GarnetâPolymer Composite Electrolytes: New Insights on Local Li-Ion

Subscriber access provided by TULANE UNIVERSITY

Article

Garnet-polymer composite electrolytes: new insights on local Liion dynamics and electrodeposition stability with Li metal anodes Jakub Zagórski, Juan Miguel Lopez del Amo, Megan Cordill, Frederic Aguesse, Lucienne Buannic, and Anna Llordes ACS Appl. Energy Mater., Just Accepted Manuscript • DOI: 10.1021/acsaem.8b01850 • Publication Date (Web): 30 Jan 2019 Downloaded from http://pubs.acs.org on January 31, 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 48 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 Energy Materials

Garnet-polymer composite electrolytes: new insights on local Li-ion dynamics and electrodeposition stability with Li metal anodes

Jakub Zagórski1, Juan Miguel López del Amo1, Megan J. Cordill2, Frédéric Aguesse1, Lucienne Buannic1, Anna Llordés1,3*

1

2

CIC Energigune, Parque Tecnológico de Álava, Miñano, Spain.

Erich Schmid Institute of Materials Science, Austrian Academy of Sciences and The Department of Material Physics, Monanuniversität Leoben, Austria. 3

IKERBASQUE, The Basque Foundation for Science. Bilbao, Spain.

KEYWORDS: Composite Electrolyte, Garnet, PEO, Lithium Metal, Solid Electrolyte, Interface, Lithium Dendrite, Solid State NMR, Elastic Modulus, Solid State Battery

ACS Paragon Plus Environment

1

ACS Applied Energy Materials 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 48

ABSTRACT

Ceramic-polymer solid electrolytes, combined with Li metal anodes, hold the promise for safer and more energetically dense battery technologies, as long as key interfacial challenges are fully understood and solved. Here we investigate a Garnet-PEO(LiTFSI) composite electrolyte system, the garnet filler being Li6.55Ga0.15La3Zr2O12 (LLZO) microparticles. A “soft” mechanical milling process ensures good miscibility between the garnet and polymer phases over a wide range of volume fraction (up to 70 vol% garnet). Excellent degree of structural and chemical homogeneity is achieved without degradation nor segregation, even at the local level, as confirmed by solid-state NMR spectroscopy, electron microscopy and gel permeation chromatography. The total Li-ion conductivity of the composites is governed by the polymer matrix, as a consequence of the high interfacial resistance (~104 Ωcm2) between the garnet particles and the PEO(LiTFSI) matrix. However, by using 7Li NMR 2D exchange spectroscopy (ESXY) in the solid state, it is shown that Li ions can locally exchange between the garnet surfaces to the surrounding polymer chains. This dynamic transfer phenomenon, occurring within the composite, seems to play a key role in kinetically stabilizing the interface with Li metal electrode, as observed from galvanostatic cycling and EIS experiments. By comparing a garnet-free PEO electrolyte with a PEO-garnet(10 vol%) composite, the latter shows key performance improvements: although the Li-ion conductivity at 70 ºC slightly decreases from 7.0·10-4 Scm-1, for PEO-LiTFSI, to 4.5·10-4 Scm1

for 10 vol.% LLZO, the composite shows up to 1 order of magnitude lower interfacial

resistance with Li metal electrode (33 cm2 vs 300 cm2), stable Li electrodeposition and no-dendrite formation. In contrast to previously believed, it is demonstrated that these


2

Page 3 of 48 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


improvements are not related to a change of the mechanical behavior but rather to a structural reorganization in the composite followed by local ion dynamics effects at the vicinity of the Li metal interface.

INTRODUCTION

Solid-state battery (SSB) technology holds the promise to be the safest alternative to the conventional Li-ion batteries (LIB) now on the market, as it would replace the flammable liquid electrolyte component. In addition to safety, SSB has the potential to double, and even triple, the energy density of the cell if Li metal is used as anode. This combination of enhanced performances and safe operation is of special interest for electric vehicle (EV) applications1. Li metal has been neglected from LIB for years due to the growth of Li dendrites and other interfacial stability issues when using liquid electrolytes2. SSB technology with stable and mechanically strong solid electrolytes has brought back the interest in a rebirth of Li metal anodes. Li metal is the only anode material combining high capacity, light weight, and lowest reduction potential, thus offering the opportunity to increase energy density when combined with a high voltage cathode in a SSB. However, the low chemical potential and high reactivity of Li metal, combined with its infinite volume change during cycling remain as key challenges 3,4.

As a result, low charge transfer kinetics at the solid anode-electrolyte interface and

especially, cell short-circuit driven by Li dendrite growth have been recently reported causes of failure in SSB5, preventing the progress of the technology towards commercialization. A solution requires research efforts on the solid electrolyte material and its interface dynamics with Li metal anode.


3


Page 4 of 48

Ceramic and polymers represent two main families of solid electrolyte materials. Among them, ceramic oxide Li7La3Zr2O7 (LZO) with cubic garnet structure is the one that holds the most interest for Li metal SSB, due to its intrinsic high ionic conductivity at room temperature (RT) and chemical compatibility with Li metal anodes6,7,8. However, ceramic electrolytes are challenging for battery operation due to poor intimate contact with the electrodes and inhomogeneous Li electrodeposition. The poor interface contact combined with the ceramic brittleness, cause high interfacial resistance, mechanical failure, dendrite growth and short-circuit7,9,10. Soft polymer electrolytes, instead, can provide better interface contact with the electrodes and mechanical stability during battery operation. Polyethylene oxide (PEO)-based solid polymer electrolytes are the most studied and already present in commercial batteries11. Additives or protective layers are typically used in combination with PEO electrolytes to create electrochemically- and kinetically-stable interfaces with Li metal anodes, which is crucial for long cycling12. However, these strategies often fail to prevent dendrite formation, especially at high current densities, causing cell short-circuiting and battery failure13,14. An alternative to ceramics and polymer electrolytes that is recently attracting much interest, is their combination into composite concepts by embedding the ion-conducting hard ceramic particles within the ion-conducting flexible polymer matrix. Ceramic-polymer composites, and in particular, those containing garnet fillers have been widely studied for the last 3 years15,16,17,18,19. However, the investigations reported so far show discrepancies and inconsistent results in terms of the influence of the ceramic filler content in the ionic transport and electrodeposition properties. Some composites show an optimal ceramic content where the Li-ion conductivity is significantly enhanced, while others show a


4

Page 5 of 48 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


monotonic decrease18,19.These differences are not well-understood, nor the detailed impact of the garnet filler on local Li mobility and resulting electrolyte performances. The poor miscibility between garnet and polymers strongly limits accessing to the high level of structural homogeneity needed at multiple length scales for an accurate investigation of the role of garnet fillers. Although some attempts have been reported to understand the local Liion dynamics in PEO-Garnet composites, for instance by using solid state Nuclear Magnetic Resonance Spectroscopy (NMR), these studies have not been supported by electrochemical and impedance tests, thus lacking of the macroscopic response to verify the proposed mechanisms 20,21. A key electrochemical process in SSBs, occurring during charging at the interface between the Li metal electrode and the solid electrolyte, is Li electrodeposition. To avoid the problematic formation of Li dendrites and consequent short-circuits and battery failure, it is of paramount importance that Li is uniformly plated onto the Li metal. However, achieving uniform Li deposition during battery operation is a great scientific challenge, especially at high current densities22. Li electrodeposition is a local-scale interfacial redox phenomenon governed by the kinetics of Li+ diffusion as well as by structural features, which can affect the charge transfer resistance23. Two main theoretical mechanisms have been proposed to explain the formation of Li dendrites. Following the Chazalviel model24, Li dendrite nucleation and propagation is driven by dissimilar transport of cations and anions in the electrolyte component. Anion depletion near the Li electrode is predicted to produce a large electric field, which facilitates electrodeposition on the Li metal surface. This model foresees that electrolytes with higher ionic conductivity and transference number will delay dendrite nucleation. The second model focuses on physical variables such as electrolyte shear


5


Page 6 of 48

modulus, surface tension and roughness which can also effect Li electrodeposition while neglecting the role of the stable Solid Electrolyte Interface (SEI) widely studied in liquid electrolytes 23,25,26,27. For instance, based on this model, it was initially predicted that a dense hard ceramic electrolyte, such as LZO, would block the growth of dendrites even if Li would be unevenly electrodeposited. However, this theory is not well-describing the dendrite formation process, as it has been experimentally observed to be irrespective of the electrolytes’ shear modulus

28.

Even in the case of elastic polymers, characterized by low

shear modulus, Li dendrites and short-circuits also occur

22,27.

Moreover, a ceramic

electrolyte with a high shear modulus cannot avoid the formation of voids at the lithium metal anode during Li stripping in the battery discharge cycle 29. Such loss of physical contact will cause preferential paths of Li deposition, i.e. avoiding the void, driving uneven Li deposition and potential dendrite growth. A variety of ceramic-polymer composite electrolytes have been investigated expecting a beneficial effect of filler introduction on mechanical and electrochemical properties. Previous reports showed an increase in cell resistance during the initial stage, which then decreased over time to zero, indicating short circuit in polymer electrolyte cells caused by dendrite growth30. Other reports highlighted that the interfacial resistance between a lithium metal electrode and polymer electrolyte is increased with the contact period and is reduced by the addition of fillers31,32,30. These results suggest that an interfacial layer might be formed in-situ between Li and the composite polymer electrolyte, being able to suppress Li dendrites only when it is stable and conductive. All in all, and given that solid-state battery technology is still at an early stage of development, the Li electrodeposition and detrimental dendrite growth remain poorly understood. This is especially true for composite electrolytes


6

Page 7 of 48 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, due to presence of ceramic fillers, introduce additional interfaces and possible interactions with mobile Li ions.

