Article Cite This: Acc. Chem. Res. 2018, 51, 80−88
pubs.acs.org/accounts
Electrochemical Interphases for High-Energy Storage Using Reactive Metal Anodes Published as part of the Accounts of Chemical Research special issue “Energy Storage: Complexities Among Materials and Interfaces at Multiple Length Scales”. Shuya Wei,† Snehashis Choudhury,† Zhengyuan Tu,‡ Kaihang Zhang,† and Lynden A. Archer*,‡ †
Robert Frederick Smith School of Chemical and Biomolecular Engineering and ‡Department of Materials Science and Engineering, Cornell University, Ithaca, New York 14853, United States CONSPECTUS: Stable electrochemical interphases play a critical role in regulating transport of mass and charge in all electrochemical energy storage (EES) systems. In state-of-theart rechargeable lithium ion batteries, they are rarely formed by design but instead spontaneously emerge from electrochemical degradation of electrolyte and electrode components. Highenergy secondary batteries that utilize reactive metal anodes (e.g., Li, Na, Si, Sn, Al) to store large amounts of charge by alloying and/or electrodeposition reactions introduce fundamental challenges that require rational design in order to stabilize the interphases. Chemical instability of the electrodes in contact with electrolytes, morphological instability of the metal−electrolyte interface upon plating and stripping, and hydrodynamic-instability-induced electroconvection of the electrolyte at high currents are all known to cause metal electrode− electrolyte interfaces to continuously evolve in morphology, uniformity, and composition. Additionally, metal anodes undergo large changes in volume during lithiation and delithiation, which means that even in the rare cases where spontaneously formed solid electrode−electrolyte interphases (SEIs) are in thermodynamic equilibrium with the electrode, the SEI is under dynamic strain, which inevitably leads to cracking and/or rupture during extended battery cycling. There is an urgent need for interphases that are able to overcome each of these sources of instability with minimal losses of electrolyte and electrode components. Complementary chemical synthesis strategies are likewise urgently needed to create self-limited and mechanically durable SEIs that are able to flex and shrink to accommodate volume change. These needs are acute for practically relevant cells that cannot utilize large excesses of anode and electrolyte as employed in proof-of-concept-type experiments reported in the scientific literature. This disconnect between practical needs and research practices makes it difficult to translate promising literature results, underscoring the importance of research designed to reveal principles for good interphase design. This Account considers the fundamental processes involved in interphase formation, stability, and failure and on that basis identifies design principles, synthesis procedures, and characterization methods for enabling stable metal anode−electrolyte interfaces for EES. We first review results from experimental, continuum theoretical, and computational analyses of interfacial transport to identify fundamental connections between the composition of the SEI at metal−electrolyte interfaces and stability. Design principles and tools for creating stable artificial solid−electrolyte interphases (ASEIs) based on polymers, ionic liquids, ceramics, nanoparticles, salts, and their combinations are subsequently discussed. Interphases composed of a second electrochemically active material that stores charge by different processes from the underlying metal electrode emerge as particularly attractive routes toward so-called hybrid electrodes that enable facile scale-up of ASEI designs for commercially relevant EES.
