Role of the carbon coating, the particle size, and the agglomeration on the electronic conductivity of LiFePO4-based electrodes for lithium ion batteries

The improvement of battery performance requires the rationale optimization of the composite electrode. The development of “new tools”, i.e. experimental techniques as well as methodologies, is needed to understand the so-called compositionarchitecture-properties and performance relationships. Otherwise, the composite electrode formulation and processing have to be optimized by “trial and error” experiments. The improvement in electronic conductivity of the composite electrode is crucial toward high rate performance [1,2]. Our previous work demonstrated the critical role of the polymer binder and electrode morphology on the electronic conductivity of the composite electrode [3]. We also showed that the carbon black/binder network can be considered as a macro-tunnel junction, with an exponential drop of electronic conductivity as a function of the thickness of the binder layer adsorbed at the contacts between the carbon black particles. However, the direct current (dc) transverse electronic conductivity (sample conductivity), that is the usually measured quantity, is a macroscopic averaged quantity. Recently, we used the broadband dielectric spectroscopy (BDS), from low (a few Hz) to microwave (a few GHz) frequencies. We demonstrated that this technique is very sensitive to the different scales of the electrode architecture involved in the electronic transport, from interatomic distances to macroscopic sizes, as well as to the morphology at these scales, coarse or fine distribution of the constituents [4]. In this work, BDS is used to study LiFePO4based composite electrodes. Model samples prepared from different LiFePO4 varying in particle size (from 50nm to 400nm) and carbon coating content (up to 4wt %) are studied. Several characterization techniques such as DRX, Mossbauer, SEM-EDX, XPS as well as TEMEELS are used in complement to the BDS measurements, for proper description of the samples and identification of the electrical relaxations. The results on the sample with 150nm particle size and 2.8wt% carbon coating are shown in Figure 1. Since all the polarizations at different scales are additive owing to their vectorial character, their contributions (relaxations) can thus be evidenced by a decomposition procedure of the Nyquist plots (e” vs. e’). Figure 1 shows the entire Cole-Cole plots of the complex permittivity (from 40 to 10GHz) of the C-LiFePO4 pellet at 300 K. Four dispersion domains are observed which could be attributed to: D1, polarizations due to the sample–metallic substrate junctions. D2 (1.68x10 Hz), polarization due to the C-LiFePO4 agglomerates (clusters). D3 (2.7x10 Hz), polarization due to hopping (Global) in the C-LiFePO4 coating (particles scale). And D4(1.47x10Hz), polarization (local) in the C-LiFePO4 carbon coating. Similarly, the Nyquist plots of the complex resistivity (ρ” vs. ρ’) of the same sample allow determining the resistivity of both the C-coating and of the clusters of CLiFePO4 (Table 1) as well as a comparison with the same material but without carbon coating. Data acquisitions as function of temperature was also carried out, Figure 2. The role of the LiFePO4 particle size (from 50nm to 400nm) and carbon coating content (up to 4wt %) on the electronic properties will be reported. Figure 1: Cole-Cole plots of the imaginary part e”(ω) vs. the real part e’(ω) of the complex permittivity at 300 K for C-LiFePO4 (150nm, 2.8wt%) pellet.