262895 Electrochemical Characterization of Passivating Films in Lithium-Ion Batteries
Electrochemical Characterization of Passivating Films in Lithium Ion Batteries
a Environmental Energy Technologies Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720
b Department of Chemical and Biomolecular Engineering, University of California, Berkeley, CA 94720-1462
Lithium-ion batteries are a promising option for electric vehicles because of their high energy and power density. However, the high cost and relatively short lifetimes of lithium-ion batteries have so far prevented widespread automotive application. One reason for the limited lifetime of a lithium-ion battery is the solid-electrolyte-interphase, or SEI. The SEI is a passivating film that forms on the graphite anode during the first few charge cycles. The electrolytes in lithium-ion batteries are thermodynamically unstable at the potential of graphite. This means that a battery without an SEI, or with a “bad” SEI, continuously grows and consumes electrolyte. The growth of the SEI means that less lithium is available for energy storage.
Although the SEI has been studied for many years, scientists still do not understand how it prevents electrolyte reduction, or what parameters are necessary for the formation of a ‘good' SEI. Because it is sensitive to air, moisture, and impurities, the SEI is very difficult to characterize. Because formation involves many competing chemical reactions, the ability of traditional electrochemical techniques to describe the film is also limited. In this work, we develop a new method to characterize the SEI using ferrocene, a redox shuttle. By comparing ferrocene kinetics in the presence and absence of passivating films, the shuttle functions as an electrochemical probe of the mechanism by which the SEI prevents reaction. The results of this method contribute to understanding an important failure mechanism in lithium-ion batteries.
We develop two electrochemical methods for characterizing the SEI. Both experimental methods are coupled with a physics-based model that is used to interpret experimental results. First, rotating disk electrode experiments are used to measure the steady-state through-film ferrocenium reduction. The high stability of both the neutral and cationic forms of ferrocene make it ideal for surface studies. Second, electrochemical impedance spectroscopy (EIS) is used to measure the frequency response of the ferrocene reaction in the presence and absence of the SEI. After developing the method on glassy carbon, a model surface, the method is extended to highly oriented pyrolytic graphite (HOPG), which more accurately resembles the carbon found in an actual lithium-ion battery.
Results and Discussion
Steady-state measurements and model fits are shown in Fig. 1. An inert glassy carbon electrode is held at low voltage, thus reducing the supporting electrolyte and building an SEI-type film on the electrode. Varying the time at which the electrode is held at low voltage controls the film thickness. After the formation hold, the electrode is moved to a solution of supporting electrolyte with dilute ferrocene and ferrocenium hexafluorophosphate. The markers show the current measured at 900 rpm after films were built on the electrode for 30 seconds, 6, 30, and 60 minute holds at 0.6 V. Dashed lines are model fits to the passivated current, and the dotted line is the reversible current, which is seen on the clean electrode. Current decreases with passivation time because the electrode has had longer to grow a “thicker” film. The model includes only three adjustable parameters: an anodic transfer coefficient αa, an exchange current density i0, and a through-film ferrocene limiting current ilim, described by the expressions below.
Cbulk refers to the concentration in the bulk solution, and the subscripts O and R represent oxidized ferrocenium and reduced ferrocene, respectively. k is a rate constant, Do,f is the diffusivity of ferrocenium inside the SEI, F is Faraday's constant, L is the film thickness, and ε is the film porosity. The shape of the curve between 3.15 and 2.5 V is given by α and i0. ilim is determined from the limit as the curve approaches very low voltages. Both i0 and ilim decrease with increased passivation time, and because both expressions contain the porosity ε, a possible explanation may be that longer formation times cause thicker but also less porous films1.
Fig. 2 shows impedance measurements of the same films as in Fig. 1 at open circuit potential and 900 rpm. Each spectrum exhibits two arcs. The high-frequency arc depends on passivation time, but the low-frequency arc does not. The high frequency arc width increases with more passivation time. Plotting the imaginary component vs. frequency (Fig. 3) shows that similarly, the low frequency peak is independent of passivation time, but the high frequency peak decreases with passivation time. The peak frequency corresponds to the time constant of the system, τ = RctCdl, where Rct is the charge-transfer resistance and Cdl is the double-layer capacitance. corresponding to a higher charge-transfer resistance or a slower reaction. These observations agree qualitatively with the steady-state findings.
Figure 1: Steady-state current vs. voltage after different lengths of passivation holds. Markers are measurements, dashed lines are model fits. Both i0 and ilim decrease.
Figure 2: Nyquist plot of electrode after different lengths of passivation holds. Longer passivation times cause higher impedance.
Fig. 3. Bode plot of electrode after different lengths of passivation holds. The high-frequency peak depends on passivation time, but the low-frequency peak does not.
Figure 4: Comparison of through-film ferrocene impedance on the edge and basal planes of graphite.
Although comparison of the experimental EIS data with a physics-based model shows that EIS does not provide as unique a fit as the steady-state measurements, the indicators of high-frequency arc width and time constant agree qualitatively with steady-state results. Impedance also has significant experimental advantages over the rotating disk electrode; it is faster, uses less material, and is less subject to variations in temperature and bulk concentration. Most importantly, it permits the use of more materials, including those actually found in lithium-ion batteries. Previous work has found that the SEI formation reactions may differ substantially on the edge and basal planes of graphite2, 3; accordingly, the current task is to use the method developed in this work to study how passivation differs with graphite orientation. A preliminary result from this study is shown in Fig. 4. Two samples of HOPG, one with an edge fraction of 0.06 (primarily the basal surface exposed) and the other with an edge fraction of 0.6 (primarily the edge fraction exposed) were cycled from 3.7 to 0.1 V vs. Li/Li+ in a solution of 2 mM ferrocene in 1.0 M LiPF6 in EC:DEC in order to form an SEI. The impedance spectra at open circuit were measured both before and after cycling. Fig. 4 shows that, before SEI formation, impedance spectra on both samples exhibit a straight line of approximately 45o slope without any high-frequency semicircles (dashed lines). The two dashed lines collapse because the kinetics are fast on both the edge and basal plane. After SEI formation, both samples show an increased impedance, but the impedance on the edge plane is much higher than that on the basal plane, despite higher electronic activity on the edge plane. These preliminary results demonstrate the ability of the developed method to characterize materials found in actual batteries.
1. M. Tang, J. Newman, J. Electrochem. Soc., 158 (2011) A530-824-832.
2. D. Bar-Tow, E. Peled, L. Burstein, J. Electrochem. Soc., 146 (1999) 824-832.
3. K. Hirasawa, T. Sato, H. Asahina, S. Yamaguchi, S. Mori, J. Electrochem. Soc., 144 (1997) L81-L84.
This work was supported by the National Science Foundation, the Department of Energy, and the Japan Society for the Promotion of Science.
See more of this Group/Topical: Materials Engineering and Sciences Division