1. INTRODUCTION
design. Instead, uncontrolled reduction/oxidation reactions between electrolyte and electrode components are utilized to produce new interphases with composition and structure that reflect the thermodynamic stability and solubility of the reaction products. In favorable situations (e.g., the lithium ion battery (LIB) anode), wherein the active graphitic carbon
Electrochemical energy storage (EES) systems, including batteries and supercapacitors, are built from multiple components that interact via interfaces. These interfaces regulate transport of charge and matter between components and are therefore critical determinants of performance, safety, and reliability. Despite their importance, the most challenging interfaces in modern EES systems (i.e., between the highly reducing anode and the electrolyte) are rarely formed by © 2017 American Chemical Society
Received: October 7, 2017 Published: December 11, 2017 80
DOI: 10.1021/acs.accounts.7b00484 Acc. Chem. Res. 2018, 51, 80−88
Article
Accounts of Chemical Research
on nanometer length scales on the order of the thickness of the SEI. Concentration of electric field lines on these so-called dendrite nucleates drives mossy or treelike deposit growth to micrometer length scales, producing the morphological instability termed dendrites. Because the processes that drive the instability are fundamentally linked to the reactivity of the electrolyte at the anode surface, the instability is present at all current densities and is most severe when poorly ionconducting interphases are formed on the electrode.5,6 At high current densities, above the limiting current, a second hydrodynamic instability determines the interface morphology. Under such conditions, the rate of migration of metal ions to the interface exceeds the rate at which diffusion replenishes counterions to maintain electroneutrality in the electrolyte. This produces a space charge region at the electrode−electrolyte interface that may extend many hundreds to thousands of ionic diameters into the electrolyte bulk (i.e., well beyond the quasiequilibrium Guoy−Chapman space charge formed spontaneously at any charged substrate in an electrolyte).7 The space charge drives dendrite formation on all metals, including Mg, which in early literature was incorrectly assumed to be inherently resistant to dendrite formation,8 by a hydrodynamic instability known as electroconvection.
particles are chemically inert and undergo only small (∼10%) changes in volume during battery cycling, spontaneously formed interphases are thin (10−100 nm), self-limiting, electronically insulating, and ionically conductive and remain intact for the battery’s lifetime. This minimizes the amount of electrolyte and electrode material consumed in forming and maintaining the interphase and allows the electrolyte to operate stably between the electrodes at potentials outside its stability limits.1 The situation is rather different for metallic anodes of contemporary interest for their ability to store much larger amounts of energy per unit mass or volume than graphite. These anodes undergo larger (60−400%) volume change during lithiation/delithiation reactions, imposing more stringent mechanical demands on spontaneously formed interphases. This is compounded by the fact that some anodes (e.g., Li, Na) are chemically unstable and their interfaces undergo cumulative changes in morphology and shape over time. As a result, spontaneously formed interphases rarely, if ever, meet the requirements for stable, long-term cycling of energetic metal anodes. Strategies for controlling the structure, composition, and evolution of such interphases are therefore urgently needed to improve the Coulombic efficiency, energy density, safety, power, and battery lifetime.2,3 In this Account, we consider the fundamental processes involved in interphase formation, stability, and failure and on that basis identify design principles and synthesis procedures for enabling stable metal anode−electrolyte interfaces for EES systems. Although the emphasis is on Li and Na metal electrodes, where the challenges are greatest, some of the ideas discussed are relevant to less reactive but comparably energetic anodes based on Si, Sn, and Zn.
3. DESIGNING SEIS FOR ACTIVE METAL ELECTRODES 3.1. Design Principles for Artificial SEIs
