Solid–Liquid Interfacial Properties Related to Ionic Conductivity of Mixtures of Metal Oxide Particles and Lithium Bis(fluorosulfonyl)amide-Sulfone Electrolytes

Abstract

We evaluated and compared the physical properties of electrolyte solutions consisting of lithium bis(fluorosulfonyl)amide and sulfone solvents (molar ratio = 1 : 3) and their mixtures containing aluminum oxide (α-Al₂O₃) or lithium lanthanum zirconate (LLZ) particles of various particle sizes. Sulfolane (SL), 3-methylsulfolane, and ethyl isopropyl sulfone were evaluated as sulfone solvents for the electrolyte solutions. The phase-change heat, phase-change temperature, and spin–spin relaxation time in nuclear magnetic resonance (NMR) measurements decrease in a mixture of SL electrolyte with a metal oxide, with an apparent average liquid thickness in the order of nm resulting from the SL electrolyte solution. This indicates a decrease in molecular mobility around the particle surface. For the α-Al₂O₃ system, no substantial changes are observed in the activation energy of ionic conductivity, self-diffusion coefficient of Li⁺ (determined via pulsed-field gradient NMR), or relative cross-peak intensities of Li⁺ and ¹H of SL in the two-dimensional NMR of the mixture. Therefore, despite its low molecular mobility, the SL electrolyte solution at the solid–liquid interface is considered to exhibit an ionic conductivity mechanism similar to that of the bulk electrolyte. It was suggested that LLZ system has a different ionic conduction mechanism than α-Al₂O₃ system.

1. Introduction

The properties of ion-conductive liquids mixed with solid materials have been extensively investigated.^1–6 Quasi-solid electrolytes are solid–liquid mixtures consisting of solid electrolyte particles and electrolyte solutions, and they can be used in lithium-ion batteries.⁷ Various materials, such as ionic liquids, have been proposed to improve the ionic conductivity between solid electrolyte particles in quasi-solid electrolyte systems.^8–11 To enhance the ionic conductivity of the quasi-solid electrolyte, the high ionic conductivity of the electrolyte solution must also be considered.¹ Sulfone solvents have a wide potential window; particularly, sulfolane (SL) has been investigated as an electrolyte for lithium secondary batteries.^12–16 Recently, the properties of concentrated SL electrolytes have been studied to determine their suitability as solvation ionic liquids.^17–19 Dokko et al.²⁰ demonstrated “Li⁺ hopping conduction”; the self-diffusion coefficient of Li⁺ (D_Li) exceeds that of the BF₄⁻ anion (D_BF4) and SL solvent (D_SL) in concentrated LiBF₄/SL electrolyte; related observations about Li⁺ hopping conduction mechanism have been reported. For example, considering a solution of lithium bis(fluorosulfonyl)amide (LiFSA) and SL, hopping conduction occurs when the concentration of Li⁺ is higher than the molar ratio of LiFSA : SL = 1 : 4.²⁰ Furthermore, a quasi-solid electrolyte consisting of fumed nano-silica and SL electrolyte has been proposed as a practical example of a concentrated SL-based electrolyte.²¹ To develop a quasi-solid electrolyte with enhanced ionic conductivity, the interaction between the solid particles and the electrolyte at their interface must be investigated.

The physical properties of liquids are known to change near the surface of solid particles; for example, numerous systematic studies have been conducted with reference to the behavior of ionic liquids on solid surfaces, activation energy (E_a) of the ionic conductivity of molten salts on metal oxide surfaces,^22,23 and differential scanning calorimetry (DSC) analysis.²⁴ In particular, the physical properties of liquids change appreciably in a limited space, such as a micropore. Previously, we evaluated mixtures of a solid oxide electrolyte, that is, garnet-type lithium lanthanum zirconate powder^25–27 (LLZ or LLZO), and FSA-based ionic liquids, with apparent average liquid thicknesses in the order of nm.²⁸ For these mixtures, we reported a decrease in the heat of fusion of the ionic liquid, based on thermal analysis, and a decrease in the spin–spin relaxation time (T₂) of FSA⁻, based on ¹⁹F nuclear magnetic resonance (NMR) evaluation.²⁸ Although these phenomena are similar to previously reported changes in liquid properties on solid surfaces,²³ the changes in the physical properties of electrolytes consisting of organic solvents (including SL) remain unclear. Elucidating the physical properties of these interfaces can potentially provide practical insights into the material design of quasi-solid electrolytes.