Here, we focus on a Garnet-PEO(LiTFSI) composite electrolyte system, the garnet fillers being Li6.55Ga0.15La3Zr2O12 microparticles (LLZO), and investigate the local structure, Li-ion dynamics and mechanical properties in order to unveil their role on the Li electrodeposition mechanism. To facilitate the interpretation and accuracy of the results, highly homogeneous composites have been prepared using a novel non-invasive wet ball milling process. This solvent mixing process enables accessing a wide range of volume fraction, up to 70 vol% ceramic (90 wt%), without phase segregation, a typical issue in ceramic-polymer composites. Selecting the composition of 10 vol% LLZO (31 wt%), the Li-ion transport at multiple length scales is analyzed, especially at the hard-soft intra-interfaces, up to now unexplored in these composite systems. Li electrodeposition dynamics at the Li metal anode have also been investigated. A combination of local/macroscopic characterization techniques, such as, solid-state Nuclear Magnetic Resonance (NMR), Differential Scanning Calorimetry (DSC), Scanning Electron Microscopy (SEM), Electrochemical Impedance Spectroscopy (EIS), galvanostatic cycling and uniaxial tensile tests have been applied to the composite electrolytes providing new insights on the local Li-ion dynamics, Li electrodeposition and mechanical behavior. This knowledge is of great value to progress towards stable Li metal solid interfaces, offering design rules for stable and conductive intrainterfaces in composite electrolytes, both key requirements for the development of SSB technology.


7


Page 8 of 48

EXPERIMENTAL SECTION Material Synthesis Ga-substituted LLZO with unit formula Li6.55Ga0.15La3Zr2O12 (LLZO) was synthesized in 3 g batches using a chelate-gel route with an aqueous solution of citric acid following the method reported in our previous work33,34. Stoichiometric amounts of LiNO3 (> 98 %, Alfa Aesar), La(NO3)3·6H2O(≥99.99 %, Sigma-Aldrich), Zr(C5H7O2)4 (> 98 %, Alfa Aesar) and Ga(NO3)3 (0,02 M solution) were used as starting materials to prepare Ga-substituted LLZO powders. Ga(NO3)3 solution was prepared by microwave digestion of Ga2O3(≥ 99.99 %, SigmaAldrich) with nitric acid (AR grade, Sigma-Aldrich). Precursors were mixed and consequently dissolved in a citric acid (99.0 %, Sigma-Aldrich) solution. The Ga doping level was fixed to 0.15 mol per formula unit to maximize ionic conductivity and 10 wt% Li excess was added to compensate for the Li losses during high temperature treatment. The obtained gel was combusted at 600 oC for 12 h in air and calcined in a tubular furnace at 950 oC for 12 h using dry oxygen atmosphere in order to synthesize the cubic phase. Calcined samples were quickly transferred into a glove box in order to avoid contact with atmospheric moisture, ground and stored in closed vials. For the sintering step, ceramic green-bodies were prepared applying 2 T uniaxial pressing in a 6 mm diameter die. Pellets were sintered in alumina crucibles for 6 h in 1200 °C under dry O2 atmosphere embedded into the mother powder. Surface impurities were removed by polishing with SiC polishing papers in an Ar filled glove box (O2 and H2O levels below 0.5 ppm). X-ray diffraction (XRD) measurement confirmed the crystallinity and cubic structure of LLZO powder and revealed presence of secondary phases, La2Zr2O7 (15 %) and Li2ZrO3 (14 %) (Figure S1) in the powder used for the preparation of the composite series. The presence of such ~30% impurities in LLZO will


8

Page 9 of 48 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


strongly affect the bulk conductivity of a sintered garnet pellet but not necessarily the conductivity of a composite membrane. As shown in Fig 4a, “multiphase LLZO pellet” has 1 order of magnitude lower conductivity than pure LLZO pellet, from 5·10-3 S/cm to 5.4·10-4 S/cm at 70ºC. However, similar values of conductivity were obtained for the composite containing multiphase and the one containing pure LLZO, i.e. 3.26·10-4 and 4.47·10-4 respectively (Figure 4a 10% content, blue and black circles). As it will be later discussed, the ion conductivity in the composite is independent of the purity of the filler.

Taking into account that the LLZO fillers contained secondary phases, the nominal ceramic filler content is higher than the content of high conductive pure cubic LLZO phase. The corrected true LLZO content has been estimated using the quantification from Rietveld analysis (Table S1). Cubic LLZO powder. was also successfully synthesized and used to prepare a composite reference sample.

Preparation of Garnet-PEO(LiTFSI) composite solid electrolyte In this work the series of composite solid electrolyte membranes with various LLZO contents in a poly(ethylene oxide) (PEO) matrix were prepared. Each slurry contained PEO (Sigma Aldrich, MW 5M), bis(trifluoromethane) sulfonimide lithium salt (LiTFSI, Sigma Aldrich, 99.95 %), LLZO and acetonitrile (ACN, 99.8 % Sigma Aldrich) as a solvent. The molar ratio of ethylene oxide monomer unit to lithium ions (EO:Li) was kept constant at 20:1. LLZO content was varied from 10 to 70 vol%. Additional polymer electrolyte samples without LLZO were prepared for comparison purposes. To obtain a composite electrolyte PEO, LiTFSI and LLZO were ball milled (Micro Pulverisette 7 premium, Fritsch) together in anhydrous


9


Page 10 of 48

ACN until obtaining homogeneous slurry. Low number of revolutions per minute (250 rpm), the number of cycles, their length and the pause between them has been chosen to avoid temperature increase and destructive effect on the polymer. For ball milling 45 mL air tight jars and ZrO2 balls were used. Jars were closed in argon atmosphere (O2 < 0.1 ppm, H2O < 0.1 ppm) and a membrane slurry was prepared by ball milling for 30 min. The resultant solution was poured inside a Teflon mold or blade casted and dried until a membrane forms. Use of “one jar” processing allows to eliminate the issue of material losses during transfer and ensure high reproducibility. The membrane was preliminary dried at room temperature and then finally dried in a dynamic vacuum for 12 h in 50 oC before use.

Structural characterization The crystallographic structure and purity of the synthesized powders were characterized by X-ray diffraction (XRD). A Bruker D8 Discovery diffractometer equipped with a Cu source Kα1 (accelerating voltage of 30 kV) in the Bragg−Brentano geometry was used. The data were collected for 2-theta angles from 15o to 80o using argon filled sample holders covered by a Kapton film. Whole Pattern Fitting Structure Refinement (Rietveld Refinement) was used as the method for structural analysis of the materials. The software FullProf 35 was used for Rietveld Refinement to identify and quantify main and secondary crystalline phase components. Morphologies of the LLZO powders and the composite membranes were examined by SEM using FEI Quanta 250. Cross-sections were prepared by mechanical cutting the membranes at room temperature in an argon glove box. Secondary Electron Imaging (SEI) and the Backscattered Electron Detector (BSED) were used to determine composite homogeneity.


10

Page 11 of 48 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


Composition and elemental distribution analysis were performed by Energy- Dispersive Xray spectroscopy (EDX) equipped with an Oxford Instrument detector. The particle size of the garnet LLZO particles were also determined by using laser particle size analyzer Mastersizer 3000E. For the light scattering measurements as-prepared garnet powder and garnet particles separated from composite electrolyte slurry by centrifuging and washing with ACN were suspended in ethanol. For data analysis Mie scattering model36 for non-spherical particles was used. After mechanical mixing, the initial LLZO particle size was reduced from Dv(50)=18.1 m down to Dv(50)= 1.39 m (Figure S2). Gel Permeation Chromatography (GPC) was used to analyze the polymer molecular weight before and after ball milling. Ultrapure water was used as a mobile phase and polyethylene glycol (PEG) as a reference. The sample was dissolved in the same mobile phase and was filtered with a syringe filter before being injected.

Solid state NMR Magic Angle Spinning Nuclear Magnetic Resonance (MAS NMR) spectra were recorded with a Bruker Avance III 500 spectrometer resonating at a 1H frequency νo = 500.24 MHz. Samples were prepared by cutting the membrane in small pieces and packed in 2.5 mm rotors. Samples were spun at an MAS frequency of 20 kHz, except for variable temperature measurements where samples were measured without MAS. 7Li, 1H and

19F

spectra were

referenced to a 0.1M LiCl solution, to a water sample resonating at 4.7 ppm, and LiF at -204 ppm, respectively. 7Li and 1H NMR experiments were performed using single pulse experiments with π/2 pulses of 2.5 and 2.3 μs, respectively. 19F NMR spectra were recorded using an echo pulse sequence with a π/2 and π pulses of 1.7 and 3.4 μs. 2D EXSY (EXchange


11


Page 12 of 48

SpectroscopY) experiments were performed using a standard three-pulse sequence with the mixing times specified in each case. Signals were deconvoluted using the Dimfit37. Thermal properties DSC measurements were performed with a DSC Q2000, TA instruments. The samples were sealed in an aluminum crucible under argon atmosphere and conditioned in 70 oC for 24 h to eliminate the sample thermal history. The samples were cooled down to -80oC and then heated up to 100 oC with a heating rate of 10 oCmin-1. Crystallinity degree was calculated following equation:

𝜒𝑐 =

Δ𝐻 ∙ 100 Δ𝐻𝑐

where: ΔH is enthalpy obtained from DCS curve and ΔHc is melting enthalpy of fully crystalline PEO38.

