Effect of Fluoroethylene Carbonate (FEC) on the Performance and

These electrodes were assembled into coin-type 2325 cells (parts obtained from ... High-purity Li-battery-grade 1 M LiPF6 electrolyte solutions compos...
0 downloads 0 Views 5MB Size
ARTICLE pubs.acs.org/Langmuir

Effect of Fluoroethylene Carbonate (FEC) on the Performance and Surface Chemistry of Si-Nanowire Li-Ion Battery Anodes Vinodkumar Etacheri,† Ortal Haik,† Yossi Goffer,† Gregory A. Roberts,‡ Ionel C. Stefan,‡ Rainier Fasching,‡ and Doron Aurbach*,† † ‡

Department of Chemistry, Bar-Ilan University, Ramat-Gan 52900, Israel Amprius, Inc., 1430 O’Brien Drive, Suite C, Menlo Park, California 94025, United States

bS Supporting Information ABSTRACT: The effect of FEC as a co-solvent on the electrochemical performance and surface chemistry of silicon nanowire (SiNW) anodes was thoroughly investigated. Enhanced electrochemical performance was observed for SiNW anodes in alkyl carbonates electrolyte solutions containing fluoroethylene carbonate (FEC). Reduced irreversible capacity losses accompanied by enhanced and stable reversible capacities over prolonged cycling were achieved with FECcontaining electrolyte solutions. TEM studies provided evidence for the complete and incomplete lithiation of SiNW's in FEC-containing and FEC-free electrolyte solutions, respectively. Scanning electron microscopy (SEM) results proved the formation of much thinner and compact surface films on SiNW's in FECcontaining solutions. However, thicker surface films were identified for SiNW electrodes cycled in FEC-free solutions. SiNW electrodes develop lower impedance in electrolyte solutions containing FEC in contrast to standard (FEC-free) solutions. The surface chemistry of SiNW electrodes cycled in FEC-modified and standard electrolytes were investigated using X-ray photoelectron spectroscopy (XPS) and Fourier transform infrared (FTIR) spectroscopy. The impact of FEC as a co-solvent on the electrochemical behavior of SiNW electrodes is discussed herein in light of the spectroscopic and microscopic studies.

1. INTRODUCTION Rechargeable Li-ion batteries find a wide range of applications in portable electronic devices, implantable medical devices, electric vehicles, and so forth.1 The development of Li-ion batteries having high specific charge capacities and energy densities is one of the main challenges in battery research.2 Because of the reversible intercalation of Li-ions without significantly changing their volume and morphology, graphite was selected as the main substitute for Li-metal anode material in rechargeable Li-ion batteries. The relatively low specific capacity of graphite (372 mA h/g compared to 3800 mA h/g for Li-metal) is its main drawback, which limits the capacity of Li-ion batteries.3,4 Consequently, intensive work has been devoted to the development of new highcapacity anode materials.5 14 Among the various Li M alloying anode materials, Sn was one of the most studied elements in replacing graphite-based intercalation anodes. There are many reports on SnO2, SnO, Sn-carbon composites, and Sn M alloys as anode materials for Li-ion batteries. However, most of these materials suffer from serious capacity fading during prolonged charge discharge cycling.6,7,9,13 Recently, silicon (Si) has been reported to be an important element in Li-ion battery anodes, which can alloy with lithium up to a stoichiometry of Li3.75Si, corresponding to a specific capacity of almost 4000 mA h/g.15,16 In addition, silicon r 2011 American Chemical Society

