Regulating Key Variables and Visualizing Lithium Dendrite Growth: An

Subscriber access provided by UNIV OF LOUISIANA

Article

Regulating Key Variables and Visualizing Lithium Dendrite Growth: an Operando X-ray Study Seung-Ho Yu, Xin Huang, Joel D. Brock, and Héctor D. Abruña J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b13297 • Publication Date (Web): 07 May 2019 Downloaded from http://pubs.acs.org on May 7, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Regulating Key Variables and Visualizing Lithium Dendrite Growth: an Operando X-ray Study Seung-Ho Yu,†,┴,‡ Xin Huang,║,§,‡, Joel D. Brock,*,║,§, and Héctor D. Abruña*,†

†

Department of Chemistry and Chemical Biology, Cornell University, Ithaca, New York

14853, USA ┴

Department of Chemical and Biological Engineering, Korea University, 145 Anam-ro,

Seongbuk-gu, Seoul 02841, Republic of Korea ║School

of Applied and Engineering Physics, Cornell University, Ithaca, New York

14853, USA §

Cornell High Energy Synchrotron Source, Cornell University, Ithaca, NY 14853, USA

‡

These authors contributed equally to this work.

1

ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

Abstract: Although Li metal has long been considered to be the ideal anode material for Li rechargeable batteries, our limited understanding of the complex mechanism of Li plating has hindered the widespread deployment of Li metal anodes. Therefore, operando studies are required to unambiguously reveal the complex mechanistic steps involved. In this study, we employed synchrotron-based X-ray imaging methods to visualize the evolution of Li plating/stripping under operando and, more importantly, practical conditions for battery operation, providing detailed insights into morphology evolution during Li plating. The effects of critical battery operating parameters, including concentration of Li salts, current density, ionic strength, various electrolytes and additives, on Li plating/stripping have been studied. The delicate interplay of these conditions on the resulting Li-metal morphology has been characterized for the first time.

2


Page 3 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


■ INTRODUCTION In Li rechargeable batteries, Li metal has been considered to be the “holy grail” of anode materials due to its remarkably high theoretical specific capacity of 3,860 mAh/g and low reaction voltage.1-4 Some of the most promising post-Li ion battery systems, such as Li-O2 and Li-S batteries,2-5 are predicated on a Li metal anode. However, several barriers to widespread adoption remain, including low Coulombic efficiency, poor cycle life and, more importantly, the growth of Li dendrites during plating, which can lead to thermal runaway and catastrophic failure.1,6 Tremendous efforts have been made to suppress Li dendrite growth. These efforts, in turn, have led to innovations and improvements in virtually every component of a Li metal battery. For example, Li dendrite growth can be suppressed, at least in part, by adjusting the Li salt concentration in the electrolyte (i.e., “solvent-in-salt”)7 or through additives, such as CsPF6,8,9 fluoroethylene carbonate (FEC),10 vinylene carbonate,11 and LiNO3 with polysulfides.12 In addition, novel electrode designs (i.e., patterning of Li metal13,14/employing a 3D current collector15,16) can suppress dendritic growth by precisely controlling the current density distribution. Similarly, dendritic growth can be suppressed by manipulating the electrode-electrolyte interfacial structure. Examples include mechanically stable nanostructures17-20 and artificial solid electrolyte interface (SEI) layers, such as Li3N,21 LiF,22 Li3PO423 and Li-rich alloys.24 It has recently been shown that nanometer sized “seeds” (e.g., Ag nanoparticles25 and nanodiamonds26) can guide Li deposition to achieve homogeneous growth. Finally, a variety of analytical techniques have been employed to unravel the mechanistic details of Li dendrite growth, including X-ray imaging,27-29 optical and electron microscopy,30-36 nuclear magnetic 3



Page 4 of 26

resonance spectroscopy/magnetic resonance imaging,37-39 plasmonic monitoring,40 and Raman scattering microscopy.41 Despite these considerable efforts, only a limited number of studies have been performed under practical operating conditions (current density >0.5 mA/cm2 and areal capacity >3 mAh/cm2) and eliminating Li dendrite growth in Li metal batteries under practical operating conditions remains elusive.1,6,30 Here, we address the critical need for a systematic study of lithium electrodeposition under practical battery conditions in terms of current density and areal capacity by using operando X-ray imaging (XRI) to visualize the plating and stripping of Li dendrites. We clearly observe the dynamics of dendritic growth and establish how the morphology of the dendrite is determined by the separator, ionic concentration, current density, electrolytes and additives.