Mechanical properties Uniaxial tensile straining experiments were performed to evaluate mechanical properties of solid electrolyte membranes. Straining was performed on an MTS Tytron 200 using a displacement rate of 5 µms-1 up to a maximum strain of 20%. Samples were cut from membranes using a scalpel, to a final size dimension of 5 mm x 40 mm, and dried for 9 h at 60 °C in a vacuum furnace (10-6 mbar). Samples were strained directly after removal from the furnace. The engineering stresses were calculated from the recorded load and the crosssectional area (sample width x thickness) and the engineering strains using the recorded displacement and the original gauge length of 20 mm.


12

Page 13 of 48 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


Electrochemical testing The cells with composite electrolytes and ceramic pellets were conditioned at 70 ºC for 24 h before performing electrochemical measurements to ensure good contact between the electrolyte and the electrodes and to remove the crystallinity factor by polymer amorphization. The Li-ion conductivity and area specific resistance (ASR) of the materials was obtained from Electrochemical Impedance Spectroscopy (EIS) using a Solartron 1260 FRA module in the frequency range from 32 MHz to 1 Hz with 50 mV excitation amplitude. For ionic conductivity measurement of the polymer-based electrolytes, symmetric cells were prepared using an 8 mm diameter electrolyte placed inside the 8 mm hole of a 16 mm diameter Kapton (0.0025 mm thick) disc and sandwiched between stainless steel blocking electrodes. For the sintered LLZO pellet, the 6 mm diameter pellets were polished, sandwiched in between two 4 mm diameter Li foils, and pressed at 0.25 T to improve the interfacial contact between the ceramic and the Li metal electrodes. For the polymer-based and sintered ceramic electrolytes, CR2032 coin cells were used with a constant pressure applied by a spring. The conductivity of the compressed powder was measured by pressing ceramic powder in between two stainless steel discs in a Swagelok cell. All the cells were assembled in an argon filled glove box and hermetically closed. These coin cells were measured in the temperature range 273−353 K, where the temperature was controlled inside a Physical Property Measurement System (PPMS). Obtained plots were analyzed with the Zview software (Scribner Associates Inc.) and the appropriate equivalent circuits were fitted. The data were converted into conductivity using the electrolyte thickness and the surface area using the following equation:


13


𝜎=

Page 14 of 48

𝑡 1 ∙ 𝐴 𝑅

where t, A, and R are the thickness, area, and resistance respectively. To investigate the interfacial resistance between the LLZO pellet and the pure PEO(LiTFSI) membrane, a CR2032 type coin cell was assembled with a LLZGO pellet sandwiched between two pure polymer electrolytes. The 10 mm diameter LLZO pellet was immobilized in the 10 mm hole of a 16 mm diameter Kapton film to avoid movement of the pellet during cell assembling and measurement. On both pellet sides, the contact area between the 12 mm diameter PEO(LiTFSI) membranes and LLZO pellet was controlled by a 16 mm Kapton disc with a 6 mm diameter hole. The PEO(LiTFSI) membranes were placed between the Kapton disc and a stainless steel disc. The cell was conditioned at 70 ºC during 24 h to get better contact between the LLZO pellet and the PEO(LiTFSI) membranes. Furthermore, symmetrical cells of 14 mm diameter Li and 16 mm diameter 10 % LLZO and 100% PEO(LiTFSI) electrolytes in CR2032 coin cell were used to evaluate the impedance evolution over time under open circuit conditions at 70 ºC. The same configuration was used to evaluate transference number of mentioned above electrolytes using Bruce & Vincent method39,40. EIS spectrum was recorded fallowed by application of 10 mV potentiostatic polarization to the cell. Current evolution was followed until steady state regime was reached and then polarization was stopped and another impedance spectrum was collected. Transference number was calculated using the following formula: 𝑡𝐿𝑖 =

𝐼𝑆𝑆(𝑉 ― 𝐼0𝑅0) 𝐼0(𝑉 ― 𝐼𝑆𝑆𝑅𝑆𝑆)

where tLi is lithium transference number, V applied potential, I0 and ISS, initial and steady state current, R0 and RSS initial and steady state resistance respectively. DC polarization was


14

Page 15 of 48 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


conducted using a Maccor battery tester on symmetric cells containing 16 mm diameter composite membrane and two 14 mm Li metal discs assembled in an argon filled glove box. DC with constant current density of 100 µAcm-2 was applied for 2 h followed by the same step but with opposite polarity.

RESULTS AND DISCUSSION

Figure 1. Schematic view of the composite electrolyte preparation using a ball milling solvent-based method.

The preparation of homogeneous composite solid electrolyte membranes is illustrated in Figure 1. Cubic Li6.55Ga0.15La3Zr2O12 garnet (Ga-substituted LLZO, hereafter named LLZO) was selected as ceramic filler as it combines high Li-ion conductivity (1.3 mScm-1)33,34 and excellent electrochemical stability with Li metal anode41,42. Homogeneous dispersions were


15


Page 16 of 48

obtained, which after casting and drying, led to self-standing flexible membranes with typical thicknesses of 150-200 µm (thickness can be tuned in the range of 35-250 µm). The homogeneous dispersions were obtained by a “non-invasive” ball milling process, i.e. using soft conditions to avoid polymer degradation, enabling wide range of membrane compositions, from polymer-rich to ceramic-rich composites, i.e. from 10 vol% (31 wt%) to 70 vol% (91 wt%) of ceramic filler (experimental section). Use of air-tight ball milling jars and dry argon atmosphere, protects the moisture-sensitive LLZO and Li salt compounds, avoiding contamination and ensuring phase stability and reproducibility. LLZO preserves the cubic structure after the milling process as shown in Figure S3, and in agreement to previously reported ball-milled Al-doped Li6.75La3Zr1.75Ta0.25O12 garnets 15. Our solvent-based mechanical mixing method allows polymer dissolution, salt dissociation, composite homogenization and ceramic particle size reduction in one step process. LLZO microparticles, extracted from a composite membrane, exhibited particle size distribution Dv(50) equal 1.39 µm (Figure S2). The mixing conditions were accurately tuned (see experimental section) to avoid detrimental polymer degradation caused by excessive mechanical stress and local heating. Gel permeation chromatography (GPC), Fouriertransform infrared spectroscopy (FTIR) and Electrochemical Impedance Spectroscopy (EIS) confirmed no detrimental effect of the milling process on the PEO matrix (Figure S4). Slight molecular weight reduction from 5·106 to 1·106 g·mol-1 was detected by GPC (Figure S5). The relatively small reduction of PEO molecular weight do not affect the ion transport properties of the polymer, as confirmed by the similar ionic conductivity we obtained for PEO(LiTFSI) membranes prepared by ball milling and magnetic stirring (4.2·10-6 Scm-1 and 4.7·10-6 Scm-1 respectively at RT) (Figure S6), which are also in good agreement with state-


16

Page 17 of 48 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


of-art values for PEO18,43,44. The process of mechanical milling does not affect the mechanical behavior of the polymer matrix as self-standing flexible membranes can be obtained. The degree of PEO crystallinity of the garnet-enriched PEO membranes was investigated by Differential Scanning Calorimetry (DSC) (Figure S7). No endothermic peak related to absorbed water was observed by DSC, which is important considering the moisture sensitivity of the garnet and metallic lithium. The PEO melting transition temperatures was reduced in the composites, from 65.2 oC for filler-free PEO(LiTFSI) to 62.6 oC, 61.5 oC and 59.2 oC for composites with 10 vol%, 30 vol% and 50 vol% LLZO respectively. This mp reduction is in agreement with the lower degree of crystallinity in the composites, e.g. from 35 % (for filler-free PEO(LiTFSI)) to 30 % and 12 % for composites with 10 vol% and 50 vol% LLZO respectively. At the highest LLZO loading (70 vol%), the degree of crystallinity was the lowest, with only 10 %, despite the fact that the melting temperature increase slightly relative to sample containing 50 vol% of LLZO (Table S2). In this case, the higher melting temperature can be understood by the need of more energy to melt the physically constricted polymer chains between the closely-packed garnet particles. For all the composites, there is no evident peak broadening for the endothermic melting process, indicating limited effect of microfillers on the PEO melting kinetics

45.

Observed melting

process is relatively slow and should not be limited by the heat transport through ceramic particles even in membranes containing high filler content. Regarding the small effect of garnet filler content on the melting temperature of the composite, this is attributed to the presence of bulky TFSI anion, which already plasticize the polymer to a level that cannot be surpassed by the embedded garnet microparticles18. Our goal, though, was not to plasticize the polymer but to investigate the effect of garnet microfillers on local ion mobility and


17


Page 18 of 48

electrodeposition properties with lithium metal anodes, which are the main limitations of PEO electrolytes.

Figure 2. Morphological characterization of composite solid electrolytes. (a) SEM images of the surface and cross-section of composite membranes with different content of LLZO. A Back-scattered Electron Detector image is shown in Figure S9. Particle size of LLZO is the same in all composites D50=1.39 µm; (b) digital photograph of each.