In this study, we focused on the changes in the physical properties near the interface of solid particles and sulfone-based electrolyte solutions, primarily SL, and attempted to evaluate and characterize their properties. Aluminum oxide (α-Al₂O₃) systems, which do not exhibit Li⁺ conductivity, and the Li⁺-conductive cubic LLZ were used as solid particles. We evaluated these particle systems at various ratios of solid and electrolyte solutions to determine common properties regardless of Li⁺ conductivity. Herein, we discuss the effect of the symmetric molecular structure of the sulfone solvents on the structure of the solid particle surface, based on evaluations of 3-methylsulfolane (MSL) and ethyl isopropyl sulfone (EiPS) (Fig. 1 and Table 1), and the influence of the lithium salt concentration in the SL system. Furthermore, the molecular mobility of the SL electrolyte components and the proximity between the SL solvents and Li⁺ are discussed based on NMR measurements.

Figure 1.

Chemical structure of sulfone solvents tested in this study.

Table 1. Physical properties of sulfone solvents.

Sulfones	Molecular Weight	Melting point/°C	Boiling point/°C	Relative permittivity	Density /g cm−3
SL	120.17	2030 2712 2829	28512 28630	42.913 43 (30 °C)30 60 (20 °C)12	1.26 (25 °C)29 1.26 (40 °C)30
MSL	134.19	629	27629	29.4 (25 °C)29	—
EiPS	136.21	<−2012 −1530	23730 265a,12	56 (30 °C)30	1.0930

^aReduced under atmospheric pressure by nomograph.

2. Experimental

2.1 Materials

α-Al₂O₃ particles of various sizes were purchased from KANTO CHEMICAL Co., Inc.; KISHIDA CHEMICAL Co., Ltd.; Kojundo Chemical Laboratory Co., Ltd.; and SUMITOMO CHEMICAL Co., Ltd. The particles were used after vacuum drying at 180 °C for 3 h. Samples of Mg- and Sr-doped cubic LLZ (target composition based on stoichiometric ratio: Li_6.95Mg_0.15La_2.75Sr_0.25Zr₂O₁₂, hereinafter “LLZ”; ionic conductivity of sintered pellet = 1.4 × 10⁻³ S cm⁻¹ at 25 °C²⁶) were prepared as described in our previous report.²⁷ LLZ samples of various particle sizes and specific surface areas were obtained by adjusting the grinding conditions in a bead mill. Table S1 lists the median diameters (D50) and Brunauer–Emmett–Teller specific surface areas of the α-Al₂O₃ and LLZ powders. It has been reported that LLZ particles forms contaminants such as Li₂CO₃ on the surfaces when exposed to the moist air.³¹ As a result of X-ray photoelectron spectroscopy (XPS) (PHI Quantes; ULVAC-PHI) analysis of the surface of this LLZ, similar C1s spectrum was obtained to the literature,³² and the area ratio of the C1s peak corresponding to carbonate was approximately 6 atom%, suggesting that the ionic conductivity of the LLZ surface is retained.

Battery-grade SL and MSL were purchased from KISHIDA CHEMICAL Co., Ltd. and used as received. Regular-grade EiPS (Tokyo Chemical Industry Co., Ltd.) was used as received only for thermal analysis. A test electrolyte was prepared by dissolving battery-grade LiFSA (KISHIDA CHEMICAL Co., Ltd.) and sulfone solvent in a predetermined molar ratio in an Ar-atmosphere glove box. Herein, for example, the test electrolyte comprising a mixture of LiFSA and SL at a molar ratio of 1 : 3 is expressed as LiFSA : SL = 1 : 3. Subsequently, α-Al₂O₃ or LLZ and the test electrolytes were mixed in an agate mortar for 10 min in an Ar-atmosphere glove box to prepare the solid–liquid mixtures.

2.2 Measurement of ionic conductivity

For the liquid electrolyte solutions, approximately 1 mL of test electrolyte was transferred into a plastic symmetric cell (SB1A-S; EC FRONTIER Co., Ltd.) with two Au-sputtered stainless steel (SUS304) plates (thickness = 20 µm, diameter = 15 mm), and the impedance was measured with a VSP-300 potentiostat (Bio-Logic Science Instruments) in the scan range of 7 MHz to 100 mHz and sine-wave amplitude of 10 mV in a temperature chamber. The ionic conductivity of the test electrolyte was calculated from the cell constant, which was obtained from the resistance of a standard sample of 0.01 M KCl aqueous solution (HORIBA, Ltd.) at 25 °C.

