Axially and radially inhomogeneous swelling in commercial 18650 Li-ion battery cells

Highlights

• X-ray computed tomography analysis of whole cylindrical cells.

• Helical trajectory scanning provides high resolution.

• Virtual unrolling is an effective tool for analysis.

• Mild battery aging causes inhomogeneous swelling of electrodes.

Abstract

Aging of lithium-ion batteries is especially important for applications such as battery electric vehicles, where they constitute a major part of the total cost and practically determine product lifetime. One of the main problems during cycle aging is the swelling of the electrode stack, as this results in increased mechanical stresses inside batteries and can further accelerate aging. Earlier studies have used X-ray tomography to address this issue and were focused on the role of large aberrations in electrode geometry in rapid capacity fade. In this study, however, we focus on batteries not exhibiting such a rapid deterioration, where only small changes to electrode geometry can be expected. Helical trajectory micro-computed X-ray tomography and virtual unrolling were used to reveal axially and radially inhomogeneous swelling of the jelly-roll electrode windings inside commercial 18650 batteries. The results supported by mathematical-physical simulations demonstrate the efficacy of the employed methods in the analysis of minute volumetric changes and show that regions inside the batteries that are comparatively unconstrained mechanically experience accelerated swelling. In particular, the top and bottom of the jelly-roll showed an elevated thickness increase, especially within the innermost windings.

Introduction

Lithium-ion batteries (LIBs) are a key technology for many kinds of mobile technologies, especially due to their high energy density and tunable power characteristics, while also providing a sufficiently long lifetime. Recently battery electric vehicles (BEVs) have become a dominant market segment for LIBs and have hence also started to drive battery technology development [[1], [2], [3]]. For BEVs, where large batteries are required for sufficient driving ranges, and the battery constitutes a large portion of the total cost, the lifetime of the battery is of crucial importance as battery replacement would often be economically unviable. This let battery/cell aging and lifetime become a focus of research interest [[4], [5], [6], [7], [8]], following many different approaches and often focusing on individual pieces of the puzzle towards a comprehensive understanding of this complex topic [9]. Cell aging manifests itself not only in a reduced electrical capacity of the cell but also in an increased thickness of the cell's active materials (electrode stack swelling) [10,11]. Capacity fade and swelling are two strongly interconnected manifestations of cell aging. Whereas the capacity fade is technologically seen as a global effect of the cell, swelling should be analyzed on a more local and, therefore, microscopic level to fully understand the driving factors of the structural changes in the active material.

Various processes are known to result in structural and mechanical changes of the cell both during charging and discharging as well as during storage. This includes plating of metallic Li at the carbonaceous anode as the main degradation mechanism, brought on especially by high charging currents at a high state of charge (SOC) and at low ambient temperatures [9]. During the first cycle of Li⁺ intercalation, a so-called solid electrolyte interphase (SEI) layer forms on the porous graphite anode and consumes a certain amount of the inventory of cyclable Li, reducing the available capacity of the cell [12]. In general, the largest volume changes occur within LIBs in the anode and cathode materials. The reversible volume expansion upon lithiating the graphite anode towards the LiC₆ phase was found to be up to 13% [13]. As the anode experiences these volume changes repeatedly during every cycle, a certain degree of SEI fracturing and regrowth consuming electrolyte species and Li-ions is inevitable. [[14], [15], [16]]. Moreover, dissolved transition metals from the cathode active materials can be reduced on the anode surface, triggering further SEI formation and hence capacity fade [17]. Volumetric changes of cathode materials are less significant and highly dependent on the type of active material. LFP LiFePO₄ (LFP) cathodes exhibit 6.8% volume changes during cycling [18] and common NMC cathodes such as NMC522 (LiNi_0.5Mn_0.3Co_0.2O₂) exhibit volumetric changes of around 2% only [19]. Nickel-rich structures, such as NCA (LiNi_0.8Co_0.15Al_0.05O₂) on the other hand show highly anisotropic volume changes upon a high degree of delithation. At voltages above 4 V vs. Li/Li⁺, the c parameter goes through a maximum before it suddenly collapses [20]. While the overall volume change is around 5% only [20], the anisotropy has been suggested to cause microcracking in the polycrystalline cathode material, resulting in exposure of more surface area to the electrolyte and material isolation. [21]. Xia et al. observed that the cracking is further depended on the applied charging rate, where increased cracking was observed for higher rates, resulting in an increased porosity and a specific surface area between about 7% and 11%. This cracking leads to overall volume expansion of the active layers of the electrode [22]. The exposure of additional surface area is detrimental as it can lead to increased CO₂ and O₂ gas formation by oxidative side reaction on the cathode at high potentials [23]. Gas formation because of abuse or degradation can result in worsened contact within the jelly roll [24] and will cause pressure build up in the cell, which can ultimately lead to swelling of the cell case or venting [25].

