What Is the Rate Limiting Mechanism in Solid-State Lithium Cells at Different Pulse Operating Conditions?

Abstract

Solid-state lithium battery could reduce the safety concerns due to thermal runaway in conventional lithium-ion batteries. As lithium has a higher gravimetric and volumetric capacity compared to graphite, the use of lithium metal as the negative electrode in the solid-state lithium batteries could also improve the cell gravimetric and volumetric energy density. In practical use, batteries are often charged and discharged with a dynamic current profile of different C-rates and durations. Understanding the different bottlenecks limiting the cell performance under realistic operational conditions such as pulse current profiles could accelerate the process of cell design for solid-state lithium batteries. This study focuses on the electrochemical modelling of the solid-state lithium cells based on Li|LiPON|LiCoO2 cell configuration, in which most of the model parameters are estimated from experimental measurements. By calculating the time constants of the electrochemical processes, which are derived from the distribution relaxation time analysis, we show that ionic conductance and charge-transfer kinetics in solid-state lithium batteries happen at the millisecond timescale. The in-situ pulse and impedance measurements show that solid diffusion in the positive electrode is limiting the cell performance as the state-of-charge decreases. Figure 1(a) shows the comparison of model prediction (yellow solid curve) to the experimental pulse measurement (black solid curve) for the current density of 0.69 A/m2. The experimentally validated model is then used to compute and analyse the contributions of different overpotentials for a range of other pulse current densities. Figure 1(b) illustrates the breakdown of different maximum overpotentials as a function of the current densities derived from Figure 1(a). At low current densities, solid diffusion overpotential in the positive electrode is the dominant overpotential, but at high current densities, the ohmic drop in the solid electrolyte becomes the rate-limiting mechanism. In Figure 1(c), the maximum overpotentials at different state-of-charge are compared. While the magnitude of different overpotentials at the state-of-charge more than 30% is similar (less than ten mV), the model prediction shows that the solid diffusion overpotential becomes the performance bottleneck as the state-of-charge decreases to less than 30%. The contributions of different overpotentials are also analysed as a function of pulse durations. Figure 1(d) shows that the ohmic drop is limiting the cell performance for a short pulse duration. However, as the pulse duration increases, solid diffusion overpotential becomes more dominant than the ohmic drop. Hence, our simulation results show that the solid diffusion in the positive electrode is the rate-limiting mechanism as the state-of-charge decreases and pulse duration increases, but the ohmic drop across the solid electrolyte becomes the dominant overpotential for a short pulse time constant and with increasing current density. Based on these rate-limiting mechanisms at different operating conditions, we propose possible cell designs for future solid-state lithium batteries. Our simulations using different cell design parameters show that the solid diffusion overpotential would increase by more than 100% if the electrode thickness is doubled. This result implies that if the phase separation behaviour in the positive electrode is not addressed in the future cell design, increasing the thickness to achieve a higher capacity will only lead to more voltage losses. We also show that the electrolyte thickness should be as thin as possible to minimise the ohmic drop in the solid electrolyte. If the bulk LiCoO2 is used as the positive electrode, further improvement in the ionic conductivity beyond 10-4 S cm-1 is shown to yield marginal reduction in the ohmic drop for both low and high current densities studied in this model. Figure 1