The ionic conductivity of the mixture was measured using a method similar to that used in our previous study.²⁸ Initially, 150 mg of the α-Al₂O₃ mixture or 120 mg of the LLZ mixture was placed between two steel punches with a diameter of 10 mm and pressed in a polycarbonate cylinder at 76 MPa for 5 min to obtain pellets. The electrochemical impedance of the pellet was measured between the punches at a pressure of 76 MPa at room temperature (25 ± 2 °C) under the same electric condition as that considered for the electrolyte solution. To measure the activation energy (E_a) of the ionic conductivity of the pellet, the impedance was measured in a temperature chamber without applying the pressure of 76 MPa on the pellet between the punches. The ionic conductivity of the pellet was calculated based on its resistance, thickness, and cross-sectional area.

2.3 DSC measurements

For the DSC measurements (Thermo plus EVO2 DSCvesta; Rigaku Corp.), approximately 10 mg of the solid–liquid mixture sample was placed in an Al sample pan under an Ar atmosphere. The temperature scan was initially performed at room temperature, followed by cooling to −120 °C and heating to 60 °C, at a scan rate of 10 °C min⁻¹. Heat flow (W g⁻¹) and phase-change heat were calculated based on the net weight of the sulfone solvent (not including weights of LiFSA and oxide particles) in the sample (described as W g_-sulfone⁻¹ or J mol_-sulfone⁻¹). The glass transition temperature (T_g) was recorded as the temperature at the inflection point in the heating curve for samples that did not exhibit a distinct phase-change peak.

2.4 NMR measurements

2.4.1 Relaxation time

For ¹H, ⁷Li, and ¹⁹F nuclei, the spin–lattice relaxation time (T₁) was evaluated via the inversion recovery method,³³ and the spin–spin relaxation time (T₂) was evaluated via the Carr–Purcell–Meiboom–Gill (CPMG) method³⁴ at 25 °C. The liquid sample was placed in a glass sample tube with an outer diameter of 5 mm and filled to a height of 40 mm. The solid–liquid mixture sample was placed in a symmetrical microsample tube (BMS-005J, SHIGEMI Co., Ltd.) filled to a height of 5 mm. These samples were analyzed using an ECZ700R instrument (JEOL RESONANCE Inc.; ¹H resonance frequency = 700 MHz, 16.45 T). The chemical shifts were calibrated using tetramethylsilane dissolved in deuterated chloroform (¹H), lithium chloride aqueous solution (⁷Li), and trifluoromethylbenzene (¹⁹F; recorded as 0 ppm for trichlorofluoromethane) as references. The detailed measurement conditions are listed in Table S2.

2.4.2 Self-diffusion coefficient of Li⁺

For both the liquid electrolyte and mixture samples, the BMS-005J sample tube was filled to a height of 5 mm and measured via a pulsed-field gradient (PFG) stimulated echo method,³⁵ using a solid-state NMR apparatus (ECA600II; JEOL RESONANCE Inc.; ⁷Li resonance frequency = 233 MHz, 14.1 T) with a diffusion probe at 25 °C. The self-diffusion coefficient of Li⁺ (D_Li+) was calculated from the slope of the signal intensity for different pulsed-field gradients. Chemical shifts of ⁷Li were calibrated as described in section 2.4.1. The detailed measurement conditions are listed in Table S3.

2.4.3 Two-dimensional NMR

For the liquid electrolyte, heteronuclear Overhauser effect spectroscopy (HOESY)³⁶ measurements were performed with the ECZ700R instrument. The sample was prepared as described in section 2.4.1; observation nucleus = ⁷Li, irradiation nucleus = ¹H (described as ⁷Li{¹H} HOESY) at 25 °C. For the mixture sample, Lee–Goldburg cross polarization (LGCP)-heteronuclear correlation (HETCOR)³⁷ measurements were performed using ECA600II with a 3.2 mm HXMAS probe. The mixture sample was placed in a ZrO₂ rotor (outer diameter = 3.2 mm) and rotated at 20 kHz for measurement; observation nucleus = ⁷Li, irradiation nucleus = ¹H (described as ⁷Li{¹H} LGCP-HETCOR) at 25 °C. Chemical shifts were calibrated as described in section 2.4.1. The detailed measurement conditions are listed in Table S4.

