Role of Inorganic Surface Layer on Solid Electrolyte Interphase

Subscriber access provided by Nottingham Trent University

Surfaces, Interfaces, and Applications

Role of inorganic surface layer on solid electrolyte interphase evolution at Li-metal anodes Ethan P Kamphaus, Stefany Angarita-Gomez, Xueping Qin, Minhua Shao, Mark H. Engelhard, Karl T. Mueller, Vijayakumar Murugesan, and Perla B. Balbuena ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.9b07587 • Publication Date (Web): 01 Aug 2019 Downloaded from pubs.acs.org on August 3, 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 18 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

ACS Applied Materials & Interfaces

Role of inorganic surface layer on solid electrolyte interphase evolution at Li-metal anodes Ethan P. Kamphaus1, Stefany Angarita-Gomez1, Xueping Qin1,2, Minhua Shao2, Mark Engelhard3, Karl T. Mueller 3,4, Vijayakumar Murugesan*3,4, and Perla B. Balbuena*1 1

Department of Chemical Engineering, Texas A&M University, United States, 77843 Department of Chemical and Biological Engineering, HKUST, Clear Water Bay, Kowloon, Hong Kong 3 Pacific Northwest National Laboratory, Richland, WA 99354, United States 4 Joint Center for Energy Storage Research (JCESR), Lemont, Illinois 60439, United States 2

ABSTRACT: Lithium metal is an ideal anode for rechargeable lithium battery technology. However, the extreme reactivity of Li-metal with electrolytes leads to solid electrolyte interphase (SEI) layers that often impede Li+ transport across interfaces. The challenge is to predict the chemical, structural and topographical heterogeneity of SEI layers arising from a multitude of interfacial constituents. Traditionally the pathways and products of electrolyte decomposition processes were analyzed with the basic and simplifying presumption of an initial pristine Li-metal surface. However, ubiquitous inorganic passivation layers on Li-metal can reduce electronic charge transfer to the electrolyte and significantly alter the SEI layer evolution. In this study, we analyzed the effect of nanometric Li 2 O, LiOH and Li 2 CO 3 as surface passivation layers on the interfacial reactivity of Li-metal using ab initio molecular dynamics (AIMD) calculations and Xray photoelectron spectroscopy (XPS) measurements. These nanometric layers impede the electronic charge transfer to the electrolyte and thereby provide some degree of passivation (compared to pristine lithium metal) by altering the redox based decomposition process. The Li 2 O, LiOH and Li 2 CO 3 layers admit varying levels of electron transfer from a Li-metal slab and subsequent storage of the electronic charges within their structures. As a result, their ability to transfer electrons to the electrolyte molecules, as well as the extent of decomposition of bis(trifluoromethanesulfonyl)imide (TFSI) anions, is significantly reduced compared to similar processes on pristine Li-metal. The XPS experiments revealed that when Li 2 O is the major component on the altered surface, LiF phases formed to a greater extent. The presence of a dominant LiOH layer, however, results in enhanced sulfur decomposition processes. From AIMD studies, these observations can be explained based on the calculated quantities of electronic charge transfer found for each of the passivating films. KEYWORDS: Li-battery, solid electrolyte interphase (SEI), lithium metal anode, X-ray photoelectron spectroscopy (XPS), Density functional theory (DFT).

1. Introduction As the most popular energy storage device, the Li-ion battery (LIB), is a critical component of modern technologies such as cellular phones,

laptops and long range electric vehicles.1-2 Developed over many decades, the operational performance of LIB batteries is approaching the theoretical thermodynamic limit of energy density (~300 W h /kg.)2-3 With increasing