■ RESULTS AND DISCUSSION The morphological evolution of plated Li on a bulk Li metal can be directly observed using operando synchrotron-based X-ray imaging techniques with a resolution of 2 µm, due to the difference in the X-ray absorption of Li and the electrolyte. As shown in Figure 1a and Movie S1, the morphology of the plated Li presents a ramified structure formed by a very large number of omnidirectionally growing small Li branches. This type of morphology is often referred to as mossy-like Li. The evolution of individual Li branches, with a diameter of less than 10 μm, is distinctly resolved (Figures 1b and S1; a selected region, white dotted box, from Figure 1a). Using X-ray absorption, we are able to quantitatively characterize the evolution of the morphological details of Li plating, such as the solidity/porosity of plated Li (see Figure S2 for details). Figure 1c 4


Page 5 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


presents the solidity (the ratio of the volume of plated Li to the entire volume) of plated Li within various regions (for example, 50-100 μm represents the region between 50 μm and 100 μm away from the bulk Li, see Figure S4-S7 for details). The figure shows that the plated Li has an upper packing limit (ca. 10% solidity under the experimental conditions employed) above which Li plating is not favored. In order to systematically study and control the factors that affect Li plating, precise control of the current density is required. Since Li metal anodes do not require foreign host materials and the plating reactions take place only on the surface, controlling the areal current density is critical (not the mass-normalized current density, mA/g). To control the areal current density, we designed a Li holder which we refer to as the “V-slot holder” (see Figure 2a and Figure S8 for experimental setup and Figure 2b for optical microscopy images of Li-unloaded and loaded V-slot holder). The V-slot holder defines the surface area of Li, enabling precise control of the areal current density. The black arrow in Figure 2b (right) indicates the X-ray propagation direction. In our case, we obtain a flat Li electrode with a surface area of 0.8×2.0 mm2, with the distance between electrodes adjustable from about 200 to 500 µm. With this electrode, the effects of critical operating parameters, such as separator, current density, Li salt concentration, and additives/coatings on the Li plating/stripping process were systematically and quantitatively characterized. The separator, which minimizes the distance between electrodes while preventing a short circuit, is an essential component of every battery. For Li metal anodes, understanding the role of the separator on Li plating is critical. However, previous Li plating studies have rarely reported on the effects of the separator during Li plating, 5



Page 6 of 26

since Li plating under the separator is extremely challenging to observe, not to mention to quantify, under operando conditions. However, with our setup we are able to observe Li plating beneath the separator (Figure 2c and Movie S2). From Figures 2c-2f, Li was first plated in one direction for 4 hours and then stripping/plating cycled at a rate of 10 mA/cm2 for 1hr (see Figure 2d for the voltage profile). A high current density of 10 mA/cm2 and a capacity of up to 40 mAh/cm2 were chosen to enable observing Li plating/stripping under harsh experimental conditions, relevant to practical batteries. Figure 2c (right) presents the morphology of plated Li before and after the initial 4 hr plating (see Figure 2e for the evolution of plated Li in a selected region). The triple layered structure of the separator can be clearly resolved. During plating, the plated Li first grows and uniformly fills the whole space under the separator (see Figure S9 for details). Then as the plating continuous, the pressure below the separator builds up (see detailed discussion in Figure S10), pushing the separator firmly against the other electrode (0-2 hr). Later (2-4 hr), Li plating takes place predominantly near the separator, leading to a much denser layer of plated Li near the separator, which is more clearly seen in the solidity plot (Figure 2f). The solidity of this denser Li layer is significantly higher (~8 times) than without a separator (The solidity, even near the bulk Li, increases. See detailed discussion in Figure S12). During additional cycles, plated Li dissolution takes place mainly in the dense Li layer near the separator (4-7.5 hr). Our results demonstrate that the separator can significantly affect Li plating/stripping. A much denser Li layer is formed close to the separator, where most Li plating/stripping takes place in the subsequent cycles, continuously stressing the separator. As a result, the separator is continuously compressed and stretched between the electrodes during 6


