bartelt, NC; Li, Y; Sugar, JD; Fenton, K; Kilcoyne, ALD; Shapiro, DA; Tyliszczak, T; Chueh, WC; El Gabaly, F | Abstract: Incorporation of ions into battery electrodes can lead to phase transformations. When multiparticle phase-transforming electrodes charge or discharge, two processes must occur in each particle: the new phase must nucleate, and then grow until the particle is fully charged or discharged. A fundamental question is which of these two processes is rate limiting. Here we construct a simple stochastic model that shows how the relative rate of nucleation compared with growth determines the particle state-of-charge distributions in the electrode. We find that the number of particles that are partially charged at any time increases as the relative nucleation rate increases. The maximum number of particles that are actively charging occurs just before the time when the first particles are becoming completely charged. By comparing measured state-of-charge distributions with the model, we determine the relative rate of nucleation. We apply this procedure to measurements of the evolution of particles in LiFePO4 cathodes and show we can account for the particle state-of-charge distribution as a function of the electrode state of charge.