3. Results and Discussion

3.1 Ionic conductivity

For the mixture consisting of LLZ particles and ionic liquids, we assumed the existence of an adsorbed layer with a relatively low molecular mobility. In this study, we defined the apparent average liquid thickness (t_liquid), which is derived from the liquid volume (V_liquid) and surface area of solid metal oxide particles (S_particle), as t_liquid = V_liquid/S_particle. We evaluated the physical properties of the liquids around the solid–liquid interface by preparing mixture samples with various t_liquid values. In order to evaluate ionic conductivity only of the electrolyte at the solid–liquid interface, without considering Li⁺ conductivity of the particle, therefore, not only LLZ but also α-Al₂O₃ of various sizes were used as solid particles to measure the ionic conductivity. First, we prepared mixed samples with various t_liquid values by changing the solid : liquid ratio of the SL electrolyte (LiFSA : SL = 1 : 3), which is considered to demonstrate Li⁺ hopping conductivity; the relationship between t_liquid and ionic conductivity was evaluated. We compared four types of α-Al₂O₃ and LLZ particles of different median diameters (D50) and specific surface areas. The ionic conductivities of these samples and their solid–liquid volume ratios are shown in Fig. 2. For example, a mixture with a composition such as that shown in Fig. 2A (a), [α-Al₂O₃ (D50 = 14.8 µm) with volume fraction = 79] and [LiFSA : SL = 1 : 3 with volume fraction = 21], is hereinafter denoted as “(D50 14.8 µm α-Al₂O₃) : (LiFSA : SL = 1 : 3) = volume ratio 79 : 21”. The ionic conductivity of this powder-rich composition (79 : 21, Fig. 2A (a)) is one-tenth of that of the liquid-rich composition (61 : 39, Fig. 2 (b)), although t_liquid of (a) is higher than that of (b). As reported previously,²⁸ this result implies that the ionic conductivity of a powder-rich mixture includes the effect of the loss of continuity of ionic conduction paths in the mixture, which can be explained by applying the percolation theory to ionic conductivity.^38–40 Therefore, the magnitude of t_liquid is not necessarily reflective to that of the ionic conductivity, particularly in powder-rich compositions. Ionic conductivity of the LLZ mixture (Fig. 2B) was evaluated as bulk ionic conductivity (σ_bulk) calculated from bulk resistance (R_bulk) of inside LLZ particle and liquid electrolyte, and total ionic conductivity (σ_total) calculated from R_bulk + interfacial resistance (R_int) including LLZ particle boundary resistance and blocking electrode interfacial resistance (Fig. S2). σ_bulk of LLZ mixture were higher than α-Al₂O₃ mixture with comparable t_liquid. σ_total of LLZ mixture, more practical value, was remarkably enhanced from pressed LLZ particle without liquid electrolyte (under 10⁻⁴ mS cm⁻¹) although it was still lower than sintered LLZ pellet (1.4 mS cm⁻¹). Therefore, liquid electrolyte is thought to be contributed in ionic conduction in solid-liquid mixture of LLZ. Solid-state NMR studies on LLZ and polymer electrolyte composites reported that Li⁺ transport pathways transit from polymer to ceramic routes with increasing LLZ content, although the ion mobility decreases.⁴¹ A similar ionic conduction mechanism is assumed for the mixture of LLZ and sulfone electrolyte.

Figure 2.

Relationships between t_liquid and ionic conductivity for different D50 particles and their solid : liquid volume ratio, (A) α-Al₂O₃ and (LiFSA : SL = 1 : 3) mixture and (B) LLZ and (LiFSA : SL = 1 : 3) mixture. (a) an example of relatively powder-rich composition; α-Al₂O₃(D50 = 14.8 µm) : (LiFSA : SL=1 : 3) = volume ratio 79 : 21, (b) an example of relatively liquid-rich composition; α-Al₂O₃(D50 = 0.804 µm) : (LiFSA : SL = 1 : 3) = volume ratio 61 : 39.