Page 7 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


cycling, which might deform the separator, and eventually lead to short circuiting of the battery (~7.5 hr). Moreover, by regulating the operating conditions of Li plating, the morphology of plated Li could be severely altered in not only the solidity, but also the main morphological features, from the commonly-observed mossy-like morphology to dendrite-like, cactus-like and even bulk-like Li. Figures 3a and 3b show, respectively, the morphology and voltage profiles of Li plated at different concentrations of LiPF6 in EC/DEC (2.0-0.05 M, see Figure S13 for the morphology in 2.0 M) under otherwise identical plating conditions (up to 4 mAh/cm2 at 10 mA/cm2). In both morphology and voltage profiles, a striking difference is observed between 0.5 and 0.1 M. At 1.0 and 0.5 M, the mossy-like morphology is observed (Figure 3a), and the voltage first reaches a maximum, then gradually decreases (Figure 3b). In contrast, at lower concentrations (0.1 and 0.05 M), the amount of plated Li is much lower for the same amount of charge, suggesting that other parasitic reactions become increasingly important at lower concentrations. Figures 3c and 3d (Movie S3) compare the evolution of the morphology of plated Li in 0.5 and 0.1 M electrolytes (see Figure S14 for 0.1 M and Figure S15 for 0.05 M). The morphology of the plated Li in the 0.1 M electrolyte is qualitatively different from the mossy-like structure. Dense Li clusters are formed, and they gradually expand, turning into “cactus” shaped structures (Figure S16). In the voltage profile, the overpotentials at 0.1M and 0.05 M are significantly higher and increase monotonically during plating. The high overpotential, as well as the lower reductive stability of the lower electrolyte concentrations, may promote the decomposition of the electrolyte and enhance the formation of the SEI layer. Based on the morphology and 7



Page 8 of 26

voltage profile, we hypothesize that the reactions leading to SEI layer formation dominate at 0.1 and 0.05 M. To test this hypothesis, we characterized the electrode using X-ray photoelectron spectroscopy (XPS) (Figures 3e, 3f, S16 and S17). The XPS chamber was directly connected to a glove box so that the electrode was never exposed to air during the entire process. From the Li 1s spectra, five Li species are present: Li metal (~52.4 eV), Li2O (~53.6 eV), ROCO2Li (~54.5 eV), Li2CO3 (~55.2 eV) and LiF (~56.1 eV). The Li metal peak (~52.4 eV) is from the plated/bulk Li. Li2O peak (~53.6 eV) arises mainly from the native oxide on the surface of bulk Li. The peaks from other three species (and possibly some of the Li2O) are from the SEI layer. This interpretation is supported by the observation that the ratio of the other three components (ROCO2Li, Li2CO3 and LiF) to Li metal/Li2O significantly decreases after surface etching, suggesting that most of the three components are present at the surface. Comparing XPS spectra in 0.5 and 0.1 M, the ratio of Li metal peak (as well as Li2O peak) to the others is much lower when using the 0.1 M electrolyte, suggesting the formation of a thicker SEI layer at 0.1 M. The SEI layer formed at lower electrolyte concentrations can lead to the formation of dense and cactus-like Li clusters, and although somewhat speculative in our part, we propose two possible ways (Figure 3d). In the first, a thicker SEI layer, formed at lower electrolyte concentrations, might be more able to mechanically withstand the volumetric expansion of Li plating, thereby suppressing the porosity of the plated Li. In the second, the SEI layer formed at a 0.1 M concentration is thicker and more LiF-dominant. It has been reported that Li ion diffusion in LiF is much slower than that in Li2CO3/Li2O,4,42 so a slower Li diffusion in the SEI layer at 0.1 M might lead to dense Li formation. 8