The ideal SEI for a reactive metal anode should be able to eliminate all three modes of instability and at the same time must be mechanically tough enough to accommodate volume changes at the electrode without cracking or disintegrating. While polymeric and ceramic coatings on the electrodes can impart many of these characteristics, the added requirements that the SEI possess high ionic conductivity and a large electrochemical stability window pose a significant challenge. Strategies for regulating transport of electrons, ions, and molecules at the electrolyte−electrode interface have therefore emerged as among the most promising routes for designing stable SEIs for metallic anodes. Previously, Tikekar et al.9 formulated a continuum theory to understand stability of metal electrodeposition and to determine the circumstances under which deposition is unstable to morphological and hydrodynamic instabilities. The key conclusions of this study are that stability can be achieved at high and low current densities: (i) if the electrolyte−electrode interface maintains a high ionic conductivity under strong bias; (ii) if the shear modulus of the interphase is large relative to the elastic modulus of the metal; and (iii) if the size of the initial nucleates is sufficiently small, in which case the surface tension is enough to prevent nucleate growth. The most straightforward approaches for creating SEIs that overcome chemical instability are unfortunately not always consistent with the requirements for arresting morphological and hydrodynamic instability. Closed-packed protective/ passivating coatings on the metal electrode using high-modulus but electrochemically stable inorganic materials such as Al2O3, SiO2, LiPON, etc. are among the most common, but these materials are effective because they prevent access of electrolyte components, which leads to poor interfacial ion transport.10 This means that deposition is largely enhanced at any cracks or pinhole defects in the coatings, which promotes electrode failure by morphological and hydrodynamic instabilities. A fine balance must therefore be struck among protecting the anode,
2. INTERPHASE FORMATION AND FAILURE MECHANISMS The composition, physical properties, and evolution of interphases formed on all electrodes are the result of chemical, morphological, and transport processes, which are responsible for interphase formation and evolution in space and time. On reactive anodes, these processes fundamentally determine the rate at which electrolyte and electrode components are consumed and are typically difficult to control at the potentials required for high-energy and high-power storage. Metallic electrodes standout because one or more of these processes are fundamentally unstable, leading to interphases that evolve too rapidly for their morphology and physical properties to be controlled. Chemical instability of a metal anode arises from a combination of two sources: the intrinsic reactivity of the metal toward electrolyte components and the high potential of the anode relative to the lowest unoccupied molecular orbitals (LUMOs) of the electrolyte. Left unchecked, both processes result in continuous reduction of electrolyte components to produce larger molecules (e.g., oligomers and cyclic polymers)1 and binary salts (e.g., LiF, Li2O, Li2S, Li3N, Li3P, etc.)4 that are thermodynamically stable in contact with the electrode and electrochemically stable at its working potential. The chemistry, composition, and structure of the resultant solid−electrolyte interphase (SEI) are therefore set by the chemical potentials of the electrolyte components, the cell running conditions, and the interfacial reaction kinetics. If the composition of the interphase is nonuniform in space, its ionic conductivity is also nonuniform. In practice this leads to localized electrodeposition of active metals (e.g., Li, Na, Al) 81
DOI: 10.1021/acs.accounts.7b00484 Acc. Chem. Res. 2018, 51, 80−88
Article
Accounts of Chemical Research
Figure 1. (a) Diffusion energy barrier at interphases composed of pure metals and of representative electrolyte reduction products. (b) Schematic and cryo-FIB-SEM images of the Lithion-protected Li anode. Reproduced with permission from ref 11. Copyright 2017 Cell Press. (c) Schematic and SEM image of an ionic-liquid-protected Na anode. Reproduced with permission from ref 5. Copyright 2017 Wiley. (d) Schematic depicting the function of the alloy-protected lithium foil and cross-sectional images of a composite Li13In3|Li foil before and after Li plating. Reproduced with permission from ref 14. Copyright 2017 Nature Publishing Group. (e) Schematic of hollow carbon sphere-protected Li and SEM images of hollow carbon nanospheres after the initial SEI formation process. Reproduced with permission from ref 15. Copyright 2014 Nature Publishing Group.