All abovementioned phenomena lead to geometrical changes in the batteries' architecture over time. Due to the fact that available space for the cell's active material is often confined, an electrode stack thickness increase usually results also in a higher mechanical pressure in the active materials, possibly even accelerating the related processes and, therefore, resulting in a severely reduced lifetime.

Evidently, analyses focused on the direct observation of geometrical changes inside the electrode stacks of LIB cells can thus provide valuable insights into cell aging, complementing the observations gained from other widely used methods using the cells' electrical characteristics [9,[26], [27], [28]] to infer aging mechanisms taking place inside LIBs. Thus, recent studies have employed micro-X-ray computed tomography (micro-CT) to non-destructively investigate the inner volume of cells that showed rapid capacity fade during cycling. Often, investigated cells were of the cylindrical 18650-type that features a so-called jelly-roll of wound cell layers, widely employed in commercial products. Waldmann et al. [29] have found that kinks in the layered structure can form in the vicinity of current collector tabs, exacerbated by high charging rates and that they could possibly be suppressed by the presence of a central pin stabilizing the layer structure. However, Pfrang et al. [30] could find the same phenomenon also in cells featuring a central pin and again attributed it to geometrical inhomogeneities in the jelly-roll, while Willenberg et al. [31] could see kinks forming also relatively far away from current collector tabs, but starting preferentially in the center of the jelly-roll. They connected the formation of kinks with the progressing formation of the SEI layer and eventual buckling of the layered structure at its weakest point in the cell. Blanc et al. [32] introduced virtual unrolling of 18650-type cells as a useful tool for visualization and analysis of morphological changes in electrodes due to battery aging. Kok et al. [33] used virtual unrolling to examine layer thickness changes without having to open a battery and could thereby show that pre-existing geometrical inhomogeneities get amplified by cycling. With the exception of the studies by Pfrang et al. and Kok et al., the geometrical changes occurring during cycling of cylindrical LIBs have only really be addressed on circular virtual cross-sections, while any observed axial inhomogeneities were ascribed to the presence of a current collector tab or progressively forming kinked regions. One reason for this lack of detail is most likely found in the trade-off between achievable resolution and imaged volume, which is necessary with classical micro-CT.

In this work, therefore, helical trajectory micro-CT was employed for the first time to image the entire volume of 18650-type LIBs with high enough resolution to distinguish all the battery's components and providing the ability to perform precise geometrical measurements throughout the entire volume of the battery. This is a major improvement compared to earlier works that had to focus on either low-resolution observation on the scale of the whole LIB cell (as in the studies listed above) or on the high-resolution analysis of only small parts like selected cross-sections, individual layers of the electrode stack or single particles/grains within the layers [34,35]. This opens up the possibility to both locate and analyze small volumetric changes taking place at the onset of aging, i.e. at a high state of health (SOH). We investigate a commercial LIB employed in a prominent BEV application, cycled in a moderate SOC range and with a moderate load profile, but with a high charging rate and compared it to a fresh and unused LIB. Pronounced radial and axial gradients could be found, even in the absence of major distortions such as kinks in the layered jelly-roll electrode structure.