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces 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

demand for power and capacity, major research efforts are focused on building the next generation of battery technologies. Limetal is a promising anode due to its ultrahigh capacity (3,860 mAh g−1) and the very low standard negative electrochemical potential (−3.04 V).3-5 However, the high reactivity and harsh dendrite growth of Li-metal anodes result in low cycling efficiency and severe safety concerns.4 In particular, spontaneous reactivity of the Li-metal with electrolyte constituents (such as solvent and counter anions) leads to a highly complex solid electrolyte interphase (SEI) layer, consisting of a wide range of inorganic and organic components.5 The most common pathway for the SEI layer formation is the reduction process activated by the low electronegativity of Li-metal, where electrolyte constituents decompose to metastable products and initiate cascading phase formation and possible phase segregations. The critical functionalities (such as electrical, mechanical and chemical properties) of an SEI layer originates from the individual components and their collective responses as part of the composite phase.3, 6 As the chemical reactions driven by electron transfer is the basis of SEI layer evolution, first principles methods are widely used to study computationally the Li-metal and electrolyte interfacial regime.7-8 Similarly, many spectroscopic and microscopic tools have been used to probe the interface.9-11 Despite numerous theoretical and experimental attempts to decode SEI layers on Li-metal, a predictive understanding of structural and chemical inhomogeneity evolution remains elusive. The underlying challenge is the multitude of reactants and possible composition(s) arising

Page 2 of 18

from the diverse surface chemistry of Li-metal leading to the widely observed inhomogeneity of SEI layers. For example, as a reactive metal, the Li surface contains ubiquitous passivating layers of inorganic components such as Li 2 O, Li 2 CO 3 , LiOH, Li 2 S, and LiF, along with organic impurities depending on manufacturing and storage conditions. Evidently, the reaction pathways and evolved products critically depend on the surface chemistry of the Li-metal, as it dictates the chemical potential and associated thermodynamics of interfacial reactions. Hence, the initial conditions for SEI layer evolution are unique to each Li-metal electrode leading to inhomogeneous compositions and structures. For simplicity, many theoretical and experimental studies make a general assumption of pristine Li-metal surfaces7-8 and focus primarily on the composition of the electrolyte constituents. However, as native inorganic layers are typically poor electronic conductors, their thickness and composition modulate the thermodynamics and kinetics of reduction processes of the electrolyte constituents. Hence, it is critical to delineate the role of common Li-metal surface layers such as Li 2 O, Li 2 CO 3 and LiOH on the electrolyte reduction process. Herein, we report a combined theoretical and experimental analysis of the reaction of dimethoxyethane (DME) solvent and bis(trifluoromethanesulfonyl)imide anion (TFSI) with the most common surface layers Li 2 O, Li 2 CO 3 and LiOH covering the Li-metal surface. With first principles computational techniques, we have modeled the interfacial reactions at a lithium metal slab with a thin inorganic (~ 1nm) layer in the presence of explicit electrolyte molecules. These models,


2

Page 3 of 18 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


which are the first of their kind, have both the Li/surface layer and surface layer/electrolyte interfaces. These results were paired with surface sensitive high resolution X-ray photoelectron spectroscopy (XPS) studies.12 This combined approach provides a comprehensive view of SEI layer evolution, while exploring an atomistic view of the complex surface chemistry of Li-metal. 2. Methodology 2.1 Computational: All calculations were done using first principles methods as implemented in the Vienna Ab-Initio simulation package (VASP.)13-15 The electronic structure calculations were performed using Density functional theory (DFT) projected augmented wave (PAW) pseudopotentials were used for electron-ion interactions and the Perdew–Burke–Ernzerhof generalized gradient approximation functional (PBE-GGA) was used to take dynamic electron correlation into account.16-18 Both ground state optimizations and ab-initio molecular dynamics (AIMD) were used in this work. The dynamics was run with the NVT ensemble using a Nose thermostat with a damping parameter of 0.5 at a temperature of 330 K. Since several different unique surface layers were investigated, the same parameters could not be used in every simulation. Instead, the same methodology was used to ensure results that were converged to a similar level of accuracy to ensure proper comparisons. See Table S1 for a comparison of the parameters used in each simulation. All of the K-point grids were of the Monkhorst-Pack variety used for surface Brillouin zone integration.19 The simulation cells were built by combining a Li-metal slab model with a thick

