Disorder-driven Sintering-free Garnet-type Solid Electrolytes

doi:10.21203/rs.3.rs-4611381/v1

Download PDF

Article

Disorder-driven Sintering-free Garnet-type Solid Electrolytes

https://doi.org/10.21203/rs.3.rs-4611381/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Oxide ceramic electrolytes for realization of high-energy lithium metal batteries generally require a series of high-temperature processes for the formation of the desired phase and inter-particle sintering. The high-temperature processing can lead to compositional changes or mechanical deformation, consequently, resulting in serious issues with material reliabilities. Here, we introduce a disorder-driven sintering-free garnet-type solid electrolyte using a novel approach for creating an amorphous matrix followed by a single-step mild heat-treatment. The softened mechanical property (yield pressure, P_y = 359.8 MPa) of disordered base materials can achieve a facile formation of a dense amorphous matrix and contributes to maintaining inter-particle connectivity during crystallization. Remarkably, the formation of the highly conductive cubic-phase garnet is triggered at a drastically lowered temperature of 350°C, leading to high ionic conductivity (σ_Li+ = 1.8 × 10^–4 S/cm at 25°C) through a single-step mild heat treatment at 500°C. The disorder-driven garnet solid electrolyte exhibits electrochemical performance similar to that of the conventional garnet solid electrolyte sintered at > 1100°C. This electrolyte exhibits the lowest processing temperature ever reported for garnet-type solid electrolytes with a high lithium ionic conductivity of ~ 10^–4 S/cm. These findings will promote the fabrication of uniform, thin, and wide solid electrolyte membranes, which is a significant hurdle in the commercialization of oxide-based lithium metal batteries, and demonstrate the untapped capabilities of garnet-type oxide solid electrolytes.

Physical sciences/Materials science/Materials for energy and catalysis/Batteries

Physical sciences/Chemistry/Inorganic chemistry/Solid-state chemistry

Physical sciences/Energy science and technology/Energy storage/Batteries

Physical sciences/Chemistry/Energy

Li metal battery

Solid-state battery

Garnet-type solid electrolyte

Amorphous material

Ceramic materials play a crucial role in various fields because of their unique combination of diverse chemical, mechanical, and electrical properties.^1–6 Establishing intimate inter-particle connections based on their rigid characteristics is necessary to preserve their outstanding intrinsic properties at particle-level into bulk-scale products.¹ In particular, intimate inter-particle connections are primarily achieved through the sintering process for oxide-based ceramic materials which hold significant importance in the field of ceramics.^7,8 The conventional sintering process requires high-temperature conditions because of their high thermal activation energy necessary for mass transportation, which not only induces complexity in the process by causing variations in composition, morphological changes or shape deformations, but also leads to an increase in processing costs.^9,10

Given the benefits from using the ceramic material, numerous attempts have been made to utilize lithium metal as an anode in the lithium battery community.^11,12 This class of materials is expected to address the persistent issue of dendrite formation in a lithium metal battery system, because of their excellent mechanical strength and lithium ion transport number close to unity.^12,13 The cohesive structure, characterized by tightly-knit inter-particle connections, plays a crucial role in suppressing dendrite formation in lithium metal battery systems while simultaneously creating a uniform lithium ionic conduction pathway.^14–16 In addition, they are particularly suitable for developing safe batteries with high-energy densities given their inherent incombustibility.¹³ Among a wide range of ceramic-based solid electrolytes, including sulfides, oxides, and halides,^12,17–19 garnet-type oxide solid electrolytes exemplified by materials such as lithium lanthanum zirconium oxide (LLZO) have attracted considerable research attention because of their exceptional properties such as high ionic conductivity (~ 10^− 4 to ~ 10^− 3 S/cm at 25°C), wide electrochemical windows, and excellent chemical compatibility with lithium metal.^20–22

Similar to most oxide ceramic materials, however, the garnet-type oxide material needs to undergo a series of high-temperature processes before it can be used as a solid electrolyte. Following a high-temperature crystallization process (> 900°C) essential for forming the desired cubic-phase with high lithium ionic conductivity, an additional sintering process at even higher temperatures (> 1100°C) needs to be accompanied for fully utilizing its high lithium ionic conductivity at the particle level into bulk-scale solid electrolyte membranes.^21,23 In the sintering process, intimate inter-particle connections form by mass transportation, thereby constructing a uniform lithium-ion conduction pathway. This series of high-temperature processes increases processing costs, induces chemical compositional changes and mechanical deformation in the solid electrolyte membranes, and generates various forms of defects.^9,24 This compromises lithium ionic conductivity and diminishes the uniformity of the solid electrolyte membranes, thereby deteriorating battery performance. The high-temperature processes involved in attaining highly conductive and dense garnet-based ceramic electrolytes become particularly critical given the stringent requirements for solid electrolytes to be thinner and wider for higher energy density coupled with the highly volatile lithium.⁹ Thus, in recent years, extensive research has focused on effectively achieving intimate inter-particle connections at low temperatures while maintaining ionic transport properties in garnet-type solid electrolytes.^14,16,25

To address this issue, considerable efforts have been devoted to lowering the sintering temperature for practical applications. Most research studies have focused on employing sintering aids such as Li₃BO₃, Al₂O₃, and SiO₂, exploiting their liquid behavior.^26–29 Considering that the precise control of chemical compositions is crucial in battery systems, the incorporation of heterogeneous compounds is an unfavorable approach because it can lead to undesirable (electro)chemical reactions or act as a resistive component. Thus far, several studies have focused on reducing the sintering temperatures without the use of foreign materials through the precise control of process parameters such as sintering atmosphere, pressure, thermal history, or particle size.^30–34 Further, a few studies on vacuum-based thin-film deposition have been reported to avoid traditional high-temperature sintering.^9,35–37 Although these approaches significantly reduce processing temperature, they still necessitate high-temperature sintering (> 900°C) or exhibit limitations in lithium ionic conductivity caused by the imperfect phase formation. Therefore, a new strategy is required to design materials that can significantly lower the processing temperature while maintaining high lithium ionic conductivity, without relying on heterogeneous materials.