Page 9 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


It is interesting that at 2.0 M LiPF6-EC/DEC, a mixed morphology of plated Li was observed (Figure S13). In some regions, Li was plated in a mossy-like shape. Overall, the Li was plated uniformly onto the bulk Li, forming a dense Li layer, which appeared as an increase in the thickness of the bulk Li. It is noteworthy that some electrolytes with high concentrated salts (e.g. 4.0 or 5.0 M LiFSI-DME) have been reported as very effective electrolytes for Li anodes, but at least in LiPF6-EC/DEC electrolyte system, 2.0 M LiPF6-EC/DEC is very viscous, like a gel, so the overpotential is higher than in 0.5 M or 1.0 M LiPF6-EC/DEC. Another critical factor influencing Li plating is the current density. The morphological evolution was characterized at four different current densities, 0.5, 1.0, 5.0 and 10.0 mA/cm2. These particular values were selected because current densities ranging from 0.5 to 10 mA/cm2 (and areal capacities above 3 mAh/cm2) are expected to be the most relevant in practical applications.1,30,43 The voltage profiles (inset of Figure 4a) demonstrate that higher current densities are correlated with higher overpotentials (as would be anticipated), but maintain similar trends with time. The voltage first reaches a maximum, then gradually decreases. Figure 4a presents the evolution of Li plating at 0.5 and 10.0 mA/cm2, respectively (See Figures S19-S22 and Movie S4 for various current densities in larger scale). Within this range of current densities, Li plating yields the mossy-like morphology. However, the individual Li branches grown at 0.5 mA/cm2 are larger than those grown at 10.0 mA/cm2 while the density of Li branches is smaller. Some have proposed that the thicker Li branches are due to the formation of a thicker SEI layer, causing a stronger confinement of the plating Li.4,6 However, the XPS results demonstrate that the ratio of the three components of the SEI 9



Page 10 of 26

layer (ROCO2Li, Li2CO3 and LiF) to Li metal/Li2O at 0.5 mA/cm2 is lower than that at 10 mA/cm2 (see Figures 4b and 4c), suggesting that the SEI layer grown at 0.5 mA/cm2 is thinner than the SEI layer grown at 10 mA/cm2. Thus, the larger size of the Li branches obtained at low current densities is not mainly caused by the confinement of the SEI layer (see Figure S23 for details). Our results support a competing hypothesis, which is that the change in size of the individual Li branches is due to higher overpotentials producing a higher number density of nuclei, leading to smaller nuclei at the same charge.44 Also, a similar relationship between the size/number of nuclei and the current density has been previously reported in galvanostatic nucleation systems, including Li-S batteries.45 A great deal of effort has gone into exploring the effects of various electrolytes, additives, and coatings on Li metal anodes.8-12,46 To characterize the effects of varying the electrolyte, on the morphology of the plated Li, Li plating in various electrolytes, under otherwise identical conditions, was conducted (see Figures 5 and S25-S28 for carbonate-based electrolytes and Figures S29, S30 and Movie S5 for ether-based electrolytes). In LiPF6−PC, one of the most commonly used carbonate-based electrolytes for battery studies, the deposited Li had significantly thicker dendritic branches (Figure 5a), relative to those in LiPF6−EC/DEC. Thus, LiPF6−PC is an ideal “base” electrolyte that can be used to compare the efficacy of various additives/coatings in suppressing dendritic Li formation. An effective idea, used in carbonate-based electrolytes with LiPF6 salt, is to promote the formation of LiF in the SEI layer.9,10,46-48 One approach is to employ additives, such as FEC,10 CsPF69 or trace water.47 Figures 5b and 5c (Movie S6) show the morphology of plated Li with additives consisting of 5 10


Page 11 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


wt% FEC and 0.05 M CsPF6 in 1.0 M LiPF6−PC, respectively. Adding FEC causes the sharp edges of the individual Li branches to turn into “blurry” clusters, exhibiting a smoother surface. (Note that, unlike dendritic Li, the smoother surface of plated Li makes it difficult to study with ex situ techniques. For example, in scanning electron microscopy, the morphology of plated Li appears similar to the surface of bulk Li. However, using our X-ray method, we can clearly monitor the morphlogical evolution during Li plating.) Using CsPF6 as an additive produces an even stronger suppression of Li growth. At low current densities (

Regulating Key Variables and Visualizing Lithium Dendrite Growth: An

Recommend Documents