film of ~1 nanometer of surface layer (LiOH/Li 2 O/Li 2 CO 3 ) followed by pure solvent or electrolyte molecules. The Li slab models used in systems were developed previously by our team.20 We selected the most favorable planes based on surface energy considerations and to minimize the degree of lattice mismatch between the SEI and Limetal. The surface layers were created by first converging computational parameters (cutoff energy and k point mesh) for the bulk systems. These parameters are unique to each system. The bulk structures were then cut by planes chosen to minimize lattice mismatch to create ~1 nm size surface layers. Even if other facets may have lower surface energies when calculated in contact with vacuum, contact with the electrolyte solution can change these preferences; therefore minimizing the interfacial lattice mismatch is a better criteria for nucleation of native films. The thin layer was then placed near the lithium metal. The initial configuration interfacial separation distance was determined by scanning over a range of values and selecting the one that gave the lowest energy. Then the combined Li/surface layer interface was optimized before introducing the electrolyte into the cell. The electrolyte with a composition of 1 M lithium bis(trifluoromethanesulfonyl) imide (Li-TFSI) in DME was incorporated by creating a vacuum layer and placing on the surface layer, followed by solvent molecules using Materials Studio Amorphous Packing tool. Finally, a monolayer of helium atoms was positioned at the bottom of the Li-metal slab to avoid bulk electrolyte interactions due to the periodic boundary conditions. For each of the passivation layers, pure DME solvent and electrolytes with1 M of Li-


3


TfSI in DME were studied. Each of these 6 systems was tested by running ab initio molecular dynamics (AIMD) simulations at 330 K to observe reactions, structural rearrangements, electronic distributions (as given by Bader charges), and any other notable behavior.21-23 The six different systems that were tested for Li 2 O, LiOH, and Li 2 CO 3 are shown respectively in Figures S3, S4 and S5. The shorter timescale of the simulations may be a potential drawback in analyzing these systems. In this work, we mainly focused on interfacial reactions without any external voltage to study the possibility of instantaneous SEI layer evolution when electrolyte meets Li-metal (even before the battery cycling process). These reactions are usually very fast and AIMD allows comparison of the passivation effects of the surface films. Further kinetically limited SEI nucleation or growth events would require other modeling approaches.24 2.2 Experimental Procedure: The as purchased Li foil (Sigma-Aldrich) was mechanically mounted onto a XPS sample holder and transfer into the ultrahigh vacuum chamber (3x10-10 Torr) using an attached recirculated Ar glove box. The Li foil was analyzed by XPS before and after e-beam exposure to 1.5 eV electrons at 20µA for 30 minutes. A Kimball Physics Inc. Model ES015 BaO Cathode was used for the electron source. The e-beam cleaned Li-metal were transferred internally from the XPS chamber to an glove box, without exposing the samples to the atmosphere (see Figure S1 and S2). One half of the surface cleaned Li-metal was dipped in electrolyte solution (1M Li-TFSI in DME) and subsequently transferred to the XPS chamber. For each sample, two spots

Page 4 of 18

were identified that represent electrolyte exposed and non-exposed areas, by monitoring in F1s region of the XPS spectrum for a CF 3 peak (representing TFSI from the electrolyte). These two spots represent the bare Li-metal, then, both before and after electrolyte exposure. Although, the e-beam cleaned sample reveals a pristine Li surface (see SI), the sample transferred to the glove box for electrolyte exposure shows small amounts of oxygen based contaminants such as hydroxide and carbonates. In particular, the bare Li-metal area still shows the presence of passivating layers of Li 2 CO 3 , LiOH as major component along with a minor amount of Li 2 O. Hence, we considered this sample as representative of Li 2 CO 3 and LiOH passivating layers. Repeated attempts to obtain pure Li-metal surfaces were unsuccessful, indicating the extremely reactive nature of Li-metal even within an Ar filled glove box. To establish a pure oxide passivating layer, we exposed the pristine ebeam cleaned Li metal within the XPS chamber to a fixed partial pressure of O 2 . This procedure created Li 2 O as a dominant passivating layer on the Li-metal, which is relatively stable in the Ar-filled glove box environment during electrolyte exposure. Two spots are chosen as described earlier to analyze the before and after electrolyte exposure regimes with a Li 2 O passivating layer. 3. Discussion As we demonstrated in our previous studies7-8, the TFSI anion would decompose completely into small anions and cations due to significant electron transfer from pristine Limetal surfaces without any passivating layer.20


