An ultrasonic imaging technique has been developed to investigate the internal changes of pouch cells nondestructively. The local ultrasonic transmittance of pouch cells has been measured and used for imaging with a new ultrasonic scanning machine designed and built in-house. The wetting process of the cells is clearly observed via such ultrasonic imaging techniques. Furthermore, ultrasonic transmission images of fresh cells and aged cells with different electrolytes and cycling conditions exhibit very different ultrasonic transmittance, which can be caused by electrolyte dry-out or ''unwetting'' due to cell swelling. The ultrasonic imaging technique is a very sensitive method to probe failure mechanisms in Li-ion pouch cells.

Inorganic surface coatings such as Al 2 O 3 are commonly applied on positive electrode materials to improve the cycling stability and lifetime of lithium-ion cells. The beneficial effects are typically attributed to the chemical scavenging of corrosive HF and the physical blockage of electrolyte components from reaching the electrode surface. The present work combines published thermochemistry data with new density functional theory calculations to propose a new mechanism of action: the spontaneous reaction of the LiPF 6 electrolyte salt with Al 2 O 3 -based surface coatings. Using 19 F and 31 P solution nuclear magnetic resonance spectroscopy, it is demonstrated that the storage of LiPF 6 -containing electrolyte solution with Al 2 O 3 produces LiPO 2 F 2 , a well-known electrolyte additive. The production of LiPO 2 F 2 is also observed for electrolyte solutions that were stored for 14 days at 40 °C with Al 2 O 3 -coated LiNi 0.6 Mn 0.2 Co 0.2 O 2 (NMC622) and LiNi 0.8 Co 0.15 Al 0.05 O 2 (NCA) materials. Given the beneficial nature of this species for the lifetime and stability of lithium-ion cells, this reaction is here proposed to similarly benefit the performance of cells that use Al 2 O 3 -coated cathode materials.

Lithium-ion cells testing under different state of charge ranges, C-rates and cycling temperature have different degrees of lithium inventory loss, impedance growth and active mass loss. Here, a large matrix of polycrystalline NMC622/natural graphite Li-ion pouch cells were tested with seven different state of charge ranges (0-25, 0-50, 0-75, 0-100, 75-100, 50-100 and 25-100%), three different C-rates and at two temperatures. First, capacity fade was compared to a model developed by Deshpande and Bernardi. Second, after 2.5 years of cycling, detailed analysis by dV/dQ analysis, lithium-ion differential thermal analysis, volume expansion by Archimedes’ principle, electrode stack growth, ultrasonic transmissivity and x-ray computed tomography were undertaken. These measurements enabled us to develop a complete picture of cell aging for these cells. This then led to an empirical predictive model for cell capacity loss versus SOC range and calendar age. Although these particular cells exhibited substantial positive electrode active mass loss, this did not play a role in capacity retention because the cells were anode limited during full discharge under all the tests carried out here. However, the positive electrode mass loss was strongly coupled to positive electrode swelling and electrolyte “unwetting” that would eventually cause dramatic failure.

The effects of different duty cycles, involving mixtures of charge-discharge cycling and open-circuit storage segments, on the lifetime of single crystal NMC532/graphite cells were studied. Charge-discharge cycling was performed at C/3 with open circuit storage times of 0, 12, 84, 180 h or 3 months applied at upper cutoff voltages (UCV) of 4.1, 4.2 and 4.3 V. Testing was made at 40 °C for a period of 2.5 years. Cells tested to the same UCV showed similar capacity loss and impedance growth with time, independent of the ratio between the time spent cycling or in storage. Differential voltage analysis showed that the vast majority of the capacity loss stemmed from lithium inventory loss at the negative electrode with little or no loss of active materials. The thickness of the pouch cells after testing increased with the fraction of time spent cycling and the amount of gas generated in the cells increased with fraction of time spent in storage at the UCV. These results show that cycled and stored cells age differently, even though a similar capacity fade rate was observed in the first 2.5 years of cell life, which may cause different failure modes at the end of cell life.