Herein, we propose a disorder-driven garnet-type (D-garnet) solid electrolytes designed to deliver high ionic conductivity while drastically lowering the process temperature. The structural disordering of starting materials achieved via mechano-chemical activation effectively renders mechanical ductility, enabling the facile creation of a dense amorphous matrix in which particles are intimately interconnected under uniaxial pressure at ambient temperature. The transformation to the cubic phase (starting at the exceptionally low temperature of 350°C) and the establishment of an inter-particle connection within the amorphous matrix occur concurrently through a single-step mild heat treatment (500°C) without the use of sintering aids or field-assisted sintering technology. High ionic conductivity (σ_Li+ = 1.8 × 10^–4 S/cm at 25°C) is attained through the single-step heat treatment without compromising the electrochemical performance of conventional garnet-type electrolytes. Through the comprehensive analysis, we revealed that the disordered amorphous nature contributes to lowering the phase formation temperature and enabling desirable inter-particle connections. Such advancements present a novel methodology for unleashing the potential of garnet-type oxide solid electrolytes, thereby ushering in a new era of energy storage technology.

Design strategy and disorder-driven soft mechanical properties

For lowering the process temperature of the garnet solid electrolyte without compromising its high ionic conductivity, there are two requirements: 1) reducing the phase formation temperature of a highly conductive pure cubic-phase, and 2) providing an environment that can effectively construct an ionic conduction pathway even at low-temperature. To achieve both objectives simultaneously, we employed a strategy of amorphizing the starting materials. Raising the initial energy state of the precursors, and thereby reducing the kinetic barrier, is expected to relieve the temperature requirement for obtaining a highly conductive cubic phase. At the same time, amorphization can facilitate the formation of compacted bulk-scale amorphous matrices through applied pressure by homogenizing particle sizes and imparting structural flexibility to the material. The dense amorphous matrix is expected to be able to aid in establishing a desirable inter-particle connection during crystal growth at moderate temperature, as observed in glass-ceramic materials.^38–40 Therefore, as shown in Fig. 1a, our approach enables the replacement of conventional high-temperature processes with a single-step mild heat-treatment. In this study, we employed a simple mechano-chemical activation method, which is one of the strategies for amorphizing materials. A Ta-doped garnet electrolyte of Li_6.5La₃Zr_1.5Ta_0.5O₁₂ (LLZTO), widely known for its high lithium ionic conductivity (> 1.0 × 10^–4 S/cm) and high stability with lithium metals, was utilized as a model system.

First, we prepared the amorphous precursors for LLZTO (a-LLZTO) using a mechano-chemical synthetic route for rendering the comprehensive structural disorder to crystalline precursors. As shown in Fig. 1b, a mixture of conventional crystalline precursors used for synthesizing LLZTO that consists of stoichiometric amounts of Li₂O, La₂O₃, ZrO₂, and Ta₂O₅ was mechano-chemically activated in a rotating jar under an Ar atmosphere (see Methods section for details). During the high-energy planetary milling process, a significant mechano-chemical energy was imparted to the crystalline precursors, thereby disrupting the regular crystal lattice and inducing atomic rearrangement. This ultimately results in the formation of a metastable amorphous structure devoid of the long-range ordering of constituent atoms in each crystalline precursor. Figure 1c shows the X-ray diffraction (XRD) patterns of the precursor mixture as a function of milling time and indicates that the pristine crystalline mixture underwent substantial amorphization over time, with the majority of crystalline peaks disappearing 2 h of milling. Eventually, after 15 h, a halo pattern without any discernible crystalline peaks was observed, thereby implying that the crystal structure of the precursor had become fully amorphized. This amorphous characteristic was further confirmed through transmission electron microscope (TEM) analysis as shown in Figure S1. The bright-field TEM image and selected area electron diffraction (SAED) profile of a-LLZTO exhibited a typical feature of the amorphous phase represented by the absence of lattice fringes or diffuse diffraction rings.

We investigated the compaction behavior of powders, which is a key factor in determining particle connectivity, to evaluate how structural difference affects mechanical properties. Crystalline cubic garnet powder (c-LLZTO) conventionally used in the sintering process and the newly synthesized a-LLZTO powder were compared because the mechanical property of a-LLZTO is important to create dense amorphous matrices in our approach. The samples were subjected to repeated compression and release using universal testing machine (UTM), with the applied pressure increasing incrementally from 50 to 1000 MPa. The Heckel model, widely used to predict powder compaction behavior, was employed to quantitatively determine mechanical properties such as yield pressure.^41–44 (see Methods section for details) Fig. 1d illustrates the relative volume changes of c-LLZTO and a-LLZTO based on the compaction profile. The relative volumes were calculated based on their true densities of 5.38 and 4.53 g/cc, respectively. Although c-LLZTO exhibited poor compaction behavior because of its rigid nature, a-LLZTO exhibited remarkably enhanced compaction behavior under the same applied pressure (e.g., under 500 MPa, normalized specific volume of 1.53 and 1.17 for c-LLZTO and a-LLZTO, respectively.). The yield pressures were calculated as 536.5 and 359.8 MPa for c-LLZTO and a-LLZTO, respectively, underscoring that deformability was significantly enhanced because of the disorder-driven structural flexibility.

We compared the cross-sectional scanning electron microscope (SEM) images of both materials prepared after applying uniaxial pressure of > 1,000 MPa (Fig. 1e) to explore their microstructure (see Methods section for details). For c-LLZTO, each particle appears to be well distinguishable with noticeable pores, which corresponds to a porosity of 35.1% (measured density = 3.49 g/cm³ and true density = 5.38 g/cm³). These particles are linked through only point-contact, critically restricting the ionic conduction pathway. In contrast, a-LLZTO exhibits strong inter-particle cohesion, thereby creating a highly connected matrix with a remarkably lower porosity of 11.5% (measured density = 4.01 g/cm³ and true density = 4.53 g/cm³). This feature of a-LLZTO is reminiscent of the microstructure achieved through the conventional high-temperature sintering process, and it clearly highlights the beneficial influence of improved deformability in achieving such a dense microstructure.

