Realtime Battery State of Health Diagnosis via Sensing of Cell Thermal Conductivity Tensor

Real-time state-of-health (SOH) and state of safety (SOS) diagnosis of battery cells is crucial for applications in electric vehicles as well as efficient recycling or repurposing of aged cells. We present a novel approach for SOH and SOS estimation by sensing the thermal conductivity tensor of lithium-ion battery (LIBs) cells using a thin-film sensor mounted on the exterior surface of a cell. Our method captures the highly anisotropic thermal conductivity of LIB pouch cells, which present a significantly higher thermal conductivity in the in-plane (<i>x</i> and <i>y</i>) direction compared to the cross-plane (<i>z</i>) direction due to the presence of multilayer internal structure.<br/>The key mechanisms of LIB degradation including lithium plating, dendrite formation, solid electrolyte interphase growth, structural decomposition, and transition metal dissolution are all expected to alter one or more elements of the thermal conductivity tensor unevenly. Therefore, a radiometric measurement of thermal conductivity tensor elements can be used as an indicator of structural and compositional changes inside the battery, and with an appropriate calibration can be employed for SOH and SOS estimation. This correlation provides a non-invasive, rapid, and reliable diagnostic tool, enhancing real-time monitoring of remaining useful life of the cell and early detection of battery failure.<br/>Our thin-film sensor consists of a 45 nm-thick Pt film deposited using electron-beam evaporation on a 25 μm-thick flexible thermally-conductive polyimide film. Subsequently, the Pt film was patterned to produce three serpentine-shaped resistors with dimensions on the order of 1 mm² and resistances in the range of 1-10 kohm. These three resistors were arranged in an L-shaped architecture with the first resistor placed at the center, the second resistor displaced in the x-direction, and the third resistor displaced in the y-direction by a few millimeters. Each of the resistors can act as both a local heater and a thermometer. For measurement of diagonal elements of thermal conductivity tensor, we employed the following approach. The central resistor deposits a localized modulated heat input while the temperature gradients established in <i>x</i>, <i>y</i> and <i>z</i> directions are measured simultaneously by conducting temperature measurements on all three resistors. Subsequently, using the measured temperature gradients, we look up a calibration dataset generated by conducting a series of finite element heat conduction simulations. By interpolating the calibration dataset, we obtained the diagonal values of thermal conductivity tensor. Next, to demonstrate the feasibility of the current sensor for LIB state estimation, we conducted charge/discharge cycling on a nickel-manganese-cobalt (NMC) LIB cell and showed that a correlation exists between the SOH and the ratio of in-plane to cross-plane thermal conductivity. The proposed non-invasive and cost-effective sensing technique aims to provide information on the internal structure and composition of battery cells which can be incorporated into battery management systems for online SOH and SOS estimation.