recent years for their ability to provide direct routes toward SEIs enriched in thermodynamically stable polymers and binary salts (e.g., LiF, Li3N, Li3P, etc.) that not only passivate the metal anode but also appear to facilitate interfacial ion transport.10 Important results from recent density functional theory analyses in fact show that binary salts are not all equal in their effectiveness as SEI components. As illustrated in Figure 1a, an SEI enriched in halide (e.g., LiF, LiI, NaBr) offers lower barriers to in-plane diffusion of ions than those computed for spontaneously formed carbonates, hydroxides, sulfides, nitrides, or phosphides. This means that well-designed, electrochemically stable interphases enriched with halide salts should simultaneously enable compact (dendrite-free) electrodeposition and low surface area for parasitic reactions.12 This prediction has been empirically investigated in a growing number of electrochemical settings and is generally found to be highly effective in promoting morphologically stable SEIs on Li, Na, and Si anodes. Interfaces formed using various salts (e.g., lithium bis(oxalato)borate, LiNO3, lithium polysulfides)13 and ionic liquids5 (Figure 1c) offer impressive benefits, but the source of the benefits remain an area of active debate. A more recent example takes advantage of the electropositivity of reactive metals such as Li and Na to enable ion exchange reactions with soluble salts containing In, Sn, or Si groups to create thin coatings of the respective metals or their Li/Na alloys on anodes. These SEIs are electrochemically active and thus play dual roles (Figure 1d).14 In so doing they reduce the weight of inactive material in the anode. While a graphitic
facilitating ion transport, and minimizing the added weight and volume of the coating. SEI designs that enable synergistic benefits are considered ideal. Tu et al.,11 for example, recently reported that coatings of lithiated Nafion (Lithion) deposited on Li anodes at thicknesses in the range 10 μm to 20 nm produce lithium metal batteries (LMBs) with near-unity lithium transference numbers, tLi+, mS/cm-level bulk ionic conductivities at room temperature, and low interfacial resistance (Figure 1b). Because such coatings offer a fixed, uniform concentration of anions anchored at the anode surface, below 200 nm thickness they are able to maintain high interfacial ionic conductivity during charging and as a result offer a fundamental mechanism for stabilizing the interface against morphological and hydrodynamic instability. Furthermore, because of the mechanical toughness of Nafion, depth-resolved experiments using cryogenic focused ion beam scanning electron microscopy (cryo-FIB-SEM) analysis revealed that Lithion artificial SEIs (ASEIs) with thicknesses below 200 nm can flex to accommodate changes in the interphase morphology and anode volume. The approach is attractive because it is in principle applicable to any metal electrode and can be extended to ASEIs based on functionalized ceramics and nanoparticles. Sacrificial electrolyte components designed to form ionconducting polymers and salts offer an older strategy for creating ASEIs for metal electrodes. Various electrolyte additives, including fluoroethylene carbonate, perfluoroethylene triphosphate, vinylene carbonate, and SEI-forming salts (e.g., LiNO3, Li3PO4, LiFSI), have gained significant popularity in 82
DOI: 10.1021/acs.accounts.7b00484 Acc. Chem. Res. 2018, 51, 80−88
Article
Accounts of Chemical Research
Figure 2. (a) Scheme of the artificial Al2O3 interphase prepared by ALD filling voids between the electrode and the solid-state electrolyte. Reproduced with permission from ref 19. Copyright 2017 Nature Publishing Group. (b) (left) Optical images of 14 nm ALD Al2O3-protected (top) and unprotected (bottom) Li metal foil exposed to air for 20 h. (right) Cycling performance Li−S cells with and without ALD-protected Li anodes. Reproduced from ref 16. Copyright 2015 American Chemical Society. (c) SEM and TEM images of a CNT−SiC sheet formed by the CVD method and the cycling performance of the materials at 100 mA/g. Reproduced with permission from ref 17. Copyright 2013 Wiley. (d) Schematic and SEM images of the preparation of a patterned LiF-coated Li metal anode and Li deposition on a patterned LiF-coated Li metal anode. Reproduced with permission from ref 18. Copyright 2017 Royal Society of Chemistry.