is the second most abundant element on earth, which makes it attractive for commercial battery applications.4 The main problem with Li Sn or Li Si alloy electrodes is the fact that full lithiation is accompanied by a volume increase of 300%.3,13,17 This means that it is hard to develop stable passivating surface films on Li Sn or Li Si electrodes, which is critically important to their long-term stability upon cycling. In addition, the pronounced volume changes during Li-alloying reactions lead to the disintegration of the active mass upon charge discharge (lithiation delithiation) cycling. It was found that the use of nanometric-sized Sn and Si particles improves the stability of anodes based on the Li Sn or Li Si active mass. Anodes consisting of various morphologies of Si or Si composites such as micrometer-sized and nanometer-sized spherical particles,18,19 submicrometer pillars,20 nanowires,3,21 and nanorods22 have been also reported by previous researchers. Because the anode operates at low potentials close to that of metallic lithium, the reduction of both the solvent and the salt of the electrolyte results in the formation of surface films on the anode surface that work as a solid electrolyte interphase (SEI).23 Received: September 22, 2011 Revised: November 21, 2011 Published: November 21, 2011 965

dx.doi.org/10.1021/la203712s | Langmuir 2012, 28, 965–976

Langmuir

ARTICLE

Figure 1. Voltage profiles of the first lithiation delithiation cycle of SiNW electrodes and their irreversible capacity loss in three electrolyte solutions at 30 °C. (a) EC-DMC/LiPF6, (b) EC-DMCFEC/LiPF6, and (c) DMC-FEC/LiPF6. (Inset) Voltage profile at the beginning of the processes.

In the case of carbonate-based solvents, this film mainly consists of insoluble reduction products such as lithium alkyl carbonates, lithium carbonate, lithium alkoxide, polycarbonates, and ethers.24 26 These surface films accumulate to a certain thickness, which passivates the anode and prevents further reaction between the electrolyte species. Li-ions can be intercalated into the anode material through the surface film without degrading the host lattice structure.27 The surface film formation on graphite anodes and their composition have already been thoroughly investigated.27 30 The formation of these surface films on graphite is an important factor responsible for reversible cycling and long-term stability.31,32 Many electrolyte additives were successfully employed to modify the surface chemistry of graphite anodes for Li-ion batteries.28,33,34 Some additives were useful in reducing the irreversible capacity loss and improving the cycle life by modifying the surface film composition.28 All of the relevant reports appearing in the literature emphasize the fact that the nature of the surface films formed on electrodes plays an important role in the long-term stability of Li-ion batteries. The use of surface fluorinated graphite in Li-ion batteries has been reported recently to improve the electrochemical performance.27,29,30,35 In addition to this, the use of various fluorinated electrolytes demonstrated the enhanced electrochemical performance of conventional graphite anodes.29,30,35,36 The advantages of using fluorinated carbonate electrolytes are their lower melting points, increased stability toward oxidation, and low flammability.35,37,38 Because the application of SiNW's as Li-ion battery anodes started only recently, little attention has been directed toward understanding the effect of fluorinated electrolytes on the surface chemistry and electrochemical performance of SiNW's. Some researchers reported that there is no surface film formation on Si electrodes during charge discharge processes.39 However, we should expect a surface film formation on any form of lithiated silicon because of the fact that these electrodes always operate at

potentials where the organic electrolyte solutions are thermodynamically unstable.40 In addition, high surface area materials such as SiNW's in their lithiated form can reduce polar aprotic electrolyte solutions at higher rates.23 Because the surface films control the electrochemical behavior, safety, shelf life, and cycle life of the Li-ion batteries, it is necessary to analyze the structure and properties of surface films formed on Si-electrodes in order to develop reliable Li-ion batteries using Si-anodes. Fluorinated solvents such as FEC, MFE, and TTFP were addressed in recent years as an important electrolyte additive for Li-ion batteries, especially with Si-anodes.37,38,41 The mechanism of how FEC works as an important electrolyte additive is not yet clear, but we believe that the work described in this article takes this issue a few steps forward. The work reported herein describes the electrochemical performance, morphology, and surface chemistry of SiNW electrodes in standard alkyl carbonates/LiPF6 solutions containing FEC as a co-solvent.