Structural evolution of amorphous precursor for garnet

We investigated the crystallization behavior of a-LLZTO to gain a comprehensive understanding of the phase-formation process. Figure 2a presents a comparative analysis of the crystallization behavior between a conventional crystalline precursor and amorphous precursor (a-LLZTO) with 20% excess lithium. In the conventional solid-state synthetic route, the majority peaks of crystalline precursor remain detectable under a temperature of 700°C, implying that the crystalline precursor decomposes at the high temperature. Further, a high thermal energy of approximately 1000°C is required to achieve the desired pure cubic-phase, and this thermal energy should be continuously supplied for an extended period (12 h). In contrast, for the amorphous precursor, a single cubic-phase can be readily obtained even under a mild heat-treatment condition of 500°C for 2 h, emphasizing the significance of the rational design of the starting materials. Figures 2b and S2 display the XRD patterns of samples using a-LLZTO, with varied heat-treatment temperatures. The cubic-phase garnet was obtained even at a low-temperature of 400°C with impurities of pyrochlore oxide (La₂Zr₂O₇, LZO). The impurities disappeared completely with an increase in the temperature to 500°C. The crystallization of the a-LLZTO occurring at a low temperature was confirmed through differential scanning calorimetry (DSC) analysis. The exothermic reaction of the a-LLZTO corresponding to crystallization into the cubic-phase initiated at ~ 400°C (Figure S3).

We conducted a systematic analysis for in-depth understanding of the phase evolution behavior using in situ heating XRD and high-resolution TEM (HRTEM) analysis during the crystallization process. XRD patterns were acquired in real-time during the heating process within the temperature range of 25–1000°C. As depicted in Fig. 2c, the cubic-phase formation starts at 400°C, indicated by the emergence of the two distinctive diffraction peaks at ~ 16° and ~ 19°, corresponding to (211) and (220) crystallographic planes of the cubic-phase garnet, respectively. The obtained result is consistent with the ex situ XRD and DSC results discussed previously. The in situ heating HRTEM observation in Fig. 2d directly verified the nucleation of nanosized (< 10 nm) crystals emerging from the amorphous precursor at ~ 350°C, followed by an extensive crystallization occurring within the temperature range from 350–500°C. The cubic-phase was observed at 350°C through the analysis of diffraction pattern observed in the SAED image, along with the presence of impurities such as unreacted lithium carbonate. Diffraction patterns corresponding to the lithium carbonate diminished with an increase in temperature, and above 500°C, only diffraction patterns associated with the cubic-phase remained. The detail SAED pattern indexing of D-garnet solid electrolytes is shown in Table S1. Based on the results of the phase evolution behavior, the schematic of reaction coordinate for the formation of cubic-phase garnet by comparing when using the crystalline and amorphous precursors as the starting materials, is illustrated in Fig. 2e. As per thermodynamics, elevating its initial energy state via a mechano-chemical activation to create the disordered phase of starting materials can intensify the thermodynamic driving force for the formation of a cubic-phase garnet (see Supplementary Note 1). Moreover, the transition to a disordered structure, where atoms are irregularly distributed, can help achieve facile atomic rearrangements caused by its higher atomic mobility. This could lower the activation energy, consequently easing the thermal energy requirement for the cubic-phase formation.

A comprehensive series of analyses were performed to explore the structural information of the D-garnet solid electrolytes. Figure 2f shows the results of the Rietveld refinement for global structure analysis, and the formation of the cubic-phase garnet (lattice constant = 12.943 Å) with a trace amount of impurities was confirmed. The structural similarity of the D-garnet electrolyte to the conventional cubic-phase garnet formed at high temperatures implies that the disordered initial state facilitates the nucleation and growth of the cubic-phase by lowering the activation energy, which was also confirmed through Raman analysis (Figure S4). Further, it indicates that the ionic transport capability within the crystal structure is comparable to that of the conventional cubic-phase garnet. Magic angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy and X-ray absorption spectroscopy (XAS) were conducted to investigate the detailed local structure. ⁷Li MAS NMR spectroscopy was performed for examining the lithium local environments in the D-garnet solid electrolytes. In Fig. 2g, ⁷Li MAS NMR spectra shows clearly distinguishable lithium local environments between the pristine a-LLZTO and D-garnet formed after the heat-treatment above 400°C. For a-LLZTO, a broad peak at a chemical shift of 3.1 ppm was observed, which was attributed to the undefined lithium sites within the disordered structure. In contrast, that of the D-garnet exhibited distinct narrow peaks at the chemical shift of 2.4 ppm, identical to that of the commercial cubic-phase garnet synthesized at high-temperature, thereby implying its similar lithium environment with the reference cubic-phase garnet. The peaks became sharper and similar to the reference peak with an increase in temperature, which is probably because of their proximity to the ordering of the perfect cubic-phase garnet with an increase in crystallinity, as observed in the XRD results. The XAS analysis was performed to observe the local structural evolution of D-garnet solid electrolytes during heat-treatment. Figure S5 shows Zr K-edge XAS spectra of initial amorphous state (a-LLZTO) and the D-garnet solid electrolytes formed at varied temperature. The Fourier transform of extended X-ray absorption fine structure (FT-EXAFS) oscillation for the Zr K-edge is shown in Fig. 2h. That of the commercial cubic-phase garnet with high crystallinity is presented together as a reference. A cubic-phase garnet exhibits two distinct peaks in the FT-EXAFS oscillation, which are represented by Zr–O and Zr–O–La bonding, respectively. In contrast, the magnitude of most peaks for a-LLZTO, except the first peak corresponding to bonding with the first nearest neighboring atoms (Zr–O), was significantly low. This implies an almost complete absence of the long-range ordering of a-LLZTO. At a heating temperature of 500°C, the second peak corresponding to the Zr–O–La bonding was observed at the same peak position as that of reference cubic garnet where the pure cubic-phase was formed. The peak profile became similar to that of the reference with an increase in heating temperature. We conducted EXAFS data fitting to obtain more detailed structural information, and the results have been presented in Table S2. We noticed that the most fitted values in the amorphous state, except for the interatomic distance of Zr–O bonding, show a significant deviation from those in the reference cubic-phase garnet. The comparable interatomic distance for the Zr–O bonding between the amorphous state (2.07 Å) and reference cubic-phase garnet (2.09 Å) can be attributed to the similar Zr–O bond lengths observed in the crystalline ZrO₂ precursor (Figure S6). We observed a tendency in which the fitted values became similar to those of the reference cubic-phase garnet when the heat-treatment temperature increased to 600°C. There was a significant discrepancy in some values such as coordination number (2.76 for the Zr–O–La bonding of the D-garnet and 3.49 for that of the reference) and Debye–Waller factor (5.02 × 10^–3 Å² for the Zr–O bonding of the D-garnet and 4.63 × 10^–3 Å² for that of the reference), which implies that the local bonding nature of the D-garnet solid electrolyte formed at 600°C differs from that of the highly crystalline cubic-phase garnet. The difference in the Debye–Waller factor, which indicates the degree of disorder, suggests that the D-garnet solid electrolytes have a certain level of static disorder compared to that of the reference cubic-phase garnet.