The developed milling process enables to solve the problem of filler agglomeration (Figure S8, S9) which was commonly reported for other ceramic-rich composite electrolytes (> 20 wt%)46,47,48. The preparation of a stable dispersion of micron-sized particles is more challenging than it is in case of nanosized particles due the tendency to precipitate. Here, the


18

Page 19 of 48 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


phase segregation problem has been solved by tuning the mechanical mixing conditions as well as the viscosity of the dispersion. The membranes obtained show homogeneous morphology, high density and flat surfaces (Figure 2). These characteristics are needed to facilitate physical contact with the electrodes and overcome current limitations of ceramicrich PEO membranes. For instance, relatively low PEO content (46 wt%) was earlier reported to be insufficient to effectively bind the components together into an homogenous membrane16. Our processing method has solved such limitation.

Solid state NMR was used to obtain insight into the local Li ion mobility as well as to analyze the local structure and interactions between LLZO, PEO and LiTFSI in the composite electrolyte. The 7Li solid state NMR spectrum of a 10 vol% LLZO composite shows two different environment of Li ions present in the PEO(LiTFSI) matrix and in the LLZO fillers (Figure 3a). Both 7Li signals can be deconvoluted in two peaks which is in contrast to previous reports49,21 attributing the presence of a third component to an interfacial environment presumably related to partial degradation of the garnet material after exposition to ambient atmosphere. In our case, the lack of a third Li component confirms that the garnet phase has not been passivated or degraded during preparation of the composite membrane. 7Li NMR

chemical shifts in Figure 3a, i.e. the polymer (-1.2 ppm) and garnet (0.8 ppm), are in

good agreement with reported values for PEO50 and pristine LLZO33,34 as well as with our control samples(Figure S10). The different 7Li NMR linewidths of the two peaks can be related to the different frequencies and dimensionality of local fluctuations in the two phases51. The narrower signal of the 7Li NMR component assigned to the PEO(LiTFSI) matrix


19


Page 20 of 48

can, thus, be explained by local polymer motion combined with the Li+ fluctuations51, as well as by the higher dimensionality of Li+ ion diffusion in the polymeric chains as compared to the restricted paths in the rigid garnet framework. The correlation between linewidths in NMR spectra and the frequency of molecular or ionic motions is however not straightforward. Variable temperature measurements combined with relaxometry methods would be necessary to quantify the kinetics52,70, which is out of the scope of this work.

To further investigate the interaction between salt-polymer matrix and garnet filler at atomic level, solid state

19F

and 1H MAS NMR spectra were recorded (Figure 3b). The

19F

NMR signal observed was fitted using a single resonance, indicating good homogeneity in the distribution of anions within the sample. In case of segregation, more than one signal would be expected as previously reported for PEO samples with non-active sepiolite fillers, in which two types of environments were observed 12. The presence of only one 19F NMR signal also indicates that there is no tendency of TFSI- anions to interact with the LLZO garnet surface. Variable temperature 1H and 19F NMR experiments were also performed in order to assess the homogeneity of the dynamics of both the anions and the host PEO polymer matrix (inset of Figure 3b). The evolution of the normalized linewidth (LW) of 1H and

19F

NMR peaks

shows similar temperature dependence, indicating that the mobility of the TFSI- anion is strongly influenced by the dynamics of the polymer chains entanglement (inset of Figure 3b), which confirms that the Li salt is completely dissociated and homogeneously dissolved in the polymeric phase. Thus, confirming that the TFSI- anions interact mostly with polymer chains and not significantly with the garnet´s surface. This hypothesis is in agreement with


20

Page 21 of 48 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


the similar Li+ transference number obtained in the composite and in the garnet-free PEO (Figure S11).


21


Page 22 of 48

Figure 3. Local structure and Li-ion dynamics in a PEO(LiTFSI)-LLZO (10 vol%) composite investigated by solid state NMR spectroscopy (a) 7Li and (b) 19F MAS NMR spectrum of the composite electrolyte. Inset of (b): temperature dependence of the normalized linewidths (NLW) of 1H and

19F

NMR peaks; (c), (d) 7Li 2D Exchange spectroscopy (EXSY) NMR

spectrum with mixing times of 0.0001 s and 0.6 s, respectively; (e) Schematic illustration of the interfacial Li+ exchange between LLZO particles and PEO(LiTFSI) matrix. Inset: PEO with ball-stick representation, where H atoms are in white, C in blue and oxygen in red. In the garnet polyhedral representation, Li ions are shown in grey/white, denoting partial occupation. La and Zr polyhedra are in orange and yellow, respectively, with oxygen in red. TFSI anions are not depicted and PEO polymer is shown as short chains for the sake of clarity.

A key question to be elucidated when combining a polymer and a ceramic material, where both are Li-ion conductors, is if Li-ions are capable to exchange between these phases. Such interfacial Li-ion transfer would be key to enable the contribution of the ceramic fillers in the macroscopic transport properties. Solid state NMR can provide information about such Liion exchange by using multidimensional correlation experiments like in 7Li 2D Exchange spectroscopy (EXSY) NMR experiment53. Figures 3c and 3d display the 7Li NMR EXSY spectra of a composite sample obtained at different mixing times. The presence of significant magnetization at the cross-peak positions (Figure 3d) with a mixing time of 0.6 s indicates that chemical exchange has occurred during this time between lithium ions at the PEO(LiTFSI) and LLZO phases. Obviously, the exchange process between both phases should occur at the interphase and this arises the question whether the exchanged Li+ population involves both bulk reservoirs or if they are restricted to surface sites in the structure. This question would also have clear implications in the Li+ transport mechanism. As it was discussed previously, the accurate deconvolution of the original 7Li NMR spectrum in Figure


22

Page 23 of 48 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


3a, and the well-defined single longitudinal relaxation times obtained for each phase are in agreement with the absence of a significant population of a third phase at the interphase between the polymer and the ceramic phase. This observation, together with the large relative intensity of the cross-peaks in Figure 3d, allows us to conclude that the exchange process observed leads to real exchange between lithium populations at the bulk of both phases. This spontaneous local Li+ transport occurring at the hard-soft interfaces, within the electrolyte, is a promising new finding in the field. If such local exchange process could occur at larger length scales, composites containing ceramic and polymeric ion conductors could potentially generate solid electrolyte materials with Li+ conductivities higher than those obtained for polymer electrolytes only. Although progress in this field will require significant effort in material optimization, this first demonstration may be of value to further develop ceramic-polymer interfaces for high performing solid-state batteries. In addition, 2D EXSY NMR appears as a valuable tool to support the rational design of novel composites offering new opportunities to progress towards controlling macroscopic transport properties. To investigate if this local Li-ion interfacial exchange between garnet and PEO has an effect on the macroscopic transport properties of composites, EIS measurements (Figure S12) were performed for the whole volume fraction range of LLZO, including filler-free PEO(LiTFSI), compressed LLZO garnet powders and sintered pellets for reference (Figure 4a and Experimental section). When increasing the content of LLZO microfiller up to 40 vol%, the total ionic conductivity, measured at 70 ºC, monotonically decreases from 6.97·10-4 Scm-1 for filler-free PEO(LiTFSI) to 4.47·10-4 Scm-1, 3.26·10-4 Scm-1, 2.10·10-4 Scm-1 and 1.42·10-4 Scm-1 for 10 vol%, 20 vol%, 30 vol%, 40 vol% LLZO, respectively. These conductivity values, similar to PEO, indicate that the embedded garnet microparticles do not contribute to the


23


Page 24 of 48

long-range transport, being thus the polymer matrix contributes the most to this process, in agreement with previous reports on other garnet-polymer systems18,19. However, these results are in contrast with other works reporting an increase in ionic conductivity with the incorporation of garnet filler15,16,45. It is worth mentioning that, in our work, the absolute conductivity values of the composites are similar to those previously reported54,19. We confirm the reproducibility and, thus, the negative or positive effect of the fillers will depend on choice of the reference PEO sample, whose value is often understated and not representing state of the art. The slight monotonic decrease observed in Figure 4a can be understood considering the less effective area of the percolative polymer phase in contact with the electrode. Normalizing the conductivity values of the composites by the polymer effective cross-sectional area (details in SI), thus, leads to a less dependent trend on the content of LLZO (Figure 4a, open circle symbols). This smoother trend further suggests that the ion conductivity is dominated by the polymer phase, remaining fairly constant up to 40 vol% LLZO microfiller. For loadings above 40 vol% LLZO, the conductivity suffers a drastic drop of 2-3 orders of magnitude, reaching similar values to those of compressed LLZO powders. At higher filler loadings, the polymer is physically constrained between the ceramic particles, resulting in denser chain entanglement and restricted chain mobility, which further limits the ion mobility55. Li-ion conductivity in the present composites is reduced mainly by a physical phenomenon that reduces the fraction of the percolative polymer phase. As shown by NMR results (Figure 3a,b), there is no evidence of chemical interaction between LLZO, PEO and LiTFSI, which could result in anion immobilization and higher conductivity56,57,58.


24

Page 25 of 48 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


The negligible contribution of garnet fillers is also evidenced when comparing a composite in which LLZO was either 100 % cubic phase or 70 % (Figure 4a, blue and black symbols). Both samples show similar conductivity values, within the measurement error, while the bulk conductivity of the corresponding LLZO pellets is two times lower. Therefore, the negligible difference in conductivity observed for the composite suggests that the transport mechanism of Li+ is independent from the garnet filler composition, as reported earlier13. Arrhenius plots further confirm the major role of the polymer in the transport mechanism (Figure 4b). Comparing a ceramic-free PEO(LiTFSI) with a 10 vol% LLZO composite and a LLZO pellet, the first two show similar temperature dependence as both are dominated by the transport within the polymer. Two regions of different activation energies are observed, separated by the PEO melting point, which is related to the PEO crystalline–to-amorphous phase transition. The conductivity trend shown by the polymer and composite samples are in agreement with the Vogel-Tammann-Fulcher (VTF) relationship59 i.e. ionic transport is facilitated by the highly conductive amorphous regions due to polymer segmental motions. Thus, in both cases the transport mechanism is controlled by the dynamics of the polymeric chains, which can be inhibited either by high loading of garnet microparticles above melting temperature, or by polymer segmental motion suppression below the melting temperature. In summary, the ion-conducting garnet microfillers do not contribute to the transport of Li ions when embedded in a polymer matrix, suggesting that a high interfacial resistance exist between both phases.


