Entropy in Li- and Mn-Rich Layered Oxides Measured Via Linear Temperature Variation

Abstract

For Li-ion battery development, Li- and Mn-rich layered oxide cathodes (LMRNCM, Li1+x[NiCoMn]1-xO2 with typically 0.1 < x < 0.2), are currently under development. Their exceptionally high gravimetric capacity is, however, accompanied by a significant voltage hysteresis, which complicates the SOC management and reduces the round-trip energy efficiency of the cell.1 In particular, a significant part of the voltage hysteresis is still present during open circuit conditions (OCV) being independent of the applied current and thus a material-specific property (see Figure 1a). The hysteresis is not only expressed in the voltage profile but also in other parameters, such as a path dependent cathode resistance2 and a hysteresis of the LMRNCM lattice parameters3.To investigate the OCV hysteresis, potentiometric entropy measurements were conducted to calculate the changes in the partial molar free entropy of the Li (de-)intercalation during charge and discharge, and from this, the reversible heat Qrev. The entropy is accessible via the temperature dependence of the OCV. To minimize errors from temperature history4 and relaxation5, a linear temperature variation method was established based on the work by Liebmann et al. 6 Using this method, entropy changes within the LMRNCM were measured as a function of state of charge (SOC) to reveal a path dependence between charge and discharge (see Figure 1b). This path dependence vanishes when the entropy is correlated to the OCV of the respective SOC indicating that the structural changes within the LMRNCM are rather a function of OCV than SOC, as can be seen in Figure 1c. This is in agreement with a previous study by Strehle et al.3 where a similar behavior was found for the lattice parameters of LMRNCM by diffraction methods. However, the herein conducted measurements did not reveal the entropy as a cause of the path-dependence (or hysteresis) but rather demonstrated it to be another indication of this phenomenon. Ultimately, the reversible heat Qrev will be compared to the energy loss correlating to the OCV hysteresis indicating that the latter heat term is significantly larger and independent from the entropy changes within the LMRNCM. Figure 1: (a) Voltage hysteresis and (b,c) changes in the partial molar free entropy for LMRNCM. A LMRNCM/Li Swagelok T-cell (equipped with a Li reference electrode) was cycled at C/10 at 25°C between 2.0-4.7 V. For the OCV curve, intermediate OCV phases of 1h were applied every ≈10% SOC (a). The entropy was measured in various T-cells via linear temperature variation after relaxation of the OCV at different SOCs during charge and discharge and is shown as a function of SOC (b) and OCV (c) with trendlines as guide for the eye.Acknowledgements:We want to acknowledge BASF SE for the support within the frame of its scientific network on electrochemistry.