Construction of ion conduction pathway within an amorphous matrix

In our approach, we anticipated that the nucleation and growth of the cubic-phase within the dense amorphous matrix would facilitate the construction of the ion conduction pathway, thereby enabling ionic conduction throughout the entire bulk solid electrolyte membrane without the need for high-temperature sintering. In this regard, the D-garnet solid electrolyte membranes were prepared in two steps: first, pelletizing the soften a-LLZTO to create the dense amorphous matrix, and second, mild heat-treatment of the pellet under atmospheric pressure to form the cubic-phase accompanying the inter-particle connections. We measured the lithium ionic conductivity and electrochemical stability against lithium metal to evaluate the feasibility of D-garnet as solid electrolytes. The ionic conductivity was determined through electrochemical impedance spectroscopy (EIS), employing a symmetric configuration with ion-blocking gold electrodes. Figure 3a presents Nyquist plots captured at 25°C for D-garnet solid electrolytes prepared under various heat-treatment conditions. All curves only displayed a diffusion tail stemming from the ion-blocking electrode, without any semicircles resulting from grain boundaries. This observation implies that the ionic pathway within the D-garnet electrolyte was constructed effectively. Surprisingly, the high ionic conductivity of 0.8 × 10^–4 S/cm was achieved for the D-garnet electrolyte, even after the mild heat-treatment at 400°C for 2 h, despite the presence of the lithium ion insulating LZO phase. Given an increase in temperature from 400°C to 600°C, the ionic conductivity of the D-garnet electrolyte increased to 2.4 × 10^–4 S/cm, which is comparable to previously reported garnet-type electrolytes sintered at high-temperatures, because of the improved phase purity and crystallinity. The high ionic conductivity of the D-garnet suggests that the static disorder observed in the low-temperature formed D-garnet, as confirmed in previous structural analyses, does not significantly impede lithium ion conduction. Moreover, the ionic conductivity was further increased to 3.3 × 10^–4 S/cm for the D-garnet electrolyte heat-treated under 600°C for 15 h. Figure 3b illustrates the Arrhenius plots of high-temperature processed conventional garnet (C-garnet) and D-garnet electrolytes (Figure S7). The C-garnet heat-treated under 600°C for 2 h, as expected, exhibited extremely low ionic conductivity of 4.2 × 10^–8 S/cm at 25°C with a critically high activation energy of 0.710 eV, attributed to imperfect ionic conduction pathway. In contrast, the D-garnet electrolytes demonstrated a remarkable improvement in ionic conductivity surpassing three orders of magnitude and exhibited a significantly reduced activation energy of 0.368 eV under the same heat-treatment condition.

We examined the stripping and plating behavior of lithium metal using the lithium symmetric cell configuration by applying a cold isostatic pressure (CIP) of 250 MPa to evaluate the stability of the D-garnet solid electrolyte against lithium metal. Figure 3c shows that the significantly reversible lithium stripping and plating behavior was observed at the current density of 0.1 mA/cm² for repeated 5000 cycles. The overpotential of 0.5 mV was consistently upheld during the repeated 5000 cycles (Figure S8), thereby affirming the remarkable (electro)chemical stability against lithium metal in the aspect of long-term durability.

We conducted a comparative experiment between two samples prepared through different sequences of densification (pelletizing) and crystallization in Fig. 3d to verify the impact of deformability of a-LLZTO. Note that a-LLZTO plays a crucial role in forming ion conduction pathways. First sample (denoted as pre-densification) underwent densification as the first step, thereby resulting in the creation of a dense amorphous matrix, followed by crystallization at 500°C for 2 h. Conversely, the other sample (denoted as post-densification) began with crystallization in a powder state under the same heat-treatment condition, followed by densification. Despite the identical overall process with differing orders, interestingly, there was a stark contrast in the observed ionic conductivity. While the first sample exhibited the desirable ionic conductivity of 1.8 × 10^–4 S/cm, the other one showed the significantly lower ionic conductivity of 4.2 × 10^–7 S/cm. Given the lack of a distinct phase differentiation between them (Fig. 3e), it is imperative to utilize deformability in constructing the matrix to unlock the potential for high ionic conductivity of garnet-type electrolytes at low temperatures.

Feasibility of disorder-driven garnet-solid electrolyte for lithium metal battery

We examined the electrochemical energy storage performance of the D-garnet solid electrolyte to validate its viability for use in a lithium metal battery. The a-LLZTO was hot-pressed under 375 MPa at 500°C for 2 h to maximize density, which resulted in a bulk-scale D-garnet solid electrolyte membrane (Figure S9). As depicted in Fig. 4a, the cross-sectional image of the D-garnet solid electrolyte illustrates a compact microstructure with a low porosity of 6.3%. We further confirmed that sacrificing the reduction in processing temperature, through hot-pressing at 600°C, resulted in a dense microstructure with a porosity of < 1.0% (Figure S10). We employed the D-garnet solid electrolyte hot-pressed at 500°C for demonstrating a hybrid lithium metal battery, paired with the LiCoO₂ (LCO) cathode. A wetted cathode with a non-aqueous ionic liquid (i.e., 2M Lithiumbis(fluorosulfonyl)imide (LiFSI) in 2 µl N-methyl-N-propylpyrrolidinium bis(fluorosulfonyl)imide (Pyr₁₃FSI)) was employed to ensure stable contact of the ionic pathway between the cathode and the solid electrolyte. In addition, to promote a uniform lithium-ion flux and homogeneous deposition on charging, we introduced a previously developed carbon interlayer with lithium metal at the anode side. In Fig. 4b, the charge–discharge profiles of the cell adopting the D-garnet solid electrolyte and the LCO cathode for initial five cycles (except the formation cycle; supplementary Figure S11) are depicted with a current density of 0.4 mA/cm² (corresponding to ~ 1.15 C-rate) at 60°C. In the first cycle, the reversible charge–discharge curve was obtained with an initial discharge capacity of 173.9 mAh/cm² with a Coulombic efficiency of 98.9%. After the first cycle, a highly reversible charge–discharge profiles were observed, and stable cycling was achieved for 100 cycles, without noticeable capacity decay, as shown in Fig. 4c. In addition, the stable cycling of the cell adopting the D-garnet solid electrolyte and LiNi_1/3Co_1/3Mn_1/3O₂ (NCM111) cathode was also attained in Figure S12.