25


Page 26 of 48


26

Page 27 of 48 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 4. Macroscopic transport properties of PEO(LiTFSI)-LLZO composite electrolytes (a) Total ion conductivities, at 70 ˚C, of composite electrolytes with different volume fractions of LLZO garnet fillers. Conductivity values correspond to the average of 3 different replicate samples, in which the standard deviation was 5 % for membranes containing 0-40 vol% LLZO and 10 % for 50 and 70 vol% LLZO. Such small dispersion value cannot be appreciated in the plot, denoting the good reproducibility between samples.

Black solid circles

correspond to LLZO “multiphase” fillers , that contain secondary phases (see experimental section and Table S1). Blue symbols correspond to samples in which the LLZO is pure cubic garnet phase. Conductivity values for compressed LLZO powders and sintered LLZO pellets are also shown for reference (square symbols). Black open circles represent values normalized taking into account the effective area of polymer matrix in contact with electrodes (SI); (b) Arrhenius plots of the pure polymer electrolyte and composite electrolyte; (c) Nyquist plots of symmetric cell consisting of LLZO garnet pellet sandwiched between two PEO(LiTFSI) membranes. Inset: schematics of the symmetrical cell.

To estimate the interfacial resistance between PEO(LiTFSI) and LLZO, a layered model system was prepared by contacting a LLZO pellet between two PEO(LiTFSI) membranes in a symmetrical cell configuration (Figure 4c). Assuming equal contribution of both hard-soft LLZO/PEO(LiTFSI) surfaces to the impedance, the LLZO/PEO(LiTFSI) interfacial resistance is extracted by fitting the Nyquist plot and subtracting the total interface resistance by the contribution of two PEO/Li metal interfaces (Figure 4c). A high interfacial resistance of 11.5 kΩ·cm2 is obtained, confirming that the LLZO/PEO interface is a limiting barrier for Li ion transport across the hard-soft interface. This high resistance is in agreement with Langer et al.60, who reported 9 kΩ·cm2 between Al-substituted Li7La3Zr2O12 and PEO-LiClO4. A high interfacial resistance between PEO and LLZO might seem in contradiction with the local Li exchange process observed by NMR. However, a high resistance value is compatible with a slow Li+ exchange interfacial process. As proved by NMR, the Li exchange between


27


Page 28 of 48

PEO and LLZO is characterized by slow kinetics, in the timescale of tenths of a second, which is 3-4 orders of magnitude slower in comparison to Li-ion mobility within PEO or bulk LLZO. In addition, the interfacial resistance has been determined from a trilayer model system, which cannot fully simulate the real interface contact in the composite. The 2 components in the composite underwent a “soft” solvent milling process to improve the homogeneity of the mixture and, likely, their intimate contact. In contrast, the trilayer system was prepared by laminating a LLZO pellet between 2 PEO membranes. Therefore, the interfacial value obtained from EIS, 11.5 kΩ·cm2, should be taken as an estimated upper limit, which is likely lower in the composite system.


28

Page 29 of 48 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 5. Electrodeposition properties and interfacial resistance with Li metal electrodes (a) Evolution of area specific resistance between Li metal and a PEO(LiTFSI)-LLZO (10 vol%) composite electrolyte, at 70oC; (b) Voltage profiles of the symmetric Li metal cell with pure PEO(LiTFSI) and composite electrolytes during stripping/plating experiment at 70oC with a


29


Page 30 of 48

current density of 0.1 mA·cm-2; (c) Engineering Stress-Strain curves for the pure PEO(LiTFSI) and composite LLZGO electrolytes. The curves demonstrate that the ultimate tensile strength and yield stresses are similar for these two materials at approximately 1.1 MPa and 1 MPa, respectively. Inset: the elastic modulus was evaluated using the initial slopes of stress-strain curves (dashed lines) for each electrolyte and found to be 31 MPa for the pure PEO(LiTFSI) and 23 MPa for the composite electrolyte.

The interfacial compatibility of Li metal electrodes and a LLZO(10 vol%)-PEO(LiTFSI) composite electrolyte has been investigated in detail, as Li metal is a high-capacity anode material key to enable high-energy solid-state batteries61,62. In contrast to the negligible effect of the LLZO in terms of conductivity, the garnet fillers show a big impact on the Li electrodeposition properties and cell cycling stability (Figure 5a). The ASR value of the reference PEO(LiTFSI) polymer, rapidly increases from 65.77 Ωcm2 to 270 Ωcm2 in 20 h, and afterwards oscillating around 300 Ωcm2. In contrast to the reference, the initial area specific resistance (ASR) of 10 vol% LLZO composite determined by EIS was 66.2 Ωcm2 and decreased to 33.4 Ωcm2 after 20 h. This drop is a direct consequence of composite melting and the contact improvement between the electrolyte and the Li metal electrodes. Moreover the interface shows stable resistance (33 Ωcm2) over 120 h with only a slight increase to 53.3 Ωcm2 after 1000 h (Figure S13), evidencing excellent long-term stability which is a major requirement for practical application in batteries. Previous reports showed a 50 % reduction of interfacial resistance, from 400 Ω·cm2 to 200 Ω·cm2 in Li7La3Zr2O12/PEO(LiTFSI) composite with stability up to two weeks achieved through elimination of water or organic solvent with fabrication procedure18. Likewise, a 37 % reduction of interfacial resistance has been reported for LLZTO-PEO composite with


30

Page 31 of 48 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


values around 325 Ω·cm2 where the improved compatibility discern in stability of LLZO toward lithium, but also from its scavenging ability16. According to the authors H2O is likely immobilized inside the LLZO particles as well as LiOH on the surface and thus is not able to diffuse from the LLZO into the polymer matrix to react with lithium electrode. In the present work, the Li metal ASR values obtained with the LLZO- PEO(LiTFSI) composite go beyond the state-of-the art, highlighting the importance of processing conditions, free of moisture (lack of characteristic signal for LiOH·H2O in 7Li NMR63), CO2 or solvent traces, which eventually could degrade electrode/electrolyte interface. To deeper understand the influence of garnet fillers on the ASR and shed light into the mechanism of interface stabilization, reversible Li plating/stripping cycles were applied to symmetric Li/electrolyte/Li cells with garnet-free and garnet(10 vol%)-PEO(LiTFSI) composite electrolytes of same thickness (Figure 5b). In this galvanometric cycling experiment, the voltage response was measured while Li metal was being alternately stripped/plated from one electrode to the opposite, in 2 h time intervals, at constant current density of 0.1 mA·cm-2. Initially, both cells respond to the electric field with similar polarization (50 mV), as expected from the similar ionic conductivities of the garnet-free and garnet(10 vol%)-PEO(LiTFSI) electrolytes (Figure 4a). However, this initial polarization evolves differently for both solid electrolytes. After less than 38 h, a drop of polarization is observed in the reference PEO(LiTFSI) electrolyte cell, which eventually leads to a short circuit (V=0) at t= 88 h (Figure 5b, red profile). This behavior is typical of metallic lithium dendrite growth creating an electronic path across the polymer electrolyte 5,13,14. In contrast, the cell containing the garnet electrolyte fillers show an increase of polarization and significant voltage instabilities that deviate from classic ohmic behavior. These voltage


31


Page 32 of 48

instabilities last for 70 h and afterwards the polarization stabilizes back to a similar initial value (Stage I and II in Figure 5b, Figure S14). To explain the different cycling behavior of the cell containing the composite electrolyte, several hypotheses could be considered (SI: Supporting text-2). During Li stripping, points of contact loss (free space) will be created at the Li metal/electrolyte interface which could be filled by the viscous polymer matrix favoring a low resistive and continuous interface. Though this effect it is expected to occur for both electrolytes, the presence of LLZO fillers in the composite has a key role as higher polarization should result from direct contact between LLZO particles and Li metal electrode. The new interface created by the hard ceramic garnet particle is more resistive than with the soft PEO polymer, increasing thus the cell voltage. The non-ohmic behavior can be understood as a response of local temperature gradients at the contacted areas between the LLZO filler and Li metal electrode. Studies on the LLZO/Li interface have recently showed that the resistive interface can build up significant mechanical stress and charge gradients that are expected to locally increase the temperature by joule effect28,64. Since the high polarization response occurs within the voltage instability region (Stage I, Figure 5b), one hypothesis is that direct contact between LLZO particles and metallic lithium occur during some stripping/plating cycles (initial 70 h). These LLZO/Li contact events can occur either during plating or stripping, modifying the interface with the electrode. In the plating process, the electrodeposited Li could reach an LLZO particle nearby the interface, while in the stripping process the free volume created followed by polymer filling could also lead to direct LLZO/Li contacts. Voltage instabilities during initial cycles is a common behavior in cells with organic liquid electrolytes and Li metal anodes23. In such well-studied systems, the so-called “formation step”, is characterized by changes at the Li


32

Page 33 of 48 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


metal surface resulting in uneven Li deposition, until it stabilizes after a certain number of cycles. Analogously, in the solid composite electrolyte, we propose that homogeneous Li electrodeposition is achieved when avoiding direct contact between LLZO particles and Li metal.