2. EXPERIMENTAL METHODS The SiNW electrodes were synthesized by Amprius Inc. (U.S.A.) via a method similar to that described by Chan et al.42 Because the SiNW's were exposed to oxygen during their production, they were covered by thin layers of silicon oxide (as analyzed by XPS). These electrodes were assembled into coin-type 2325 cells (parts obtained from NRC, Canada) using a Celgard 3401 polypropylene separator, 1 M LiPF6 electrolyte solutions (in alkyl carbonate mixtures), and Li-foil as the counter electrode. The electrochemical cells were assembled in a glovebox (VAC Inc.) filled with high-purity argon (99.9995%). The gloveboxes were also equipped with O2 and H2O absorbers and detectors in order to bring the oxygen and moisture contents in the atmosphere to below the parts per million level. After assembly, the two-electrode cells were stored at room temperature for 24 h to ensure the complete impregnation of the electrodes and the separators with the electrolyte solution. 966

dx.doi.org/10.1021/la203712s |Langmuir 2012, 28, 965–976

Langmuir

ARTICLE

Figure 2. First consecutive cyclic voltammograms of SiNW electrodes in three electrolyte solutions at 30 °C. (A) EC-DMC/LiPF6, (B) DMC-FEC/ LiPF6, and (C) EC-DMC-FEC/LiPF6. Scan rate: 1st cycle = 0.02 mV/s, 2nd 20th cycles = 0.2 mV/s). High-purity Li-battery-grade 1 M LiPF6 electrolyte solutions composed of dimethyl carbonate (DMC), and ethylene carbonate (EC) (1:1 w/w) (Ube Industries, Japan) were used as received. Fluoroethylene carbonate (Solvay, Li battery grade) was mixed (10%) separately with a solution of DMC and EC-DMC (1:1) to form binary and ternary solvent mixtures, respectively. The HF and H2O contents in these solutions were not more than 30 and 10 ppm, respectively. A Maccor 2000 multichannel battery tester was used for the galvanostatic charge discharge cycling of the two-electrode cells in the potential range of 0.05 1.0 V at a 1 C rate. Other electrochemical measurements, such as cyclic voltametry and impedance spectroscopy, were carried out with a 1470 battery test unit model coupled to a Solartron 1255 frequency response analyzer (driven by Corrware and Zplot software from Scribner Associates, Inc.). All of the potentials expressed in this article are given vs Li/Li+ unless otherwise stated. The electrode capacity calculations were made in triplicate, and the results were within the 5% error limit. X-ray photoelectron spectroscopy (XPS) analysis of SiNW's cycled in various electrolytes was carried out using a Kratos Axis HS spectrometer (England) equipped with an Al Kα X-ray radiation source (photon energy 1486.6 eV). A homemade transfer system equipped with a gate valve and a magnetic manipulator were used for the transfer of the highly sensitive samples from the highly pure argon atmosphere of the glovebox to the XPS system. The high-resolution scans were recorded at a resolution and pass energy of 0.1 and 23.5 eV, respectively. The binding energies of all of the elements present were determined by setting the CC/CH component of the C 1s peak at 284.5 eV. Quantitative surface chemical analysis was performed using high-resolution core-level spectra after the removal of a nonlinear Shirley background. The spectra obtained were analyzed and deconvoluted using Vision Software (Kratos). Overlapping signals were analyzed after deconvolution into

Gaussian/Lorenzian-shaped components. The components of the surface films formed on cycled electrodes were characterized by FTIR spectroscopy (Nicolet 6700 spectrometer) placed in a glovebox (made by Ravona Inc., Israel) under moisture- and CO2-free air (purified by a Balston Inc. air purifier). The FTIR spectra of these electrodes were obtained with a diffuse reflectance accessory from Pike Technologies (in the range of 4000 400 cm 1). Scanning electron microscopy (SEM) studies were conducted on pristine and cycled electrodes using an Inspect FEI microscope working at an accelerating voltage of 15 kV and a magnification of 10 000. TEM measurements were carried out using the Technai G2 TEM system from FEI Inc.