We compared the ionic conductivity of the D-garnet solid electrolytes as a function of processing temperature with previously reported garnet-type solid electrolytes in Fig. 4d and Supplementary Note 2. A majority of the garnet-type solid electrolytes were positioned at the top right, as high-temperature processing exceeding 1000°C is mandatory to achieve high ionic conductivity. Moreover, it can be confirmed that the current research trends are shifting toward lowering processing temperatures on the left side, as indicated by the recent reports on the garnet-type solid electrolytes. However, certain limitations continue to persist because of the unsatisfactory processing temperature or low ionic conductivity. It should be emphasized that the D-garnet solid electrolytes developed in this work successfully achieved both the low processing temperature and high ionic conductivity simultaneously. The design strategy presented here will help demonstrate a low-temperature-processed solid electrolyte by detouring the need for conventional sintering and opening up the potentials for widespread large-scale application in the overall cell design.

A novel sintering-free garnet was successfully discovered by a simple mechano-chemical reaction accompanying comprehensive structural disordering. The sintering-free garnet enhances deformability for the formation of the amorphous matrix and lowers crystallization temperature. The amorphous structure exhibited drastically enhanced deformability, thereby creating a highly interconnected matrix reminiscent of the microstructure achieved through conventional high-temperature sintering process. In addition, the amorphous precursor enabled the facile formation of a desired cubic-phase at a notably lower temperatures compared to that of the conventional crystalline precursor. The combined in-depth analysis unveiled that the formation of the cubic-phase initiated at 350°C, with the emergence of the nano-sized crystalline particles within the amorphous matrix. The nucleation and growth within the compact matrix enabled the D-garnet electrolyte to achieve the ionic conductivity of 0.8 × 10^− 4 S/cm after a mild heat-treatment of 400°C for 2 h, and this further increased to 3.3 × 10^− 4 S/cm upon heat-treatment at 600°C for 15 h. The feasibility of the D-garnet solid electrolyte used in lithium metal batteries was verified based on stable electrochemical cycling in the lithium symmetric cell and hybrid full-cell. The current research highlights the strategy for achieving a low processing temperature by introducing a structural disorder in materials, which has significant potential for advancing the field of lithium metal batteries employing solid electrolytes.

Preparation of D-garnet solid electrolytes

Li₂O (99.5%), ZrO₂ (99.9%), Ta₂O₅ (99.85%), and La₂O₃ (99.9%) (Thermo Fisher Scientific, USA) were used as precursors to synthesize a-LLZTO. La₂O₃ was prepared by heat treatment in air at 900°C for 15 h before use. Solid electrolyte powders and pellets of LLZTO were purchased from Toshima Manufacturing Co., Ltd. (Japan) as a reference of the conventional cubic garnet.

The a-LLZTO was mechano-chemically synthesized by mixing the precursor powders in stoichiometric ratios with 0, 10, 20, and 50 mol% excess of the lithium source. The powders were ball-milled using a planetary mill under 370 rpm for 15 h with silicon nitride balls (Pulverisette-7 Premium Line, Fritsch, Germany). The ball-milling jar was sealed in an Ar-filled glove box to minimize air exposure.

The synthesized a-LLZTOs were pelletized under a uniaxial pressure of 1.3 GPa to create the dense amorphous matrix used to prepare the D-garnet solid electrolytes. Subsequently, the a-LLZTO pellets were crystalized under various heat-treatment conditions using a box furnace (AJ-SB4, Ajeon Furnace Control, Korea) in the air atmosphere. For a single-step hot-pressing process, the D-garnet solid electrolytes were prepared using a hot-press furnace (Ajeon Furnace Control, Korea) at 500 and 600°C under a pressure of 375 MPa for 2 h in an Ar gas atmosphere with a flow rate of 3 L/min. The prepared D-garnet solid electrolytes were treated in a dry room, where the dew point was maintained under − 60°C.

Characterizations

The crystal structure of synthesized materials was investigated using XRD. The XRD patterns were collected using a D8 Discover (Bruker, Germany) diffractometer with Cu–Kα radiation in the 2θ range of 10–90° at 1°/min. An in situ XRD analysis was conducted using an Empyrean (Malvern Panalytical, UK) diffractometer equipped with an HTK 1200N (Anton Paar, Austria) high-temperature chamber in aerobic atmosphere. Diffraction patterns were recorded every 100°C up to 1000°C with a heating rate of 10°C/min, holding the temperature for 10 min before the measurement at each step, employing Cu–Kα radiation in the 2θ range of 10–60° with a scan rate of 2.5°/min.

The mechanical property based on structural difference was investigated by comparing the compaction behavior of c-LLZTO and a-LLZTO using UTM (Instron, USA). Typical powder compaction behavior, following Heckel equation, shows three different regions: rearrangement, deformation and work hardening.⁴⁵ Among the three regions, yield pressure can be calculated from the deformation region, assuming that the compaction behavior follows first-order reaction kinetics.

$$\text{ln}\frac{\text{1}}{\text{ε}}\text{=kσ}{}_{\text{ax}}\text{+A}$$

$$\text{k=}\frac{\text{1}}{\text{3}{\text{σ}}_{\text{0}}}\text{=}\frac{\text{1}}{{\text{P}}_{\text{y}}}$$

where $\text{ε, σ}{}_{\text{ax}}\text{, }{\text{σ}}_{\text{0}}\text{,}\text{ }\text{and}\text{ }\text{P}{}_{\text{y}}\text{ }$represent porosity, axial compression stress, yield strength, and yield pressure, respectively. 0.1 g of sample powder was evenly placed inside the press mold with a diameter of 10 mm. The sample powder was compressed under displacement speed of 1 µm s^− 1 with relaxation until the pressure reaches 50, 100, 250, 500, 750 and 1000 MPa. The true density of sample powder was measured through gas pycnometer (AccuPyc II 1340, USA) for the UTM analysis.

