Research Article www.acsami.org
Cite This: ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Multistage Mechanism of Lithium Intercalation into Graphite Anodes in the Presence of the Solid Electrolyte Interface Franz Dinkelacker,†,§ Philipp Marzak,†,‡,§ Jeongsik Yun,†,‡ Yunchang Liang,† and Aliaksandr S. Bandarenka*,†,‡ †
Physik-Department ECS, Technische Universität München, James-Franck-Straße 1, 85748 Garching, Germany Nanosystems Initiative Munich (NIM), Schellingstraße 4, 80799 Munich, Germany
‡
S Supporting Information *
ABSTRACT: A so-called solid electrolyte interface (SEI) in a lithium-ion battery largely determines the performance of the whole system. However, it is one of the least understood objects in these types of batteries. SEIs are formed during the initial charge−discharge cycles, prevent the organic electrolytes from further decomposition, and at the same time govern lithium intercalation into the graphite anodes. In this work, we use electrochemical impedance spectroscopy and atomic force microscopy to investigate the properties of a SEI film and an electrified “graphite/SEI/electrolyte interface”. We reveal a multistage mechanism of lithium intercalation and deintercalation in the case of graphite anodes covered by SEI. On the basis of this mechanism, we propose a relatively simple model, which perfectly explains the impedance response of the “graphite/SEI/electrolyte” interface at different temperatures and states of charge. From the whole data obtained in this work, it is suggested that not only Li+ but also negatively charged species, such as anions from the electrolyte or functional groups of the SEI, likely interact with the surface of the graphite anode. KEYWORDS: lithium-ion batteries, solid electrolyte interface, electrochemical impedance spectroscopy, graphite anodes, three-stage mechanism, SEI thickness organic electrolytes.5−8 Still, the SEI remains perhaps the least understood part in Li-ion batteries. Recently, we utilized analysis of electrochemical impedance spectroscopy (EIS) data to reveal the nature of the mechanisms of interfacial charge and mass transfer during (de-)intercalation in various battery systems and identified, for example, that these are significantly influenced by the electrolyte composition.9−13 Because of the dissimilar time constants of different electrochemical processes, EIS may directly provide valuable information about battery systems in situ and in operando.14−25 Although there have been numerous studies of graphite electrodes using EIS, fitting of the impedance spectra is often missing because of the difficulty of their interpretation and analysis.26−32 In this work, we reveal a multistage mechanism of lithium intercalation and de-intercalation in the case of graphite electrodes covered by SEI. On the basis of this mechanism, we propose a relatively simple impedance model, which perfectly explains the EIS response of the graphite/SEI/
1. INTRODUCTION After the successful commercialization of Li-ion batteries, the demand for these batteries has been continuously increasing because of their applications in numerous electric devices. Furthermore, with growing concerns about environment and sustainability of resources, traditional energy sources have been being replaced with renewable energy ones and the regulation on automobile industries has been tightened up as years go by. With the wider propagation of the electric vehicles, the demand for Li-ion batteries is forecasted to progressively increase.1 Rechargeable Li-ion batteries were first commercialized using oxide cathodes and graphite anodes by Sony in 1991.2−4 After more than two decades of research, many advanced cathode materials have been proposed. However, graphite still occupies the firm position as the main anode material in Li-ion batteries. Graphite shows a reasonable specific capacity, flat potential profile, and low discharge potential close to Li/Li+, and it is relatively abundant. However, its performance is largely affected by the growth of the so-called solid electrolyte interface (SEI) on the electrode surface. Therefore, there has been a large number of studies to understand the role of SEI layers in the interfacial charge and mass transfer between graphite and © XXXX American Chemical Society
Received: December 9, 2017 Accepted: March 14, 2018 Published: March 14, 2018 A
DOI: 10.1021/acsami.7b18738 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 1. (A) Experimental PAT-core cell from EL-CELL. (B) Sketch of an assembled cell including housing and cell components. (C) Typical potential curve for the graphite electrodes at C/10, where the three intercalation plateaus associated with different occupancy states of the graphene layers are marked as states I, II, and III. To secure stable measurement conditions, all impedance measurements were performed in a temperature chamber with thermoelectric Peltier technology “KT 53” (Binder) ensuring temperature fluctuations not greater than 0.1 °C. For exemplary determination of the dielectric constant of a SEI, highly doped silicon wafer electrodes with a surface area of ∼1.13 cm2 and highly oriented pyrolytic graphite (HOPG) were used. The wafer electrodes were assembled in a half-cell configuration with a lithium counter electrode using the Swagelok cell design with the 260 μm glass fiber separator. As an electrolyte, 1 M LiPF6 in fluoroethylene carbonate (FEC)/EMC 3:7 (v/v) with 2 wt % vinyl carbonate was used. Again, a Li−metal point reference electrode enabled threeelectrode impedance measurements. The passivating SEI film was formed onto the wafer surface with a controlled voltage profile. The