to protect the Li anode from parasitic reactions with soluble polysulfide species in Li/S batteries (Figure 2b).16 Selfterminated CVD methods have likewise been used by Fu et al.17 to create uniform coatings of Si on carbon nanotube (CNT) sheets that show good cycling stability (Figure 2c). Lu and co-workers18 reported a RF magnetron sputtering procedure that can be used to create uniform LiF coatings on Li in a range of thicknesses at ambient temperature. The authors observed that at a LiF thickness of around 150 nm, both the shape and peak-to-peak voltage profiles obtained in galvanostatic cycling of LiF-protected Li anodes in liquid electrolytes exhibit long-term stability (Figure 2d), testifying to the importance of LiF as a reduced salt in spontaneously formed SEIs. Building upon these approaches, Kim et al.20 proposed a facile Langmuir−Blodgett scooping (LBS) method that can be used to create uniform and ordered coatings of preformed nanoparticles and polymers on both the electrode and separator. The method relies on Marangoni stresses produced by surface tension gradients at liquid−gas interfaces to rapidly create highly ordered coatings of SiO2, TiO2, graphene oxide, and polymeric particles that can be transferred to metal electrodes by dip-coating under ambient conditions. Taking advantage of the softness of the most reactive of the
carbon coating has similar traits, its low capacity and high electronic conductivity means that it can be plated by electrodeposited Li or Na at high current densities, which defeats the purpose of the coating (Figure 1e).15 Appropriate steps are nonetheless required to minimize metallic SEI thicknesses to lower the interfacial resistivity and improve the ability of the SEI to flex to accommodate volume changes. 3.2. Creating Artificial SEIs on Reactive Metal Electrodes
Physical and chemical approaches for creating artificial SEIs in any of the above designs have been developed in tandem in recent years. Anode coatings based on functionalized ceramics, nanoparticles (SiO2, Al2O3, TiO2, carbon), polymers, and salts introduced as discrete phases in liquid electrolytes to protect Li and Na anodes as well as to impart stable artificial SEIs on Si and Sn particles used in composite electrodes have been reported by several authors.10 Atomic layer deposition (ALD), chemical vapor deposition (CVD), and plasma vapor deposition (PVD, or radiofrequency (RF) magnetron sputtering) have all been applied to create void-free coatings on length scales ranging from 10 μm to 10 M salts in the solvent, fundamental properties of the electrolyte can be altered, thus providing attractive electrochemical stability and Coulombic efficiency.36 A similar effect can be observed in concentrated aqueous electrolytes, for which a voltage stability window comparable to aprotic electrolytes is reported to enable aqueous lithium ion batteries (Figure 3d).37
milieu of sulfides, oxides, carbonates, and other inorganics that form the SEI have likewise been reported to produce high overpotentials and low cycling efficiencies in cells utilizing SSEs as a result of solid-state reactions.29 Even in the absence of volume changes in the electrode, it will evolve with time and in space. Thus, a well-defined SEI at the initial stage can quickly deteriorate by physical rupture or fragmentation of the electrode.30 In other cases, Li anodes have been reported to nucleate dendrites that easily proliferate in SSEs by growing along grain boundaries where the conductivity is the highest.31 SSEs that are able to stretch and shrink with the electrode are a requirement for good interfacial contact and transport. Recent advances in gel/solid-state electrolytes (GEs/SSEs) provide a path toward electrolytes that can meet this need while at the same time addressing other challenges associated with leak-free, nonflammable, and safe rechargeable batteries.26 A continuous network of liquid electrolyte in gels requires compliant polymer frameworks with an open structure, which provide open pathways for dendrite proliferation in high-energy alkali-metalbased batteries. In contrast, solid-state polymer electrolytes offer multiple advantages, including weight, cost, toughness, and manufacturability, but rarely achieve high-enough roomtemperature ionic conductivity and high-temperature stability to elicit serious consideration for all solid-state batteries, even in the simplest lithium ion chemistries.25 Cross-linked polymer electrolytes that may also be used as electrode binders have recently been reported (Figure 3a,b).32,33 In this context, crosslinking is also thought to provide mechanical toughness to accommodate volume changes at the metal electrode. The poor oxidative stability of the most commonly used polymers, including poly(ethylene glycol), nonetheless poses significant barriers to applications in which the metal anode is paired with a high-voltage cathode.26 Electrochemical interfaces in SSEs are not limited only to the