(Invited) The Reservoir Effect: When Capacity Measurements Cannot Track Cell Aging

Abstract

The initial capacity that is provided by a Li-ion battery is limited by the amount of Li+ that is supplied by the cathode. During the life of a cell, growth of the solid electrolyte interphase (SEI) will slowly consume Li+, which generally leads to capacity fade. Capacity and coulombic efficiency measurements are commonly used in battery research to quantify the evolution of SEI growth. In this talk, we show that, in reality, such measurements can often be unreliable sources of information about the extent of aging endured by the cell.[1,2] A core problem with relying on capacity measurements to track aging is that they may not probe all the Li+ that is present in the cell. See for example Figure 1a, which shows the voltage profiles of three types of anodes (graphite, silicon-graphite and SiOx) during the discharge of a full-cell. At the end of discharge, we assume that all anodes will experience a same potential of 0.6 V vs Li/Li+, which is the point in which the profiles intersect in the figure. Portions of the voltage profiles to the right of this point of intersection indicate Li+ that will remain in the anode even after the cell is “fully” discharged. That is, Li+ that was inserted into the anode during the initial charge, but that cannot be recovered at the voltage cutoff selected for cell discharge. The magnitude of this “reservoir” of Li+ depends on the shape of the voltage profile of the anode at high potentials. Failure to account for the existence of this reservoir will, for example, distort estimates about the initial Li+ consumption to form the SEI. To make matters worse, capacity stored in this Li+ reservoir can slowly be accessed as the cell ages, blurring the correlation between capacity measurements and SEI growth. During cell charging, most cathode materials of commercial interest will experience an increase in potential as they delithiate (Figure 1b). If some of that Li+ is consumed at the SEI, then the cathode cannot be restored to its original state of lithiation once the cell is fully discharged. In other words, aging will cause the cathode to experience increasing potentials once the cell meets the discharge cutoff voltage. Assuming that voltage cutoffs are constant for all cycles, an increase in cathode potential by ΔU would prompt an equal increase in the potential of the anode. A key consequence is that this bump in potential will lead to partial delithiation of the Li+ reservoir stored in the anode, releasing “extra” Li+ that can go back into the cathode. Accessing this extra capacity will offset some of the measurable Li+ losses to the SEI. Since the size of this reservoir is larger in Si than in graphite, capacity measurements in silicon cells can largely underestimate the true extent of SEI growth, and may conceal relatively high levels of cell aging under the appearance of satisfactory cycling performance. The example above shows that the reliability of using capacity measurements for conveying information about aging depends on the anode used in the cell. This same general behavior can also arise when comparing different cathode materials, or even different operating conditions (such as cycling rate and depth of discharge) for a same type of battery. This talk will discuss some of these cases, and how they can affect the diagnostic of aging in Li-ion cells.Figure 1. a) Assumed voltage profiles for anodes during the discharge of hypothetical full-cells. At the discharge cutoff voltage, all anodes will initially experience a typical value of ~0.6 V vs. Li/Li+ (indicated by the intersection point). Regions to the right of that point indicate capacity that remains stored at the anode even when the full-cell is nominally fully discharged. b) Example of successive charge and discharge half-cycles in a NMC811 cathode. The solid black line indicates portions of the voltage profile that are actively utilized in the cathode. When Li+ is lost to the SEI, the cathode cannot be restored to the Li+ content it exhibited at the beginning of charge, leading to an increase in cathode potential when the full-cell reaches the cutoff voltage.