3. RESULTS AND DISCUSSION 3.1. Electrochemical and Microscopy Analysis. The basic electrochemical behavior of SiNW electrodes was studied in three electrolyte solutions (EC-DMC/LiPF6, DMC-FEC/LiPF6, and EC-DMC-FEC/LiPF6) using chronopotentiometry and voltammetry (vs Li counter and reference electrodes). Figure 1 shows typical voltage profiles of the electrodes in these solutions during the first cycle. The inset of Figure 1 enlarges the scale of the presentation in order to exhibit the initial part of the voltage profiles. The irreversible charges in these first cycle experiments are emphasized in this figure by the vertical dashed lines. In general, the voltage profiles corresponding to the first discharge (the first Li intercalation of SiNW's) of these half cells containing fluorinated and nonfluorinated electrolyte solutions were similar, reflecting surface film formation and Li Si alloying. The initial reduction processes were amazingly short. The potential drops in 967

dx.doi.org/10.1021/la203712s |Langmuir 2012, 28, 965–976

Langmuir

ARTICLE

Figure 4. Effect of the electrolyte solutions on the SiNW electrodes’ polarization during galvanostatic cycling at 30 °C. (a) EC-DMC/LiPF6, (b) EC-DMC-FEC/LiPF6, and (c) DMC-FEC/LiPF6.

Figure 3. Galvanostatic cycling performance of SiNW electrodes in three electrolyte solutions at 30 °C (1 C rate). See the voltage limits in Figure 1. (a) EC-DMC/LiPF6, (b) DMC-FEC/LiPF6, and (c) ECDMC-FEC/ LiPF6.

demonstrate very prolonged cycling. In parallel studies, SiNW's were cycled hundreds of cycles in half-cell experiments (constant capacity mode, capacity >1000 mA h/g), demonstrating very stable cycling. Also, many hundreds of cycles could be demonstrated with full cells comprising Si electrodes, Li-insertion cathodes, and EC-DMC-FEC/LiPF6 solutions. This work is in progress and is beyond the scope of this article. In this work, experiments were terminated after 30 cycles because it was found that this number of cycles is enough to bring the SiNW electrodes to steady morphology and surface chemistry. Considerable capacity fading was identified in the case of FECfree standard electrolyte solutions. It is evident that the use of DMC-FEC/LiPF6 and EC-DMC-FEC/LiPF6 solutions enabled the attainment of reversible capacities of 1224 and 1159 mA h/g in contrast to 343 mA h/g for the EC-DMC/LiPF6 solution after 30 cycles. In addition to the superior reversible capacities exhibited by cells containing FEC, a significant decrease in the electrode polarization was also identified (10 and 100 mV decreases for EC-DMC-FEC/LiPF6 and DMC-FEC/LiPF6, respectively, in comparison to that for EC-DMC/LiPF6) (Figure 4). This means that surface films with better electrical properties are developed on the SiNW electrodes in electrolyte solutions containing FEC.35 Typical cycling data of these electrodes in three different electrolyte solutions at 60 °C are also presented (Figure S1). These measurements reflect the general good cyclability of SiNW electrodes in alkyl carbonate/LiPF6 solutions that seem to be superior, compared to other types of Si or Si composite-based electrodes.4,23,47,48 A cycling efficiency very close to 100% was demonstrated, and the positive effect of FEC on the performance of these electrodes was spectacular. Electrochemical impedance spectroscopy (EIS) measurements can provide valuable information on all of the time constants related to the electrode processes, including transport phenomena such as Li-ion migration through surface films, charge transfer, phase transition, solid-state diffusion, capacitive behavior, and so forth.49 Because the information provided by EIS is not specific at all, the main problem associated with EIS measurements is the ambiguous interpretation of the spectral features.50 Nevertheless, EIS can serve as an efficient tool for comparing the general behavior of electrode/solution interfaces. For example, generally a

these galvanostatic polarizations were very fast to the low potentials, in which Li Si alloying occurs (