SEM images were obtained using an SU-8030 FE-SEM (Hitachi, Japan), coupled with an EDS and EDS spectrometer (X-max 80, Oxford Instrument, UK), operated at an applied voltage of 3 kV. Cross-sectioned samples were prepared using a cross-section polisher (CP) (IB-19520CCP, JEOL, Japan) that polishes using an Ar⁺ ion beam at 6 kV with 200 µA for 8 h and all sample treatment was held. To prevent the contamination from air exposure, sample preparation was conducted in an Ar-purged glove box.

The DSC analysis was performed on a DSC 8000 (PerkinElmer, USA) in the temperature range of 25–500°C at a heating rate of 10°C/min.

In situ heating TEM analysis was conducted using an E-chip based Aduro heating holder with AXON system (Protochips Inc., USA) for image recoding and processing in double Cs-corrected Titan3 G2 60–300 microscope (Thermo Fisher Scientific, USA) at 300 keV. The temperature of the E-chip was increased in three steps: RT–350°C (10°C/s), 350–500°C (10°C/min), and 500–800°C (10°C/min), with a holding time of 10 min for each step to acquire TEM and SAED images.

The Raman spectra were generated using an inVia Raman microscope (Renishaw, UK) equipped with a 514-nm-excitation-laser source (~ 1 mW power).

Rietveld refinement was performed using FullProf software.

Solid-state MAS NMR experiments were performed on a Bruker Avance HD-III consoles corresponding to a 1H Larmor frequency of 700.52 MHz (B0 field 16.4T). Commercial Bruker double-resonance 2.5 mm MAS probes that allow spinning frequencies up to 35 kHz were used for all experiments. ⁷Li MAS NMR experiments were performed with a spinning frequency of 25 kHz.

Zr K-edge XAS were collected on BL10C beam line (using multiple wiggler source) at the Pohang light source (PLS-II) with top-up mode operation under a ring current of 250 mA at 3.0 GeV. The monochromatic X-ray beam could be obtained using liquid-nitrogen cooled Si(111) double crystal monochromator (Bruker ASC, Germany). For Zr K-edge XAFS measurements (absorption edge of 17998 eV), X-ray absorption spectroscopic data were recorded in a transmittance mode using ionization chamners (IC SPEC, FMB Oxford Ltd., UK) as photon detector. Higher order harmonic contaminations were eliminated by detuning to reduce the incident X-ray intensity by ~ 30%. Energy calibration has been carried out with reference Zr metal foil.

The XAFS data analysis were performed through the standard XAFS procedure.^46–49 Using AUTOBK module in UWXAFS package⁵⁰, the k³-weighted Zr K-edge EXAFS spectra, k³χ(k), have been obtained through background removal and normalization processes. The k³χ(k) spectra have been Fourier-transformed (FT) in the k ranges between 3.0 and 14.0 Å⁻¹. The experimental FT spectra have been inversely Fourier-transformed with the hanning window function in the r space range between 1.0 and 4.0 Å. To determine the EXAFS structural parameters for the first bond pairs, the curve-fitting process has been carried out by using the single bonding model. Theoretical single scattering paths of the first shells around central Te element have been calculated with FEFF9 code^51,52 under the space groups of I 41/a c d for the tetragonal Li₇La₃Zr₂O₁₂ model. In the EXAFS curve fitting process with FEFFIT module, total amplitude reduction factor, S₀², were fixed to 0.95 for the Zr K-edge XAFS, which were obtained after EXAFS fitting for metallic Zr metallic phase. The EXAFS structural parameters, interatomic distance (r), coordination numbers (N), Debye-Waller factor (σ²), have been determined within allowed R-factor value which is quality of the fit with {ReΔχk² + ImΔχk²}/{Re(χ_kdata)²+Im(χ_kdata)²}, where χ(k) is EXAFS-function) and Δχ(k) means χ(k)_data - χ(k)_{best−fitted}.

Computational details

Vienna Ab initio Simulation Package (VASP)⁵³ was used to conduct density functional theory (DFT) calculations. We used projector augmented wave pseudopotentials⁵⁴ as implemented in VASP, and generalized gradient approximation parametrized by Perdew-Burke-Ernzerhof⁵⁵ was utilized as an exchange-correlation functional. In order to model an a-LLZTO phase, we used a simulated melt-quench method using ab initio molecular dynamics (AIMD) simulation. First, crystalline cubic LLZTO structure was rapidly heated from 0 K to 3000 K in 2 ps, and the temperature was held at 3000 K for additional 6 ps. Then, the system was quenched to 300 K in 4 ps, and the structure was finally equilibrated for 2 ps at 300 K. All AIMD simulations were conducted under npt ensemble, and all input parameters and k-point meshes for DFT and AIMD calculations were generated with pymatgen⁵⁶, using pymatgen.io.vasp.sets module.