open-circuit potential (OCP) of the cell was ∼1.2 V versus Li/Li+ after assembly and was maintained for 4 h as a waiting period. The following potential profile consisted of a ramp down from the OCP to 200 mV versus Li/Li+ with a scan rate of 15 mV/s and a holding period of 20 h at 200 mV versus Li/Li+. The chosen potential assured no incipient lithiation of the silicon wafer which takes place at potentials smaller than 150 mV versus Li/Li+. The SEI formation on HOPG was performed in a three-electrode setup using HOPG as a working electrode and Li−metal as reference and counter electrodes.33 The electrolyte contains a 1 M LiPF6 solution in EC/EMC 3:7 (v/v) with 2 wt % vinyl carbonate. After the assembly, the OCP of the cell exhibited ∼2.5 V versus Li/Li+ which was maintained for 4 h as a waiting period. The following potential profile consisted of a ramp down from the OCP to 100 mV versus Li/Li+ with a scan rate of 15 mV/s and a holding period of 20 h at 100 mV versus Li/Li+. After the SEI formation, the electrode potential was applied to 2.0 V versus Li/ Li+ for 1 h to fully delithiate the HOPG for better SEI thickness measurements. For the SEI thickness determination, the cells were opened inside the glovebox and rinsed carefully using FEC/EMC 3:7 (v/v) and EC/ EMC 3:7 (v/v) without conducting salt or additives in case of the Si and HOPG electrodes, respectively. This step is necessary to remove excess of the conducting salt from the surface after evaporation of the solvents. Subsequently, half of the film in the area not in contact with the separator was removed using a razor blade. Previous to this procedure, the razor was applied to a fresh silicon wafer and thickness profiling revealed no residual marks at the surface. This is important to avoid confusion of steps in the film thickness with steps on the wafer
electrolyte interface at different temperatures and states of charge (SOC). Using this model, we propose that not only Li+ but also negatively charged species, such as functional groups of SEI or anions, can likely interact with the surface of the anode.
2. EXPERIMENTAL SECTION The battery cells were assembled using both commercially manufactured nickel cobalt manganese (NCM) dioxide cathodes and graphite anodes with a 260 μm glass fiber separator. The cathode material contained a ratio of 6:2:2 (LiNi0.6Co0.2Mn0.2O2) where 4% of a conductive agent and 2% of binder were added. The graphite anode consisted of 96.7% graphite, 1% conductive additive, and 2.3% binder. Both electrode materials were compressed for higher volumetric energy density (so-called calendaring), with a resulting porosity of ∼30%. The thickness of the NCM electrode was ∼49 μm, and the graphite electrode was ∼44 μm; both electrodes had a geometrical surface area of ∼2.54 cm2. A commercial EL-CELL PAT-core cell (Figure 1A) was used for electrochemical measurements. The PATcore cell was assembled using a sleeve equipped with a built-in ringshaped Li reference electrode allowing reliable impedance measurements using a three-electrode configuration. The electrolyte was composed of a 1 M LiPF6 solution in ethylene carbonate (EC)/ethyl methyl carbonate (EMC) 3:7 (v/v) with 2 wt % vinyl carbonate. To ensure low cell resistance, a defined force is applied to the “sandwich” using a spring. The scheme of this assembly is shown as a sketch in Figure 1B. After assembly, the SEI formation was performed for two cycles at a C/10 rate. The potential curve of graphite is shown in Figure 1C. The cycling followed a so-called “cc−cv protocol” of the battery between 3.0 and 4.2 V. The cv step was limited by the current falling below C/ 40. A capacity determination cycle was performed at C/2, where the capacity was calculated from the discharge step. Subsequently, the cell was charged and again discharged to the desired SOC. All battery tests (full charge/discharge, cycling and capacity measurements) were performed using a multichannel “Material Development System” (BaSyTec) or “Cell Test System” (BaSyTec). Impedance spectroscopy was performed using an impedance analyzer “IM6ex” (Zahner-elektrik). The spectra were collected in the potentiostatic mode in the frequency range between 100 kHz and 10 mHz. The excitation amplitude was set to 10 mV. To validate the measured impedance data, Kramers−Kronig checks were performed. B
DOI: 10.1021/acsami.7b18738 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
100 −1 kHz, approximation is given in red) and the midfrequency region (MFR, 1 kHz to 1 Hz, approximation is given in blue). This curvature is divided into one semicircle in the HFR and one in the MFR, which can be clearly observed at low temperatures (see Figure 2B). The inflection point between MFR and low-frequency region (LFR, 1 Hz to 10 mHz, green area) is followed by a linear slope toward the very low frequencies. As shown in Figure 2B, x-intercepts of the impedance spectra in the HFR, which are related to Ohmic losses, such as the electrolyte uncompensated resistance, contacts of active materials, and current collectors, as well as the diameters of semicircles in the MFR that are correlated with the charge-transfer resistance, tend to increase as the temperature decreases. This fact might be interconnected with the reduced mobility of ions in the electrolyte at lower temperatures, additionally slowing down the kinetics of the otherwise thermally activated Li+ intercalation and de-intercalation. On the contrary, the slopes of the impedance spectra in the LFR appear to be unchanged irrespective of the temperature. As shown in Figure 2C, characteristic loops can be observed in the transition between MFR and LFR at low SOCs (