Based on the above results shown in Fig. 2, ionic conductivity was measured and E_a was evaluated for the composition with the highest possible liquid ratio within the range where pellets could be formed; simultaneously, α-Al₂O₃ particles with large surface area were selected to control t_liquid at the lowest possible value to reflect the solid–liquid interfacial properties. We also evaluated the effects of the sulfone solvent structure. Figure 3 shows the ionic conductivity (σ_bulk) and their E_a of the electrolyte solutions, LiFSA : SL = 1 : 3 and LiFSA : MSL = 1 : 3, and those of the pellets with t_liquid = 3.1 nm, (D50 0.11 µm α-Al₂O₃) : (LiFSA : SL = 1 : 3) = volume ratio 6 : 4 and (D50 0.11 µm α-Al₂O₃) : (LiFSA : MSL = 1 : 3) = volume ratio 6 : 4. The ionic conductivities of the SL-based samples were 1.5–1.8 times higher than those of the MSL-based samples, irrespective of mixing with α-Al₂O₃. The higher molecular weight of MSL (Table 1) possibly increases the intermolecular force and the viscosity. However, we did not focus on comparing the ionic conductivities because we prepared samples with constant molar ratios of lithium salt and sulfone solvent (instead of volume molar concentration) to emphasize the consistency of the hopping conduction mechanism. The E_a values associated with the ionic conductivity did not appreciably increase owing to solid–liquid mixing in both sulfone electrolyte systems, compared with the results of an α-Al₂O₃ and ionic liquid system investigated by Mizuhata et al.²² Therefore, although ionic conduction may depend on the chemical species of the liquid, the ionic conduction mechanisms in the vicinity of the solid–liquid interface and in the bulk electrolyte are considered similar, at least for the mixture of α-Al₂O₃ and sulfone-based electrolytes. For the case of mixture of LLZ ((D50 0.872 µm LLZ) : (LiFSA : SL = 1 : 3) = volume ratio 6 : 4) shown in Fig. 4, especially, the E_a of σ_bulk decreased by about 30 % compared to the electrolyte alone (LiFSA : SL = 1 : 3). E_a of σ_total in the mixture of LLZ was approximately similar to the LiFSA : SL = 1 : 3. These results suggest that the ionic conduction mechanism and ion conduction path are different from those of the electrolyte solution. Additionally, in Figs. 3 and 4, α-Al₂O₃ and LLZ particles were each chosen to represent the minimum t_liquid, respectively. This α-Al₂O₃ is corresponding to the experiments in the section 3.4. Therefore, since α-Al₂O₃ in Fig. 3 and LLZ in Fig. 4 mutually differ in particle size and specific surface area, it should be noted that the magnitudes of these ionic conductivity and E_a values cannot be directly compared.

Figure 3.

Temperature dependency of ionic conductivity and the corresponding E_a for sulfone electrolytes (t_liquid = 3.1 nm) and liquid-rich mixtures with α-Al₂O₃. (a) σ_bulk of LiFSA : SL = 1 : 3, (b) σ_bulk of LiFSA : MSL = 1 : 3, (c) σ_bulk of (D50 0.11 µm α-Al₂O₃) : (LiFSA : SL = 1 : 3) = volume ratio 61 : 39, t_liquid = 3.1 nm, (d) σ_bulk of (D50 0.11 µm α-Al₂O₃) : (LiFSA : MSL = 1 : 3) = volume ratio 61 : 39, t_liquid = 3.1 nm.

Figure 4.

Temperature dependency of ionic conductivity and the corresponding E_a for sulfone electrolytes (t_liquid = 3.1 nm) and liquid-rich mixtures with LLZ. (a) σ_bulk of LiFSA : SL = 1 : 3, (e) σ_bulk of (D50 0.872 µm LLZ) : (LiFSA : SL = 1 : 3) = volume ratio 60 : 40, t_liquid = 32 nm, and (f) σ_total of (D50 0.872 µm LLZ) : (LiFSA : SL = 1 : 3) = volume ratio 60 : 40, t_liquid = 32 nm.

3.2 Evaluation of phase-change behavior using DSC

Thermal analysis is useful for analyzing liquids at solid–liquid interfaces. For example, interactions between metal oxide and eutectics have been discussed using differential thermal analysis.²³ In the present study, for the combination of two metal oxides (α-Al₂O₃, or LLZ) and three types of electrolytes consisting of sulfone solvents (SL, MSL, or EiPS), we evaluated the interactions between metal oxide and sulfone electrolyte via DSC measurements. Initially, the differences in the thermal behaviors of sulfone solvents and sulfone electrolyte solutions were compared. The DSC curves of the sulfone solvents and their solutions with the lithium salt (solvent : LiFSA = molar ratio 1 : 3) at increasing temperatures are shown in Fig. 5. For example, SL solvents exhibit endothermic peaks below its melting point, which is attributed to solid-state phase change; additionally, a sharp endothermic melting peak is observed. Although LiFSA : SL = 1 : 3 shows peaks that can be attributed to phase change, melting, or eutectic point, only a glass transition is observed for LiFSA : MSL = 1 : 3 and LiFSA : EiPS = 1 : 3, which is similar to previously reported results.⁴² The Li⁺-coordinated structures of MSL and EiPS are considered difficult to crystallize because of their relatively high conformational freedom, whereas SL has a symmetric molecular structure. This is related to the order of their melting points (EiPS < MSL < SL), as listed in Table 1. However, it should be noted that even in the system in which only the glass transition was observed, the possibility of a first-order transition occurring depending on the temperature scanning conditions cannot be denied.