Based on all the above, the mechanism of stable Li electrodeposition proposed here is that a reorganization process occurs, in which the composite electrolyte pursues to create a continuous thin polymer layer to separate the surface of LLZO particles from direct contact with the Li metal anode. In agreement with this hypothesis, the polarization value for the composite after the voltage instability region is similar to that of the reference polymer cell before short circuit (Figure 5b). After reorganization, the composite electrolyte cell exhibit stable performances for at least for 750 h which is big improvement in comparison with the LLZO-free cell (short-circuit at t= 88 h). The mechanical properties of solid electrolytes can play a key role in the Li electrodeposition process by blocking the eventual Li dendrites. Generally, it is accepted that a high shear modulus is required to stop growth of lithium dendrites across a solid electrolyte65,66. However, for pure ceramic LLZO electrolytes, the mechanical stress built-up at the Li metal interface during electrodeposition, can lead to cracking and mechanical failure followed by dendrite growth and short-circuit9,67,28. The LLZO pellet scenario cannot be applied to the present LLZO-PEO composite electrolyte since, the stress will be accommodated by the elastic polymer matrix.


33


Page 34 of 48

To investigate the possible influence of mechanical properties in the homogeneous Li electrodeposition process, engineering stress-strain curves were recorded for both the reference PEO and the LLZO-PEO composite electrolyte (Figure 5c). Both electrolytes show similar elastic behavior with elastic modulus of 31 MPa and 23 MPa for the pure polymer and 10 vol% LLZO, respectively. The shear modulus, G, is also expected to be similar based on its dependency on the PEO Poisson’s ratio: 𝐺 = 𝐸 2(1 + 𝜐). Based on the similar elastic and shear modulus, the better cycling stability of the composite/Li metal interface is not likely related to mechanical characteristics of the electrolyte, as previously believed65,66. Rather than an intrinsic material property, the improvement of Li electrodeposition when using a LLZO-PEO(LiTFSI) composite electrolyte can be related to a localized interfacial phenomenon during the electrodeposition process, which favors homogeneous current distribution (Figure 6).


34

Page 35 of 48 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 6. Proposed model of local Li+ transport at various stages of Li electrodeposition in (a,b) PEO(LiTFSI) reference electrolyte and (c,d) LLZO-PEO(LiTFSI) composite electrolyte. (a,c) show the initial stages. (b) Concentration gradients in PEO(LiTFSI) lead to inhomogeneous Li plating and dendrite growth. The brownish layer indicates the formation of a resistive interfase (d) Minimization of the concentration gradient in the composite is facilitated by local Li+ exchange at the LLZO-PEO interface (solid blue arrow), leading to homogenous electrodeposition. Dotted blue arrow represent the Li+ diffusion front.

Homogeneous Li+ transport and electrodeposition requires a low resistive and stable interface layer between the electrolyte and Li metal electrode. Though, in our case the global ionic conductivity was slightly decreased by the introduction of LLZO, these fillers have demonstrate to play a key role on improving the Li electrodeposition dynamics and stability with Li metal electrodes. Considering the above discussion and results (Figure 3 and Figure 5), we propose that the improvement is related to the fact that Li+ can exchange between the


35


Page 36 of 48

garnet and the PEO phases, as demonstrated by 2D EXSY NMR (Figure 3c,d,e). This local Li+ dynamics will enable the minimization of charge gradients near the electrolyte/electrode interface facilitating homogeneous Li electrodeposition (Figure 6d). Minimization of concentration gradients has been previously demonstrated to greatly improve the cycling stability of Li metal in highly concentrated liquid electrolytes68. Similarly, in our solid composite electrolyte, the increased availability of Li+ as a result of the LLZO surface exchange will balance the Li+ concentration near the electrode. Under an applied current density, the Li+ exchanged between the LLZO particle surface and the surrounding PEO can equilibrate the Li+ concentration and reduce the perturbation of Li motion. Such minimization of the concentration gradient is of great practical interest to avoid Li dendrites as they typically start to form when the ionic concentration drop to zero at the negative electrode69. In addition, the lower interfacial impedance obtained in the LLZO-PEO composite (Figure 5a) is in agreement with the proposed model, suggesting that garnet has the ability to locally enhance Li+ ion transport near the electrode/electrolyte interface and reduce anion depletion. Our study shows that there is a favorable structural/chemical/dynamic feature present in LLZO/PEO composite that provides, low ASR, stable electrodeposition and dendrite-free operation with Li metal electrodes. The mechanism of this multi-step process needs to be further understood in order to find the design rules for high performing solid electrolytes of great interest in various battery technologies.

CONCLUSIONS


36

Page 37 of 48 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


The introduction of ion-conducting garnet LLZO microparticles into PEO polymer electrolyte is found to greatly enhance the interfacial stability with Li metal anode upon electrochemical cycling, offering new opportunities to prevent Li dendrites in solid-state batteries. Stable Li electrodeposition was obtained due to the structural reorganization at the interface between Li metal and composite, homogeneous diffusion of lithium ions in the electrolyte achieved through complete salt dissociation and minimized current gradients thanks to the Li ions exchange between PEO and LLZO phases. The elastic and shear modulus of the electrolyte do not play a key role in the homogeneous Li electrodeposition, which is instead dominated by the interfacial composition, assisted by local Li+ dynamics. This knowledge is of interest for elucidating the design rules for other solid composite electrolyte systems to enable stable interfaces with metal anodes, which is essential in different battery technologies – Li metal SSB but also metal-air and Li-S. In addition, 2D EXSY NMR appears as a valuable tool to support the rational design of novel composite electrolytes. Proper tuning of the ceramic-polymer interfaces will offer future opportunities to enhance the macroscopic transport properties, up to now limited by the low-conducting polymeric matrix phase.

REFERENCES (1)

Janek, J.; Zeier, W. G. A Solid Future for Battery Development. Nat. Energy 2016, 1 (9), 1–4.

(2)

Aurbach, D.; Zinigrad, E.; Cohen, Y.; Teller, H. A Short Review of Failure Mechanisms of Lithium Metal and Lithiated Graphite Anodes in Liquid Electrolyte Solutions. Solid State Ionics 2002, 148 (3–4), 405–416.

(3)

Dollé, M.; Sannier, L.; Beaudoin, B.; Trentin, M.; Tarascon, J.-M. Live Scanning Electron


37


Page 38 of 48

Microscope Observations of Dendritic Growth in Lithium/Polymer Cells. Electrochem. Solid-State Lett. 2002, 5 (12), A286. (4)

Kerman, K.; Luntz, A.; Viswanathan, V.; Chiang, Y.-M.; Chen, Z. Review—Practical Challenges Hindering the Development of Solid State Li Ion Batteries. J. Electrochem. Soc. 2017, 164 (7), A1731–A1744.

(5)

Fan, L.; Wei, S.; Li, S.; Li, Q.; Lu, Y. Recent Progress of the Solid-State Electrolytes for High-Energy Metal-Based Batteries. Adv. Energy Mater. 2018, 1702657, 1702657.

(6)

Thangadurai, V.; Narayanan, S.; Pinzaru, D. Garnet-Type Solid-State Fast Li Ion Conductors for Li Batteries: Critical Review. Chem. Soc. Rev. 2014, 43 (13), 4714.

(7)

Hofstetter, K.; Samson, A. J.; Narayanan, S.; Thangadurai, V. Present Understanding of the Stability of Li-Stu Ff Ed Garnets with Moisture , Carbon Dioxide , and Metallic Lithium. J. Power Sources 2018, 390 (March), 297–312.

(8)

Ma, C.; Cheng, Y.; Yin, K.; Luo, J.; Sharafi, A.; Sakamoto, J.; Li, J.; More, K. L.; Dudney, N. J.; Chi, M. Interfacial Stability of Li Metal-Solid Electrolyte Elucidated via in Situ Electron Microscopy. Nano Lett. 2016, 16 (11), 7030–7036.

(9)

Aguesse, F.; Manalastas, W.; Buannic, L.; Del Amo, J. M. L.; Singh, G.; Llordés, A.; Kilner, J. Investigating the Dendritic Growth during Full Cell Cycling of Garnet Electrolyte in Direct Contact with Li Metal. ACS Appl. Mater. Interfaces 2017, 9 (4), 3808–3816.