4

Page 5 of 18 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


To understand the roles of Li 2 O, LiOH and Li 2 CO 3 as surface passivating layers on Limetal reactivity, we constructed a thin layer (~1 nm) of each passivation layer sandwiched between (a) the Li-metal and pure DME solvent and (b) an electrolyte solution containing 1 M Li-TFSI in DME. The electronic and structural changes at the two interfaces (i.e., the Li-metal/passivation layer interface and the passivation layer/electrolyte interface) were analyzed to study the interfacial reaction processes. Each of the passivating layers is discussed below in detail. 3.1 Role of Li 2 O surface layer: As a first step we examined pure DME solvent molecules interacting with a thin Li 2 O layer on a Li-metal substrate. As electron transfer is the basis of any solvent decomposition, monitoring the changes in the Bader charges is an effective tool for analyzing the reactivity of DME molecules. The absence of any significant changes in our calculation indicates that the DME molecules are relatively stable at this interface without an applied potential. In our previous work, we established that the DME molecules can undergo decomposition reactions in the electron rich environments (representing an applied potential) of pristine Li-metal yielding C–O bond scission.22 This absence of any DME decomposition indicates that the presence of a Li 2 O layer might reduce the electron availability at the interface with solvent molecules. Although, the bulk Li 2 O has large band gap (~5 eV) and poor electronic conductivity, the band structure of nanometric Li 2 O could be different due to its interfacial interactions with the low electronegativity of the Li-metal slab. Hence, we analyzed the charge density at the Li/Li 2 O interface and found 8 electrons were released from the Li-

metal slab, and the majority of these charges were distributed in the Li 2 O layer. This electron transfer across the Li-Li 2 O interface is observed instantaneously in our AIMD simulations, as shown in Figure 1a. The interfacial arrangement of the Li-Li 2 O interface, where the Li 2 O layer was placed with the oxygen facet (100) exposed to the Limetal, provides a stable conformational structure. This observation is in agreement with natural oxidation processes, where surface metal atoms are active sites for oxide layer growth. Notably, we could argue that the top most Li-metal layer becomes part of the Li 2 O structure by shifting electronic charges into the Li 2 O sub lattice. Almost seamless transition at the solid-solid interfaces allows electrons to easily transfer to the Li 2 O film upon contact. Thus, the simulation starts with some charge already transferred due to the highly favorable conditions at the Li/Li 2 O interface. This change is quantified by examining the density of states (DOS) which shows how the electronic structure is affected. The DOS in Figure S6 shows that the electronic structure of the lithium slab did not significantly change but the DOS for Li 2 O displays quite interesting behavior. In the initial configuration, there is a DOS peak at the Fermi energy which fits the definition of a conductor. After ~10 ps of simulation time, this peak disappears and is replaced with a band gap that is indicative of an insulator. The extensive number of electrons already transferred to the Li 2 O at the beginning of the simulation could be related to the conductorlike band structure of the Li 2 O in the initial configuration. Figure 1b displays the partial charges of the Li and O in Li 2 O layer, allowing for


5


Page 6 of 18

Figure 1. a) Net Bader charges of system components for pure DME with Li/Li2O b) Bader charges of elements in Li2O layer by layer. Layers are numbered as shown in (c.) c) Side view of the initial (top) and final (bottom) configuration showing the complete simulation cell. The middle image displays the atomic density profile corresponding to the top and bottom images. The highlight illustrates the coating region. Color code for atoms: Li: purple; O: red. DME molecules are displayed as a background.

