Thermal Behaviour of LixMeO2 (Me=Co or Ni+Mn+Co) Cathode Materials

Abstract

The main challenge for electric vehicles, smaller portable devices and high capacity stationary storage facilities is their safety behaviour under normal and abusive operation conditions. The most commonly used cathodes in lithium-ion batteries (LIB) for portable applications contain lithium cobalt oxide (LiCoO2) as active material. The susceptibility of LiCoO2 cathodes to thermal runaway was previously reported in the framework of safety studies for energy storage systems [1]. The partial substitution of Co by other transition metals was much exploited. One of the main benefits of cells containing electrodes with new chemistries, such as lithium nickel manganese cobalt oxide (Li(NiMnCo)O2, isostructural with the layered LiCoO2), is the improved thermal stability in the fully charged state (delithiated cathodes). LIB with cathodes containing Li(NiMnCo)O2 perform better during thermal runaway tests than those with LiCoO2 [2,3]. Understanding the heat balance between the amounts of heat generated and dissipated in LIB during operation, when the concentration of Li in the electrode varies, is a critical issue. To improve battery safety, thermal management systems (TMS) were introduced to control the operation of high energy and high power batteries. Therefore, knowledge of bulk electrode thermal properties such as thermal conductivity, thermal diffusivity and specific heat capacity are required to optimize TMS design models and for increasing the predictability of multiscale simulations [4]. The thermodynamic properties of LiMeO2 (Me=Ni+Mn+Co) materials for LIB with layered structure were measured by calorimetry up to 973 K. Heat increment and specific heat capacity of samples with Ni:Mn:Co ratio of 1:1:1 and 4:4:2 were determined by transposed drop calorimetry and differential scanning calorimetry. These studies on active materials were extended to cathode level for samples with 1:1:1 stoichiometry. The cathodes under investigation were tape-casted composite thick films, both containing LiMeO2 active material mixed with additives (binder and carbon black), deposited on aluminium current collector foils. To compare the currently studied alternative material (Me=Ni+Mn+Co and Ni:Mn:Co=1:1:1) with the traditional LiCoO2, cathodes with Me=Co were also prepared using the same method and additionally investigated. Due to the presence of the additives, the thermophysical properties of the cathodes were studied as a function of temperature up to 573 K [5]. Thermal diffusivity was measured directly on cathode samples by laser flash analysis, whereas specific heat capacity measurements were performed using differential scanning calorimetry on samples prepared from the composite contained in the cathode sheet. These heat capacity data, together with those for aluminium, were used to predict the cathodes respective heat capacities. Furthermore, the transport properties were also studied as a function of lithium concentration x in the Li x MeO2 cathodes (0≤x≤1). The compositions, which were investigated in this work, were produced by electrochemical titration. The values of x were determined by laser induced breakdown spectroscopy through surface and bulk analysis for homogeneity in lithium concentration [6]. All cathode samples for the laser flash measurements were used in a dry state, without liquid electrolyte in the pores of the composite. Since most of the cell and battery pack models consider a homogeneously distributed mixture of liquid electrolyte in the porous electrodes, data provided in this study could contribute to the improvement of such models through the implementation of dry areas which coexist within cell layers [7]. These areas represent sources of ageing and degradation and lead to thermal gradients across the cells which are up to now neglected in the state of the art of the TMS. The results presented in this work are the first experimental data of this kind which combine application oriented and fundamental research. Furthermore, our temperature and concentration dependent data are critical for significantly improving simulation studies of the thermal behaviour of LIB, including thermal runaway, since current approaches assume bulk properties to be independent of temperature and lithiation degree due to lack of measured data.