(10) Cheng, L.; Chen, W.; Kunz, M.; Persson, K.; Tamura, N.; Chen, G.; Doeff, M. Effect of Surface Microstructure on Electrochemical Performance of Garnet Solid Electrolytes. ACS Appl. Mater. Interfaces 2015, 7 (3), 2073–2081. (11) Vervaeke, M.; Calabrese, G. Prospective Design in the Automotive Sector and the Trajectory of the Bluecar Project: An Electric Car Sharing System. Int. J. Veh. Des. 2015, 68 (4), 245. (12) Mejía, A.; Devaraj, S.; Guzm, J.; López Del Amo, J.-M.; García, N.; Rojo, T.; Armand, M.; Tiemblo, P. Scalable Plasticized Polymer Electrolytes Reinforced with Surface-


38

Page 39 of 48 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


Modified Sepiolite Fillers - A Feasibility Study in Lithium Metal Polymer Batteries. 2016, 306, 772–778. (13) Yang, T.; Zheng, J.; Cheng, Q.; Hu, Y. Y.; Chan, C. K. Composite Polymer Electrolytes with Li7La3Zr2O12Garnet-Type Nanowires as Ceramic Fillers: Mechanism of Conductivity Enhancement and Role of Doping and Morphology. ACS Appl. Mater. Interfaces 2017, 9 (26), 21773–21780. (14) Yang, C.; Liu, B.; Jiang, F.; Zhang, Y.; Xie, H.; Hitz, E.; Hu, L. Garnet/Polymer Hybrid IonConducting Protective Layer for Stable Lithium Metal Anode. Nano Res. 2017, 10 (12), 1–10. (15) Zhao, C.-Z.; Zhang, X.-Q.; Cheng, X.-B.; Zhang, R.; Xu, R.; Chen, P.-Y.; Peng, H.-J.; Huang, J.-Q.; Zhang, Q. An Anion-Immobilized Composite Electrolyte for Dendrite-Free Lithium Metal Anodes. Proc. Natl. Acad. Sci. 2017, No. October, 201708489. (16) Cheng, S. H.-S.; He, K.-Q.; Liu, Y.; Zha, J.-W.; Kamruzzaman, M.; Ma, R. L.-W.; Dang, Z.-M.; Li, R. K. Y.; Chung, C. Y. Electrochemical Performance of All-Solid-State Lithium Batteries Using Inorganic Lithium Garnets Particulate Reinforced PEO/LiClO 4 Electrolyte. Electrochim. Acta 2017, 253, 430–438. (17) Fu, K. (Kelvin); Gong, Y.; Dai, J.; Gong, A.; Han, X.; Yao, Y.; Wang, C.; Wang, Y.; Chen, Y.; Yan, C.; et al. Flexible, Solid-State, Ion-Conducting Membrane with 3D Garnet Nanofiber Networks for Lithium Batteries. Proc. Natl. Acad. Sci. 2016, 113 (26), 7094– 7099. (18) Keller, M.; Appetecchi, G. B.; Kim, G. T.; Sharova, V.; Schneider, M.; Schuhmacher, J.; Roters, A.; Passerini, S. Electrochemical Performance of a Solvent-Free Hybrid Ceramic-Polymer Electrolyte Based on Li7La3Zr2O12in P(EO)15LiTFSI. J. Power Sources 2017, 353, 287–297. (19) Langer, F.; Bardenhagen, I.; Glenneberg, J.; Kun, R. Microstructure and Temperature Dependent Lithium Ion Transport of Ceramic-Polymer Composite Electrolyte for Solid-State Lithium Ion Batteries Based on Garnet-Type Li7La3Zr2O12. Solid State


39


Page 40 of 48

Ionics 2016, 291, 8–13. (20) Zheng, J.; Hu, Y. Y. New Insights into Compositional Dependence of Li-Ion Transport in Polymer-Ceramic Composite Electrolytes. ACS Appl. Mater. Interfaces 2018, acsami.7b17301. (21) Zheng, J.; Dang, H.; Feng, X.; Chien, P.-H.; Hu, Y.-Y. Li-Ion Transport in a Representative Ceramic–polymer–plasticizer Composite Electrolyte: Li 7 La 3 Zr 2 O 12 –polyethylene Oxide–tetraethylene Glycol Dimethyl Ether. J. Mater. Chem. A 2017, 5, 18457–18463. (22) Takeda, Y.; Yamamoto, O.; Imanishi, N. Lithium Dendrite Formation on a Lithium Metal Anode from Liquid, Polymer and Solid Electrolytes. Electrochemistry 2016, 84 (4), 210–218. (23) Bieker, G.; Winter, M.; Bieker, P. Electrochemical in Situ Investigations of SEI and Dendrite Formation on the Lithium Metal Anode. Phys. Chem. Chem. Phys. 2015, 17 (14), 8670–8679. (24) Chazalviel, J. N. Electrochemical Aspects of the Generation of Ramified Metallic Electrodeposits. Phys. Rev. A 1990, 42 (12), 7355–7367. (25) Crowther, O.; West, A. C. Effect of Electrolyte Composition on Lithium Dendrite Growth. J. Electrochem. Soc. 2008, 155 (11), A806. (26) Monroe, C.; Newman, J. The Impact of Elastic Deformation on Deposition Kinetics at Lithium/Polymer Interfaces. J. Electrochem. Soc. 2005, 152 (2), A396. (27) Zhang, X.-W.; Li, Y.; Khan, S. A.; Fedkiw, P. S. Inhibition of Lithium Dendrites by Fumed Silica-Based Composite Electrolytes. J. Electrochem. Soc. 2004, 151 (8), A1257. (28) Porz, L.; Swamy, T.; Sheldon, B. W.; Rettenwander, D.; Frömling, T.; Thaman, H. L.; Berendts, S.; Uecker, R.; Carter, W. C.; Chiang, Y.-M. Mechanism of Lithium Metal Penetration through Inorganic Solid Electrolytes. Adv. Energy Mater. 2017, 7 (20), n/an/a. (29) Stone, G. M.; Mullin, S. A.; Teran, A. A.; Hallinan, D. T.; Minor, A. M.; Hexemer, A.; Balsara,


40

Page 41 of 48 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


N. P. Resolution of the Modulus versus Adhesion Dilemma in Solid Polymer Electrolytes for Rechargeable Lithium Metal Batteries. J. Electrochem. Soc. 2012, 159 (3), A222–A227. (30) Liu, S.; Imanishi, N.; Zhang, T.; Hirano, A.; Takeda, Y.; Yamamoto, O.; Yang, J. Lithium Dendrite

Formation

in

Bis(Trifluoromethanesulfonyl)Imide

Li/Poly(Ethylene and

Oxide)–Lithium

N-Methyl-N-Propylpiperidinium

Bis(Trifluoromethanesulfonyl)Imide/Li Cells. J. Electrochem. Soc. 2010, 157 (10), A1092. (31) Scrosati, B.; Croce, F.; Persi, L. Impedance Spectroscopy Study of PEO-Based Nanocomposite Polymer Electrolytes. J. Electrochem. Soc. 2000, 147 (5), 1718. (32) Tecnologie, D.; Avanzate, E. Electrochemical Characterization of a Composite Polymer Electrolyte with Improved Lithium Metal Electrode Interfacial Properties. 1999, 5, 59–63. (33) Bernuy-Lopez, C.; Manalastas, W.; Lopez Del Amo, J. M.; Aguadero, A.; Aguesse, F.; Kilner, J. A. Atmosphere Controlled Processing of Ga-Substituted Garnets for High LiIon Conductivity Ceramics. Chem. Mater. 2014, 26 (12), 3610–3617. (34) Buannic, L.; Orayech, B.; López Del Amo, J.-M.; Carrasco, J.; Katcho, N. A.; Aguesse, F.; Manalastas, W.; Zhang, W.; Kilner, J.; Llordés, A. Dual Substitution Strategy to Enhance Li + Ionic Conductivity in Li 7 La 3 Zr 2 O 12 Solid Electrolyte. Chem. Mater. 2017, 29 (4), 1769–1778. (35) J. Rodríguez-Carvajal. FullProf Suite http://www.ill.eu/sites/fullprof/index.html. (36) Stratton, J. A. Electromagnetic Theory; 1941. (37) Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calvé, S.; Alonso, B.; Durand, J. O.; Bujoli, B.; Gan, Z.; Hoatson, G. Modelling One- and Two-Dimensional Solid-State NMR Spectra. Magn. Reson. Chem. 2002, 40 (1), 70–76. (38) Sanchez-Soto, P. J.; Gines, J. M.; Arias, M. J.; Novak, C.; Ruiz-Conde, A. Effect of Molecular


41


Page 42 of 48

Massa on the Melting Temperature , Enthalpy and Entropy of Hydroxi-Terminated PEO. J. Therm. Anal. Calorim. 2002, 67, 189–197. (39) Bruce, P. G.; Vincent, C. A. Steady State Current Flow in Solid Binary Electrolyte Cells. J. Electroanal. Chem. 1987. (40) Bruce, P. G.; Evans, J.; Vincent, C. A. Conductivity and Transference Number Measurements on Polymer Electrolytes. Solid State Ionics 1988. (41) El Shinawi, H.; Janek, J. Stabilization of Cubic Lithium-Stuffed Garnets of the Type “Li 7La3Zr2O12” by Addition of Gallium. J. Power Sources 2013, 225, 13–19. (42) Jalem, R.; Rushton, M. J. D.; Manalastas, W.; Nakayama, M.; Kasuga, T.; Kilner, J. A.; Grimes, R. W. Effects of Gallium Doping in Garnet-Type Li7La3Zr2O12 Solid Electrolytes. Chem. Mater. 2015, 27 (8), 2821–2831. (43) Simonetti, E.; Carewska, M.; Maresca, G.; De Francesco, M.; Appetecchi, G. B. Highly Conductive, Ionic Liquid-Based Polymer Electrolytes. J. Electrochem. Soc. 2017, 164 (1), A6213–A6219. (44) Kim, G. T.; Appetecchi, G. B.; Alessandrini, F.; Passerini, S. Solvent-Free, PYR1ATFSI Ionic Liquid-Based Ternary Polymer Electrolyte Systems. I. Electrochemical Characterization. J. Power Sources 2007, 171 (2), 861–869. (45) Buvana, P.; Vishista, K.; Shanmukaraj, D.; Murugan, R. Lithium Garnet Oxide Dispersed Polymer Composite Membrane for Rechargeable Lithium Batteries. Ionics (Kiel). 2017, 23 (3), 541–548. (46) Liu, W.; Liu, N.; Sun, J.; Hsu, P. C.; Li, Y.; Lee, H. W.; Cui, Y. Ionic Conductivity Enhancement of Polymer Electrolytes with Ceramic Nanowire Fillers. Nano Lett. 2015, 15 (4), 2740–2745. (47) Mallakpour, S.; Khadem, E. Recent Development in the Synthesis of Polymer Nanocomposites Based on Nano-Alumina. Prog. Polym. Sci. 2014, 51, 74–93. (48) Arya, A.; Sharma, A. L. Insights into the Use of Polyethylene Oxide in Energy