Figure 5.

DSC curves of sulfone solvents and their electrolyte solutions with LiFSA. (a) SL, (b) LiFSA : SL = 1 : 3, (c) MSL, and (d) LiFSA : MSL = 1 : 3, (e) EiPS, (f) LiFSA : EiPS = 1 : 3. Heat flow is based on sulfone weight (g_-sulfone); temperature scan rate = 10 °C min⁻¹.

For the LiFSA : SL = 1 : 3 mixture, which exhibits phase-change peaks, the DSC curves (Figs. S3 and S4), phase-change heat, and phase-change temperature (Figs. 6A and 6B) are shown for samples with various t_liquid values. Assuming that the phase-change heat is caused only by the sulfone solvents, the heat flow (W g_-sulfone⁻¹) and phase-change heat (J mol_-sulfone⁻¹) are calculated from the net weight of the sulfone solvent (g_-sulfone). As shown in Fig. 5, t_liquid is considered one of the dominant parameters for determining the thermal properties of sulfone electrolytes on the metal oxide surface, because particles with different D50 values and chemical compositions demonstrate similar relationships between t_liquid and thermal behavior, in contrast to the ionic conductivity discussed in section 3.1. In this system, the phase-change heat and temperature decrease when t_liquid < 200 nm. Therefore, the region of t_liquid below 200 nm may be affected by the interaction between the metal oxide surface and the sulfone electrolyte. For the mixture of sulfones with lower symmetry in the molecular structure, the DSC curves (Figs. S5–S8) and the relationship between t_liquid and T_g (Figs. 6C and 6D) are evaluated. With reference to the LiFSA : EiPS = 1 : 3 mixture, T_g values for t_liquid < 10 (Figs. S8 (g) and S8 (h)) were not obtained because the inflection points of the DSC were unclear. The region influenced by the metal oxide surface does not appear to be considerably affected by the molecular structural symmetry of the sulfone solvent, although the inflection point of T_g versus t_liquid is unclear for the LiFSA : MSL = 1 : 3 mixture.

Figure 6.

t_liquid dependence of (A) phase-change heat in LiFSA : SL = 1 : 3 mixture, (B) phase-change temperature in LiFSA : SL = 1 : 3 mixture, (C) T_g in LiFSA : MSL = 1 : 3 mixture, and (D) T_g in LiFSA : EiPS = 1 : 3 mixture.

3.3 Influence of Li salt concentration and solid–liquid interaction

Tsurumaki et al.¹¹ reported that DSC peak intensity of an ionic liquid was prominent when the ionic liquid was mixed with aluminum-doped LLZ. The structural order of the ionic liquid was found to be enhanced in the LLZ composites,¹¹ suggesting that the electrostatic interaction between the ions and the solid surface affects the thermal behavior of the electrolyte solution. Similarly, we confirmed sharp DSC peaks for the solid–liquid mixture (Figs. S3 (b)–S3 (j)), possibly indicating that sulfone electrolytes also form a structural order on the surface of the metal oxides. However, because the DSC peaks are broadened in the nanometer region of t_liquid (Figs. S3 (k)–S3 (m)), the formation of a structural order cannot be concluded from only the DSC peak associated with the sulfone electrolyte system. The effect of the Li salt concentration on the DSC profile was then evaluated to assess the influence of the interaction between the ions and the metal oxide surface. Besides, a similar sharpening of the exothermic peak was observed when the temperature decreasing, but the phase-change temperature of the peak fluctuated by approximately ±10 °C regardless of the t_liquid. Therefore, thermal behavior in temperature rising was used for discussions.