Preparation and assembly of electrochemical cells

The prepared D-garnet pellets were polished in the dry room using 400-, 800-, 3000-, and 7000-grit SiC abrasive paper for evaluating the electrochemical performance. After polishing, the pellet surface was immediately blow dried with dry air. To measure the ionic conductivity, a symmetric cell configuration was employed using blocking gold electrode sputtered on each surface of the pellet. For preparation of the Li symmetric cell, Li metal electrodes, 20-µm-thick Li metal on a 10-µm-thick copper foil (Honjo Metal Co. Ltd., Japan) with a diameter of 5 or 9 mm, were placed onto both sides of prepared pellet and the assembly was vacuum sealed in the dry room. Then, CIP of 250 MPa was applied for 3 min to induce the intimate contact between the pellet and lithium metal electrodes. Subsequently, the Li symmetric cell was assembled using 2032-type coin cell and sealed using manual coin cell hand crimper (Hohsen Corp., Japan). We fabricated the hybrid Li metal cell using a 4 mm diameter LCO cathode (Samsung SDI, Korea) (loading capacity: ~0.35 mAh/cm², active material: 80 wt%, Super-P carbon: 10 wt%, poly-vinylidene fluoride binder: 10 wt%), and an NCM111 cathode (loading capacity: ~3.2 mAh/cm², active material: 96 wt%, Samsung SDI, Korea) wetted with 3 µl of ionic liquid electrolyte (2M LiFSI (99.99%, PANAX ETEC Co., Ltd., Korea) in Pyr₁₃FSI (99.9% Kanto Chemical Co. Inc., Japan)). For preparing the anode side, Li metal and carbon interlayer, which was previously adopted,⁵⁷ were attached to the prepared D-garnet pellet by the CIP process. We positioned the cathode on the opposite side of the prepared pellet, and the hybrid lithium metal cell was constructed following the same procedure as the lithium symmetric cell. The cell performance was evaluated without additional external pressure.

Electrochemical measurement

Potentiostatic electrochemical impedance spectroscopy (PEIS) measurements were conducted using a frequency response analyzer (SI 1255 FRA, AMTEK SI, USA) in conjunction with a potentiostat (SI 1287 ECI, AMTEK SI, UK) to measure the ionic conductivity of the prepared solid electrolytes. The PEIS measurements were performed at an open-circuit voltage using an alternating current perturbation of 10 mV. The frequency ranges from 1 MHz to 1 Hz was used at various temperatures (25–60°C). For cycling a Li symmetric cell and hybrid Li metal cell, the galvanostatic electrochemical measurement was conducted at 60°C using a battery cycler (TOSCAT-3100, Toyo System, Japan). The cells were (dis)charged with a current density of 0.1, 0.4, and 0.6 mA/cm² for the Li symmetric cell, hybrid Li metal cell with LCO, and hybrid Li metal cell with NCM111, respectively.

Supporting Information

Supporting Information is available from or from the author.

Acknowledgements

This research was supported by funds from Samsung Electronics Co. Ltd.

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.

Conflict of interest

There are no conflicts to declare.