42

Page 43 of 48 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


Storage/Conversion Devices: A Critical Review. J. Phys. D. Appl. Phys. 2017, 50 (44). (49) Zheng, J.; Tang, M.; Hu, Y. Y. Lithium Ion Pathway within Li7La3Zr2O12-Polyethylene Oxide Composite Electrolytes. Angew. Chemie - Int. Ed. 2016, 55 (40), 12538–12542. (50) Tan-Vu Huynh. NMR study of lithium mobility in polymer electrolytes. Other. Université d’Orléans, 2015. English. . (51) Vinod Chandran, C.; Heitjans, P. Solid-State NMR Studies of Lithium Ion Dynamics Across Materials Classes. In Annual Reports on NMR Spectroscopy; 2016. (52) Kuhn, A.; Kunze, M.; Sreeraj, P.; Wiemhöfer, H. D.; Thangadurai, V.; Wilkening, M.; Heitjans, P. NMR Relaxometry as a Versatile Tool to Study Li Ion Dynamics in Potential Battery Materials. Solid State Nucl. Magn. Reson. 2012, 42, 2–8. (53) Yu, C.; Ganapathy, S.; Eck, E. R. H. V.; Wang, H.; Basak, S.; Li, Z.; Wagemaker, M. Accessing the Bottleneck in All-Solid State Batteries, Lithium-Ion Transport over the Solid-Electrolyte-Electrode Interface. Nat. Commun. 2017, 8 (1), 1–9. (54) Zhang, J.; Zhao, N.; Zhang, M.; Li, Y.; Chu, P. K.; Guo, X.; Di, Z.; Wang, X.; Li, H. Flexible and Ion-Conducting Membrane Electrolytes for Solid-State Lithium Batteries: Dispersion of Garnet Nanoparticles in Insulating Polyethylene Oxide. Nano Energy 2016, 28, 447–454. (55) Schneider, G. J. Dynamics of Nanocomposites. Curr. Opin. Chem. Eng. 2017, 16, 65–77. (56) Long, L.; Wang, S.; Xiao, M.; Meng, Y. Polymer Electrolytes for Lithium Polymer Batteries. J. Mater. Chem. A 2016, 4 (26), 10038–10069. (57) Jiang, C.; Li, H.; Wang, C. Recent Progress in Solid-State Electrolytes for Alkali-Ion Batteries. Sci. Bull. 2017, 62 (21), 1473–1490. (58) Tikekar, M. D.; Archer, L. A.; Koch, D. L. Stability Analysis of Electrodeposition across a Structured Electrolyte with Immobilized Anions. J. Electrochem. Soc. 2014, 161 (6), A847–A855.


43


Page 44 of 48

(59) Diederichsen, K. M.; Buss, H. G.; McCloskey, B. D. The Compensation Effect in the VogelTammann-Fulcher (VTF) Equation for Polymer-Based Electrolytes. Macromolecules 2017, 50 (10), 3831–3840. (60) Langer, F.; Palagonia, M. S.; Bardenhagen, I.; Glenneberg, J.; La Mantia, F.; Kun, R. Impedance Spectroscopy Analysis of the Lithium Ion Transport through the Li 7 La 3 Zr 2 O 12 /P(EO) 20 Li Interface. J. Electrochem. Soc. 2017, 164 (12), A2298–A2303. (61) Liu, B.; Zhang, J. G.; Xu, W. Advancing Lithium Metal Batteries. Joule 2018, 2 (5), 833– 845. (62) Lin, D.; Liu, Y.; Cui, Y. Reviving the Lithium Metal Anode for High-Energy Batteries. Nat. Nanotechnol. 2017, 12 (3), 194–206. (63) Meyer, B. M.; Leifer, N.; Sakamoto, S.; Greenbaum, S. G.; Grey, C. P. High Field Multinuclear NMR Investigation of the SEI Layer in Lithium Rechargeable Batteries. Electrochem. Solid-State Lett. 2005, 8 (3), A145. (64) Manalastas, W.; Rikarte, J.; Chater, R.; Brugge, R.; Aguadero, A.; Buannic, L.; Llordés, A.; Aguesse, F.; Kilner, J. Mechanical Failure of Garnet Electrolytes during Li Electrodeposition Observed by In-Operando Microscopy. J. Power Sources. 2019 412, 287-293. (65) Khurana, R.; Schaefer, J. L.; Archer, L. A.; Coates, G. W. Suppression of Lithium Dendrite Growth Using Cross-Linked Polyethylene/Poly(Ethylene Oxide) Electrolytes: A New Approach for Practical Lithium-Metal Polymer Batteries. J. Am. Chem. Soc. 2014, 136 (20), 7395–7402. (66) Wang, S.; Xu, H.; Li, W.; Dolocan, A.; Manthiram, A. Interfacial Chemistry in Solid-State Batteries: Formation of Inter-Phase and Its Consequences Interfacial Chemistry in Solid-State Batteries: Formation of Inter- Phase and Its Consequences. 2017, No. 3. (67) Sharafi, A.; Meyer, H. M.; Nanda, J.; Wolfenstine, J.; Sakamoto, J. Characterizing the Li– Li 7 La 3 Zr 2 O 12 Interface Stability and Kinetics as a Function of Temperature and Current Density. J. Power Sources 2016, 302, 135–139.


44

Page 45 of 48 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


(68) Qian, J.; Henderson, W. A.; Xu, W.; Bhattacharya, P.; Engelhard, M.; Borodin, O.; Zhang, J. G. High Rate and Stable Cycling of Lithium Metal Anode. Nat. Commun. 2015, 6. (69) Brissot, C.; Rosso, M.; Chazalviel, J. N.; Lascaud, S. Dendritic Growth Mechanisms in Lithium/Polymer Cells. J. Power Sources 1999, 81, 925–929. (70) Rettenwander,D.; Langer J.; Schmidt, W.;, Arrer, C.; Harris, K.J.; Terskikh, V.; Goward, G.R.; Wilkening, M.; Amthauer, G. Site Occupation of Ga and Al in Stabilized Cubic Li7– 3(x+y)GaxAlyLa3Zr2O12 Garnets As Deduced from 27Al and 71Ga MAS NMR at Ultrahigh Magnetic Fields. Chem. Mater., 2015, 27 (8), 3135–3142.

ASSOCIATED CONTENT Supporting Information Table with nominal and corrected LLZO content in composites; XRD patterns of pure cubic LLZO and sample containing secondary phases; LLZO particle size distribution before and after ball milling; XRD patterns of PEO(LiTFSI) membrane, LLZO powder and 30 vol% LLZO composite membrane; FTIR spectra of PEO powder and PEO membrane; Molecular weight of PEO in composite; EIS spectra of PEO(LiTFSI) electrolytes prepared using magnetic stirrer and ball milling methods; DSC measurements; Table with melting temperature and crystallinity degree; SEM images of composite electrolytes; SEM image with Back Scattered Electron Detector (BSED) for Z-contrast; 7Li MAS NMR spectra of single components;, DC polarization experiment; Nyquist plots of composite electrolyte at various temperatures; Nyquist plots after 100 h and 1000 h of storage time; Voltage profiles of 3 different symmetric cells during stripping/plating experiment; Supporting text with effective contact


45


Page 46 of 48

area coefficient calculation; Supporting text with hypothesis to explain the different cycling behavior of the cell containing the composite. (PDF) AUTHOR INFORMATION Corresponding Author *E-mail: [email protected] Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Funding Sources This work was carried out at CIC Energigune (Spain) and funded by Gobierno Vasco (ELKARTEK CIC ENERGIGUNE 15) and by MINECO (SELIS project ENE2015-64907-C2-1R). A.L. also thanks IKERBASQUE for financial support.

ACKNOWLEDGMENT The authors thank Dr. Cristina Luengo Vilumbrales for assistance with DSC measurements and Maria Jáuregui for service of PPMS and XRD platform. We are also grateful to Prof. John Kilner, Prof. Michel Armand and Dr. Devaraj Shanmukaraj for fruitful discussions ABBREVIATIONS


46

Page 47 of 48 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


ASR: area specific resistance; EIS: electrochemical impedance spectroscopy; GPC: gel permeation

chromatography;

LIB:

lithium-ion

batteries;

LiTFSI:

lithium

bis(trifluoromethanesulfonyl)imide; LLZO: Li6.55Ga0.15La3Zr2O12; LZO: Li7La3Zr2O12; PEO: poly(ethylene oxide); PPMS: physical property measurement station; SEI: solid electrolyte interface; SSB: solid-state batteries


47


Page 48 of 48

Figure For Table of Contents Only:


48

GarnetâPolymer Composite Electrolytes: New Insights on Local Li-Ion

Recommend Documents

GarnetâPolymer Composite Electrolytes: New Insights on Local Li-Ion