exploration of how the electrons are distributed in the Li 2 O film. The Li 2 O layers are numbered as indicated in Figure 1c. For layer 1, closest to the lithium slab, the oxygen partial charges are slightly more negative than their average value. This demonstrates that the top Li-metal layers rearrange to match the stoichiometry of the Li 2 O layer, causing the Li to form ionic bonds with the oxygen atoms. Despite the charge transfer, most of the initial interfacial structure is retained with the exception of changes to the first layer of the Li-metal and the outer Li layer of the Li 2 O (as shown in Figure 1c). On the other side of the Li 2 O layer, (i.e. at layer 6, the Li 2 O/electrolyte interface), both the oxygen and lithium atoms have relatively electron-rich partial charges when compared to the rest of the layers. In fact, some oxidized Li atoms are pulled from

the outer layer of Li 2 O into the electrolyte by the solvent molecules as part of the Li+ solvation process. To further analyze the structural evolution, we computed the atomic density as a function of the surface layer thickness (i.e. Z direction) that allowed facile structural analysis of the interfaces. The computed atomic density profiles shown in Figure 1c (middle) reveal that despite atomic displacements and electron density changes, overall crystallographic structural order in the Li 2 O phase is mostly preserved. As a next step, we exposed the 1M LiTFSI containing electrolyte system to study the lithium salt decomposition at the Li 2 O surface layer. Similar to the pure DME system (vide infra) the charge is instantaneously transferred from the lithium metal to the Li 2 O. However, unlike the pure DME system, some of the


6

Page 7 of 18 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


Figure 2. a) Snapshot of reactions occurring in Li/Li2O with DME and LiTFSI b) Snapshot of the system at the end of dynamics. c) Net Bader charges for the different system components d) Bader charges of elements in Li2O layer by layer. Color code for atoms: Li: purple; O: red; S: yellow; N: blue; F: light blue; C: grey. DME molecules are displayed as a background.

electrons distributed within the Li 2 O layer are subsequently transferred to the LiTFSI molecules, thus reducing the salt into smaller compounds. The LiTFSI molecule shown in Figure 2 gains 2e- within the first 3 ps by N-S bond breaking leaving a SO 2 CF 3 and a NSO 2 CF 3-2 fragment. No further reactions occur for this system during the extent of this simulation (up to 13 ps.) The initial and final states of the system are displayed in Figure S7. The charges in the Li 2 O film display very similar properties to the model without lithium salt, but the magnitudes of the partial charges are different. For example, at the Li 2 O/electrolyte interface (layer 6) the Li and O atoms have relatively less electron density than in the pure DME system, probably as a result of the electron transfer to the LiTFSI molecules.

The DOS calculations for the LiTFSI system show the same behavior as identified previously with the DME-only simulation (see Figure S6.) Upon immediate creation of the Li 2 O/Li-metal interface, Li 2 O shows conductor-like behavior by accepting electrons from the Li-metal. However, the DOS shows insulator-like behavior after transferring electrons to the LiTFSI molecule (Figures S6 and S8.) Based on these results, we hypothesize that a thin layer of Li 2 O at the surface may not be fully passivating but can act as conduit for electron transfer depending on the interfacial constituents. Aurbach discussed that materials like Li 2 O and LiOH can be nucleophilic themselves, which could cause the reactions that we have observed.25 3.2 LiOH surface layer: LiOH as a thin layer (~1 nm) on Li-metal shows substantially


7


Page 8 of 18

Figure 3. a) Net Bader charges of system components for DME with Li/LiOH b) Bader charges of elements in LiOH layer by layer. The layers are numbered as in Figure 3c. c) Side view of the initial (top) and final (bottom) configurations showing the complete simulation cell. The middle image displays the atomic density profile corresponding to the top and bottom images. The highlight illustrates the passivation layer. Color code for atoms: Li: purple; O: red; H: white. DME molecules are displayed as a background.

different electronic structure evolution (in terms of both magnitude and kinetics) than the Li 2 O layer when exposed to the pure DME solvent. The LiOH layer receives significantly less electron transfer (

Role of Inorganic Surface Layer on Solid Electrolyte Interphase

Recommend Documents