Boch, P. & Ni, J.-C. Ceramic materials: Processes, properties, and applications. Vol. 98 (John Wiley & Sons, 2010).
Buchanan, R. C. Ceramic materials for electronics. (CRC press, 2018).
Otitoju, T. A. et al. Advanced ceramic components: Materials, fabrication, and applications. J. Ind. Eng. Chem. 85, 34-65 (2020).
Conrad, H. J., Seong, W.-J. & Pesun, I. J. Current ceramic materials and systems with clinical recommendations: a systematic review. J. Prosthet. Dent. 98, 389-404 (2007).
Xiao, Z. et al. Materials development and potential applications of transparent ceramics: A review. Materials Science and Engineering: R: Reports 139, 100518 (2020).
Kharton, V., Marques, F. & Atkinson, A. Transport properties of solid oxide electrolyte ceramics: a brief review. Solid State Ion. 174, 135-149 (2004).
Bowen, P. & Carry, C. From powders to sintered pieces: forming, transformations and sintering of nanostructured ceramic oxides. Powder Technol. 128, 248-255 (2002).
Coble, R. & Burke, J. Sintering in ceramics. Prog. Ceram. Sci. 3, 197-251 (1963).
Hood, Z. D. et al. A sinter-free future for solid-state battery designs. Energy Environ. Sci. 15, 2927-2936 (2022).
Kuang, X., Carotenuto, G. & Nicolais, L. A review of ceramic sintering and suggestions on reducing sintering temperatures. Adv. Perform. Mater. 4, 257-274 (1997).
Krauskopf, T., Richter, F. H., Zeier, W. G. & Janek, J. r. Physicochemical concepts of the lithium metal anode in solid-state batteries. Chem. Rev. 120, 7745-7794 (2020).
Manthiram, A., Yu, X. & Wang, S. Lithium battery chemistries enabled by solid-state electrolytes. Nat. Rev. Mater. 2, 1-16 (2017).
Bachman, J. C. et al. Inorganic solid-state electrolytes for lithium batteries: mechanisms and properties governing ion conduction. Chem. Rev. 116, 140-162 (2016).
Banerjee, A., Wang, X., Fang, C., Wu, E. A. & Meng, Y. S. Interfaces and interphases in all-solid-state batteries with inorganic solid electrolytes. Chem. Rev. 120, 6878-6933 (2020).
Janek, J. & Zeier, W. G. Challenges in speeding up solid-state battery development. Nat. Energy 8, 230-240 (2023).
Xu, L. et al. Interfaces in solid-state lithium batteries. Joule 2, 1991-2015 (2018).
Wang, Y. et al. Design principles for solid-state lithium superionic conductors. Nat. Mater. 14, 1026-1031 (2015).
Kamaya, N. et al. A lithium superionic conductor. Nat. Mater. 10, 682-686 (2011).
Li, X. et al. Progress and perspectives on halide lithium conductors for all-solid-state lithium batteries. Energy Environ. Sci. 13, 1429-1461 (2020).
Wang, C. et al. Garnet-type solid-state electrolytes: materials, interfaces, and batteries. Chem. Rev. 120, 4257-4300 (2020).
Thangadurai, V., Narayanan, S. & Pinzaru, D. Garnet-type solid-state fast Li ion conductors for Li batteries: critical review. Chem. Soc. Rev. 43, 4714-4727 (2014).
Zhao, N. et al. Solid garnet batteries. Joule 3, 1190-1199 (2019).
Kim, K. J., Balaish, M., Wadaguchi, M., Kong, L. & Rupp, J. L. Solid‐state Li–metal batteries: challenges and horizons of oxide and sulfide solid electrolytes and their interfaces. Adv. Energy Mater. 11, 2002689 (2021).
Balaish, M. et al. Processing thin but robust electrolytes for solid-state batteries. Nat. Energy 6, 227-239 (2021).
Chen, R., Li, Q., Yu, X., Chen, L. & Li, H. Approaching practically accessible solid-state batteries: stability issues related to solid electrolytes and interfaces. Chem. Rev. 120, 6820-6877 (2019).
Xie, H. et al. Consolidating the grain boundary of the garnet electrolyte LLZTO with Li 3 BO 3 for high-performance LiNi 0.8 Co 0.1 Mn 0.1 O 2/LiFePO 4 hybrid solid batteries. J. Mater. Chem. A 7, 20633-20639 (2019).
Zhang, S. et al. Enhanced densification and ionic conductivity of Li-garnet electrolyte: efficient Li2CO3 elimination and fast grain-boundary transport construction. Chem. Eng. J. 393, 124797 (2020).
Wang, Y. et al. Effect of Al2O3 on the sintering of garnet-type Li6. 5La3Zr1. 5Ta0. 5O12. Solid State Ion. 294, 108-115 (2016).
Janani, N., Ramakumar, S., Kannan, S. & Murugan, R. Optimization of lithium content and sintering aid for maximized Li+ conductivity and density in Ta‐doped Li7La3Zr2O12. J. Am. Ceram. Soc. 98, 2039-2046 (2015).
Ohta, S. et al. Li+ conducting garnet-type oxide sintering triggered by an H+/Li+ ion-exchange reaction. J. Mater. Chem. A 8, 8989-8996 (2020).
Akimoto, J., Akao, T., Nagai, H. & Kataoka, K. Low-Temperature Sintering of a Garnet-Type Li6. 5La3Zr1. 5Ta0. 5O12 Solid Electrolyte and an All-Solid-State Lithium-Ion Battery. ACS Appl. Mater. Interfaces 15, 18973-18981 (2023).
Han, S. et al. A Full Oxide-Based Solid-State Lithium Battery and Its Unexpected Cathode Degradation Mechanism. ACS Energy Lett. 8, 4794-4805 (2023).
Ihrig, M. et al. Low temperature sintering of fully inorganic all-solid-state batteries–Impact of interfaces on full cell performance. J. Power Sources 482, 228905 (2021).
Huang, X. et al. Manipulating Li2O atmosphere for sintering dense Li7La3Zr2O12 solid electrolyte. Energy Storage Mater. 22, 207-217 (2019).
Pfenninger, R., Struzik, M., Garbayo, I., Stilp, E. & Rupp, J. L. A low ride on processing temperature for fast lithium conduction in garnet solid-state battery films. Nat. Energy 4, 475-483 (2019).
Sastre, J. et al. Blocking lithium dendrite growth in solid-state batteries with an ultrathin amorphous Li-La-Zr-O solid electrolyte. Commun. Mater. 2, 76 (2021).
Song, S. et al. Isomeric Li–La–Zr–O Amorphous–Crystalline Composite Thin-Film Electrolytes for All-Solid-State Lithium Batteries. ACS Appl. Energy Mater. 4, 8517-8528 (2021).
Shelby, J. E. Introduction to glass science and technology. (Royal society of chemistry, 2020).
Holand, W. & Beall, G. H. Glass-ceramic technology. (John Wiley & Sons, 2019).
Fernandez‐Martin, C. et al. Nucleation and growth of nanocrystals in glass‐ceramics: an in situ SANS perspective. J. Am. Ceram. Soc. 95, 1304-1312 (2012).
Paronen, P. Heckel Plots as Indicators of Elastic Properties of Pharmaceuticals. Drug Dev. Ind. Pharm. 12, 1903-1912 (1986).
Mahmoodi, F., Klevan, I., Nordström, J., Alderborn, G. & Frenning, G. A comparison between two powder compaction parameters of plasticity: the effective medium A parameter and the Heckel 1/K parameter. Int. J. Pharm. 453, 295-299 (2013).
Sonnergaard, J. A critical evaluation of the Heckel equation. Int. J. Pharm. 193, 63-71 (1999).
Hooper, D., Clarke, F., Mitchell, J. & Snowden, M. A modern approach to the Heckel Equation: The effect of compaction pressure on the yield pressure of ibuprofen and its sodium salt. J. Nanomed. Nanotechnol. 7, 2 (2016).
Ghori, M. U. & Conway, B. R. Powder compaction: compression properties of cellulose ethers. Br. J. Pharm. 1, 19-29 (2016).
Rehr, J. J. & Albers, R. C. Theoretical approaches to x-ray absorption fine structure. Rev. Mod. Phys. 72, 621 (2000).
Rehr, J. J. & Albers, R. C. Scattering-matrix formulation of curved-wave multiple-scattering theory: Application to x-ray-absorption fine structure. Phys. Rev. B 41, 8139 (1990).
Newville, M. IFEFFIT: interactive XAFS analysis and FEFF fitting. J. Synchrotron Rad. 8, 322-324 (2001).
Ravel, B. & Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Rad. 12, 537-541 (2005).
Stern, E., Newville, M., Ravel, B., Yacoby, Y. & Haskel, D. The UWXAFS analysis package: philosophy and details. Physica B: Condensed Matter 208, 117-120 (1995).
Rehr, J. J., Kas, J. J., Vila, F. D., Prange, M. P. & Jorissen, K. Parameter-free calculations of X-ray spectra with FEFF9. Phys. Chem. Chem. Phys. 12, 5503-5513 (2010).
Ankudinov, A. L., Ravel, B., Rehr, J. & Conradson, S. Real-space multiple-scattering calculation and interpretation of x-ray-absorption near-edge structure. Phys. Rev. B 58, 7565 (1998).
Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169 (1996).
Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953 (1994).
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865 (1996).
Ong, S. P. et al. Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis. Comput. Mater. Sci. 68, 314-319 (2013).
Kim, S. et al. High-power hybrid solid-state lithium–metal batteries enabled by preferred directional lithium growth mechanism. ACS Energy Lett. 8, 9-20 (2022).

There is NO Competing Interest.

aLLZTOSI.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Disorder-driven Sintering-free Garnet-type Solid Electrolytes

Status:

Version 1

Abstract

Figures

Introduction

Results and discussion

Design strategy and disorder-driven soft mechanical properties

Structural evolution of amorphous precursor for garnet

Construction of ion conduction pathway within an amorphous matrix

Feasibility of disorder-driven garnet-solid electrolyte for lithium metal battery

Conclusion

Methods

Preparation of D-garnet solid electrolytes

Characterizations

Computational details

Preparation and assembly of electrochemical cells

Electrochemical measurement

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1