We compared the thermal behavior of the system with a low Li salt concentration (LiFSA : SL = 1 : 10) and the SL solvent without the Li salt. The corresponding DSC curves are shown in Figs. S9–S12, and the relationship between t_liquid and the phase-change heat is shown in Fig. 7. The DSC curve of the LiFSA : SL = 1 : 10 system changes substantially with t_liquid, in contrast to that of the LiFSA : SL = 1 : 3 system. Therefore, only the phase-change heat is evaluated for the LiFSA : SL = 1 : 10 system because the phase-change temperature of the LiFSA : SL = 1 : 10 system is difficult to determine. The inflection point of the phase-change heat versus t_liquid is ∼200 nm, similar to that of the LiFSA : SL = 1 : 3 system. However, for the SL solvent without the Li salt, both the phase-change heat and temperature decrease below t_liquid = ∼50 nm. The solid–liquid interaction in the system consisting of SL solvent without Li salt is decreased compared with that of Li salt-containing systems. These results indicate that the ions in the SL electrolyte play an important role in this interaction. In addition, the parameter t_liquid is an average value that is different from the actual individual thickness of the liquid on the metal oxide particle; therefore, the observed t_liquid may be influenced by the extremely small thickness of the liquid. The region of the organic electrolyte affected by the electrode surface is calculated to be approximately 1 nm.⁴³

Figure 7.

t_liquid dependence of (A) phase-change heat in LiFSA : SL = 1 : 10 mixture^a and (B) phase-change temperature in SL mixture. ^aPhase-change heat data for t_liquid > 180 nm in (A) (Figs. S9 (a)–S9 (c) and Figs. S10 (a)–S10 (c)) were evaluated using both endothermic and exothermic peaks.

With reference to the metal oxide, for example, the oxide composition affected E_a of ionic conductivity in a system of metal oxides and molten salts.²³ In the present study, the thermal behavior of α-Al₂O₃ and LLZ is not considerably different in SL, MSL, or EiPS systems, although the ionic character based on the electronegativity of constituent elements of α-Al₂O₃ and LLZ is different. Accordingly, more precise measurements and analyses are desirable in future studies.

3.4 NMR evaluation of SL electrolyte at the solid–liquid interface

For evaluating the molecular mobility via NMR, for example, the self-diffusion behavior can be estimated from the plot of T₁ against the reciprocal of temperature.⁴⁴ Furthermore, PFG-NMR has been used to evaluate the self-diffusion coefficient (D) of electrolytes.^45–47 T₂ indicates molecular mobility;⁴⁸ a low molecular mobility decreases T₂ owing to the increased interaction among surrounding nuclei. In our previous study, we observed that T₂* (T₂ value calculated from the signal width) considerably decreased in a mixture of α-Al₂O₃ and the ionic liquid as t_liquid decreased.²⁸ In this study, we evaluated LiFSA : SL = 1 : 3 and (D50 0.110 µm α-Al₂O₃) : (LiFSA : SL = 1 : 3) = volume ratio 61 : 39 (t_liquid = 3.1 nm), with the smallest t_liquid among those prepared in this study, to compare the molecular mobility of the bulk electrolyte and on the metal oxide surface. For these samples, T₁ was evaluated via the inversion recovery method, T₂ via the CPMG method, and D_Li+ via PFG-NMR. Only α-Al₂O₃ was used as the metal oxide powder owing to the difficulty in distinguishing between Li⁺ of LLZ and Li⁺ of LiFSA via NMR. Table 2 presents the results of the study. The T₁ for the electrolyte components in the solid–liquid mixture (t_liquid = 3.1 nm) are longer than those in the electrolyte solution, implying that the solid–liquid mixture presents a greater difference between the resonance frequencies of each nucleus and their time scale of molecular motion. Thus, changes occur in the molecular motion between the electrolyte solution and solid–liquid mixture. T₂ is observed to decrease substantially at less than 1/100 in the solid–liquid mixture (t_liquid = 3.1 nm) compared with that of the electrolyte solution, indicating that the electrolyte near the solid–liquid interface exhibits low molecular mobility, similar to the solid phase. This corresponds to a decrease in the phase-change heat in the DSC measurement. In contrast, the decrease in D_Li+ in the solid–liquid mixture is only 20 % of that in the electrolyte solution. The ionic conductivity of the solid–liquid mixture is affected by the continuity of the ionic conduction paths (Section 3.1), whereas the evaluation of D through PFG-NMR is considered to be less influenced by the distribution of the electrolyte in the mixture. Assuming that the transference number of Li⁺ does not change between this solid–liquid mixture and the bulk electrolyte, we speculate that the ionic conductivity is retained, although the molecular mobility decreases in the region around t_liquid = 3.1 nm.