junction between the electrode and electrolyte but also include boundaries between micro- and nanoscale phases in the bulk SSE. Though less studied, the difficulty of maintaining good ion transport across these interfaces is a critical factor ensuring steady and fast ion transport. This feature is particularly critical in polymer electrolytes composed of heterogeneous phases, as in block copolymer electrolytes, where restricted transport across the phase boundaries can produce large resistances if the electrolytes are improperly prepared.34 Careful tuning of these structures nonetheless offers a flexible platform to prepare various polymer SSEs with balanced mechanical strength and transport properties. Short- and long-range ordering of the polymer chains raises novel possibilities for anisotropic ion transport in nanoscopic ion-conducting domains that can regulate ion motions.27 Accompanying the development and understanding of the interphase at SSEs, promising advances have also been recently achieved on more conventionally applicable liquid electrolytes for metal-based rechargeable batteries. This requires careful tuning of the liquid composition that forms a desirable SEI with the electrode without compromising fast bulk ion transport. It has been reported that a trisalt-based electrolyte with a small amount of LiPF6 generates a conductive and electrochemically stable SEI in the Li−NCM battery (Figure 3c).35 Further investigation of the as-formed SEI revealed a dominant polycarbonate composition that is believed to be mechanically robust as well. A recent innovative class of efforts to design electrolytes involve blending of a large quantity of highly soluble salts in the solvent to form superconcentrated
4. DESIGNING ELECTRODES FOR STABILITY An electrochemical cell is conventionally envisioned as a set of nominally 2D electrode layers intersected by electrolyte. However, interphases formed at the electrode−electrolyte interface are dominantly 3D structures. To realize electrodes with appropriate structural characteristics that can sustain morphological and structural changes in a reactive metal electrode during macroscopic expansion and contraction in the charging and discharging processes, new approaches are required. In this regard, multidimensional carbon hosts assembled at the electrode−electrolyte interface have been proposed, which can be used as 3D containers for electrodeposition of reactive metals and at the same time ensure good electrical contact, even after sections of the metal break away to form orphaned phases (Figure 4a).21,38 In addition to protecting the metal anode from direct contact with electrolyte, the approach lowers the local current density, which increases the compactness of electrodeposits during battery recharge. Additionally, unlike the planar ASEIs discussed in the previous sections, the 3D design mitigates volume change. Although the specific capacity of the electrode is reduced, the hybrid electrode can have significantly enhanced performance in comparison to a 2D lithium foil. Forming the 3D host using metals such as Sn−C or Si−C composites offers a potential solution to this problem, but these approaches are still under development.39 The combination of a metal anode and high-energy conversion cathode, including elemental sulfur or oxygen, leads to additional challenges when conversion products dissolve in the electrolyte. Interlayers between a sulfur cathode and the separator have been reported to provide a powerful strategy for absorbing and reutilizing soluble polysulfides by serving as upper current collectors.40 The approach opens a path to Li−S cells with high sulfur loadings (loadings up to 32 mg/cm2 have already been demonstrated).41,42 Interlayer coatings created using the LBS method discussed earlier are reported to exhibit enhanced stability by formation of a dense and close-packed structure (Figure 4b).20 5. CONCLUSION AND OUTLOOK Results from experimental, continuum theoretical, and computational analyses of interfacial transport have in the past few years identified fundamental connections between the composition of the solid−electrolyte interphase (SEI) at the metal−electrolyte interface and stability. On the basis of such studies, artificial SEIs (ASEIs) based on single- or near-singleion conductors formed from polymers, ionic liquids, ceramics, nanoparticles, salts, and their combinations have started to appear. ASEIs based on materials that completely block transport of matter (electrolyte, ions, and electrons) have been known for some time, but the adoption of state-of-the-art deposition tools from the semiconductor industry, including atomic layer deposition, RF magnetron sputtering, and Langmuir−Blodgett scooping, now makes it possible to achieve 86
DOI: 10.1021/acs.accounts.7b00484 Acc. Chem. Res. 2018, 51, 80−88
Article
Accounts of Chemical Research
supported by the King Abdullah University of Science and Technology (KAUST) through Award KUSC1-018-02.