Single-crystal LiNixMnyCozO2 (NMC) materials have recently garnered significant academic and commercial interest as they have been shown to provide exceptional long-term charge-discharge cycling stability in Li-ion cells. Understanding the degradation mechanisms occurring in conventional polycrystalline NMC materials in comparison to the more stable single-crystal equivalents has become a topic of significant interest. In this study, we demonstrate how multi-scale, in-situ computed tomography can be used to characterize important changes occurring in wound pouch cells containing polycrystalline or single-crystal NMC. These changes include cell-level phenomena (such as deformation of the jelly roll and electrolyte depletion) as well as electrode-scale phenomena (such as electrode thickness growth and electrode cracking). A set of twenty-one cells were scanned in total, consisting of three different electrodes: polycrystalline NMC622, single-crystal NMC811, and single-crystal NMC532. These studies were designed to characterize the effects of varying C-rate, depth of discharge, and duty cycle, so this work includes an analysis of these factors as they relate to physical changes taking place at the cell and electrode level.

Mechanical degradation of electrode materials is an important potential failure mode in lithium-ion batteries. High-energy-density cathode materials like nickel-rich NMC (LiNixMnyCozO2) undergo high anisotropic volume expansion during cycling that applies significant mechanical stress to the material. Computed tomography (CT) of cells can be used to image cell-level and electrode-level changes that result from long-term cycling, without the need for cell disassembly or destructive sampling. Previous work by our group has used synchrotron CT to show cathode thickness growth and depletion of liquid electrolyte after long-term (> 2 years) cycling of polycrystalline NMC622/graphite cells. These phenomena were attributed to cathode microcracking, but direct evidence of this was not available at the time. In this study, we present in-situ, sub-micron CT of these unmodified pouch cells, providing new insights into the morphological changes occurring at the particle level. These results confirm that extensive microcracking and dramatic morphological changes are occurring in the cathode that were not previously observed. Combined with the cell-level and electrode-level scans presented previously, this study provides a complete, multi-scale picture of cathode microcracking and how its effects propagate throughout the cell.

Electrolyte additives are a practical route to improving the lifetime and performance of lithium-ion cells. It is not well understood what makes a good additive; thus, the discovery of new additives poses a significant challenge. Computational methods have the potential to streamline the search for new additives, but it is important to compare predicted additive behavior with experimentally measured results. A new electrolyte additive, 1,3-dimethyl-2-imidazolidinone (DMI), has been evaluated in LiNi 1-x-y Mn x Co y O 2 (NMC)/graphite pouch cells as a single additive and with the co-additive vinylene carbonate (VC). This work compares the density functional theory (DFT)-predicted behavior of DMI with experimental results, including differential capacity analysis (dQ/dV), electrochemical impedance spectroscopy (EIS), high-temperature storage, gas chromatography-mass spectrometry (GC-MS) and long-term cycling tests. The DFT-calculated reduction potential of DMI is −0.63 V vs Li/Li + , consistent with the experimental observation that it reduces at a lower potential than ethylene carbonate (EC), ∼0.80 V vs Li/Li + . Although DMI turns out not to be a competitively useful additive, the good match between many aspects of the experimental results and theoretical predictions is a good indication that it is possible to understand aspects of the behavior of additives. This can guide future researchers.

Finding new electrolyte additives could help create lithium-ion batteries with better performance at high voltage, allowing higher energy density. However, finding the perfect additive remains a tremendous challenge, since researchers still don’t understand how to predict their performance. A new group of dioxazolone electrolyte additives have been tested in lithium-ion batteries alone or in combination with well-known co-additives. The new additives consist of a 3-phenyl-1,4,2-dioxazol-5-one (PDO) parent structure with or without (methoxy, fluoro and nitro) functional groups on the para position of the phenyl ring. It is found that PDO (no functional group) and p-(4-nitrophenyl)-1,4,2-dioxazol-5-one (pNDO) are the best performing dioxazolones overall and show promising results.