Understanding internal state non-uniformity that occurs across the electrodes in large-format Lithium-ion batteries (LIBs), and among parallel-connected cells, is a critical part of the cell/module design process [1]. The reduction of non-uniform current distributions is critical as this leads to non-uniform utilization, which effectively reduces energy density, and can eventually cause non-uniform degradation, leading to earlier cell/module failure. Several studies, performed particularly at high charge/discharge rates (i.e., above 1C-rate), show single cell temperature differences greater than 20°C, for large-format Lithium-ion cells [2]. As engineers maximize the energy density of Lithium-ion cells, and push them to extreme limits (e.g., Tesla Supercharging), the problems of non-uniform currents and temperatures will continue to be exacerbated. Therefore, high-rate battery systems generally require the integration of a battery thermal management system (BTMS). Coupling the optimization of cell/module, and BTMS designs, to efficiently minimize thermal non-uniformity and packaging requirements, requires spatially-resolved coupled thermal-electrical-electrochemical models. However, little published research exists regarding the measurement of spatially distributed effects in Lithium-ion batteries and the experimental validation of spatially distributed models.

Under uniform thermal conditions, the in-situ internal current/State of Charge (SOC) distributions within a segmented-cathode (LiFePO4) cell have been measured by Zhang et al. [3]. Yang et al. studied the non-uniform SOC buildup between two 18650 LiFePO4 cells connected in parallel, where each cell was held at a different temperature [4]. Klein et. al. studied the non-uniform temperature effects on the bulk cell resistance under constant current pulses and dynamic current charge-depleting cases for a LiNiMnCo pouch cell [5].

This study presents the non-uniform temperature effects on in-situ current/SOC distributions among modules of five parallel-connected cells. Two commercially available 18650 cell chemistries were used; LiFePO4/Graphite (LFP) and LiNiMnCo/Graphite (NMC). For each chemistry, the five cells were connected in parallel with a current measuring shunt in series with each cell. The test system is illustrated in Figure 1a, and the module is in Figure 1b. Average module temperatures of 15, 25, and 35°C were used with temperature differences across the module being controlled to 0, 5, 10, and 20°C. Direct-current pulse tests were performed at 20, 50, and 80% SOC using 10sec pulses at constant C/5, C/2, 1C, and 2C-rates. Additionally, C/5, C/2, and 1C-rate constant current full-capacity discharges were performed under the same thermal conditions to observe the effects of non-uniform SOC among the cells.

The pulse results indicate that the LFP cell resistance has a lower temperature sensitivity than the NMC, and therefore experiences less current non-uniformity for the same amount of temperature non-uniformity. Larger current non-uniformity is observed at lower average temperatures, where the resistance is more temperature sensitive. The current distribution develops quickly (i.e., <2sec) and remains fairly constant during the 10sec pulse. However, the constant current full discharge results are quite different. Figure 1c-e plots the measured normalized current for the left/middle/right (cold/mean/hot) LFP cells for the average module temperature case of 25°C and the C/5, C/2, and 1C-rates, respectively. The instantaneous current distributions measured at the same average module SOC as the pulse tests are more uniform in the full capacity discharge tests. Therefore, the amount of maximum non-uniform SOC is lower than would be expected from extrapolating the pulse test current distribution results. Additionally, the NMC unit, which had a larger current distribution in the pulse testing compared to LFP, has a substantially lower current-distribution than the LFP unit for the full-capacity discharges. Finally, using the measured SOC profiles for each cell, a tolerable amount of temperature uniformity for each chemistry is interpolated from the results. Figures 1f and 1g present these results for two maximum tolerable SOC non-uniformities of 2% and 5%. The NMC chemistry provides a higher tolerance to temperature non-uniformity for a given SOC non-uniformity tolerance, as compared to the LFP chemistry. The results of this study are aimed at aiding both the validation of spatially resolved models, and to quantify the thermal non-uniformity targets for these common commercially available Lithium-ion cell chemistries.

[1] T. Waldmann, G.Bisle, B.-I. Hogg, S. Stumpp, M.A. Danzer, M. Kasper, P. Axmann, M. Wohlfarht-Mehrens, Jounal of The Electrochemical Society 162 (6) (2015) A921-A927

[2] C. Veth, D. Dragicevic, C. Merten, Journal of Power Sources 267 (2014) 760-769

[3] G.Zhang, C.E. Shaffer, C.-Y. Wang, C.D. Rahn, Journal of The Electrochemical Society 160 (4) (2013) A610-A615.

[4] N. Yang, X. Zhang, B. Shang, G. Li, Journal of Power Sources 306 (2016) 733-741.

[5] M. Klein, S. Tong, J. Park, Applied Energy 165 (2016) 639-647.