Table 2. T₁, T₂ and D_Li+ of electrolyte solution and mixtures (25 °C).

Samples	T1/s			T2/ms			DLi+/m s−2
	1H	7Li	19F	1Ha	7Li	19F
LiFSA : SL = 1 : 3	1.32	1.26	0.140	1620	505	40.3	1.35 × 10−11
α-Al2O3 mixtureb	1.55	2.16	0.167	0.753	1.64	0.358	1.08 × 10−11

^aEvaluated at 2.257 ppm.

^b(D50 0.11 µm α-Al₂O₃) : (LiFSA : SL = 1 : 3) = volume ratio 61 : 39, t_liquid = 3.1 nm.

These results indicate that the Li⁺ coordination structure in the solid–liquid mixture is similar to that in the bulk electrolyte. For confirmation, we used two-dimensional NMR to evaluate the proximity of Li⁺ and SL in these systems. HOESY experiments are typically used to evaluate interactions among ions, for example, cations and anions in ionic liquids,⁴⁹ carbonate electrolyte,⁵⁰ and electrode material.⁵¹ We subjected the quasi-solid electrolytes to HOESY and HETCOR experiments. The results of the ⁷Li{¹H} HOESY experiment for LiFSA : SL = 1 : 3 electrolyte and ⁷Li{¹H} LGCP-HETCOR experiment for (D50 0.110 µm α-Al₂O₃) : (LiFSA : SL = 1 : 3) = volume ratio 61 : 39 (t_liquid = 3.1 nm) are shown in Figs. 8A and 8B, respectively. The signals of hydrogen atoms were assigned according to the literature.⁵² The chemical shifts of Li⁺ and H atoms in the solid–liquid mixed sample (Fig. 8B) increase by approximately 1 ppm from the liquid electrolyte (Fig. 8A). Li⁺ is closer to the α-hydrogen atoms of the sulfonyl group of SL (H_a) than to the β-hydrogen atoms (H_b) in both systems, possibly coordinating with the sulfonyl group, because H_a has a higher cross-peak intensity with Li⁺ than H_b. Therefore, the solid–liquid interface also possibly has a solvation structure and hopping conduction mechanism similar to that of the bulk electrolyte solution; consequently, D_Li+ may approach that of the bulk electrolyte, despite the lower molecular mobility at the solid–liquid interface.

Figure 8.

(A) ⁷Li{¹H} HOESY experiment for LiFSA : SL = 1 : 3 and (B) ⁷Li{¹H} LGCP-HETCOR experiment for (D50 0.110 µm α-Al₂O₃) : (LiFSA : SL = 1 : 3) = volume ratio 61 : 39, t_liquid = 3.1 nm.

4. Conclusions

In this study, sulfone electrolytes on metal oxide particles exhibited higher ionic conduction and diffusion properties than expected from their lower molecular mobility owing to changes in thermal properties and NMR relaxation behaviors. The phase-change behavior indicates that the t_liquid region is less than 200 nm and is influenced by the interaction with the solid surface, although the low liquid thickness region also affects the actual system. The presence of ions in the electrolyte solution considerably influences its interaction with the solid surface. The differences in the chemical structures of the sulfone solvents also affect their thermal phase-change behaviors, whereas the difference in the metal oxide species, such as α-Al₂O₃ and LLZ, does not have a remarkable influence. For the mixture of α-Al₂O₃, the ionic conduction mechanism on the metal oxide surface is possibly similar to that in the bulk electrolyte solution because of the similar E_a of ionic conductivity and two-dimensional NMR spectra. It is conceivable that a conductive path different from the electrolyte solution contributes to the bulk conductive component of the mixture of LLZ. Furthermore, it will be important to decrease ion transfer resistance passing through the LLZ surface and the liquid electrolyte. These results are expected to contribute to the development and design of quasi-solid electrolyte materials based on the properties of solid–liquid interfaces. We anticipate further development of evaluation methods applicable to various interfaces in complex and dynamically changing solid–liquid interfaces, as in actual battery systems.

Acknowledgments

The authors would like to thank to Mr. Junpei KONDO (Niterra) for his suggestion and development of the mixture system of LLZ-sulfone based electrolyte, and Mr. Masanori NAKANISHI (Niterra) for preparation of various size of LLZ powder. The authors would like to thank to also R&D members of Niterra for helpful advices and discussions. We would like to thank Editage (www.editage.com) for English language editing of a draft of this manuscript.

Data Availability Statement

The data that support the findings of this study are openly available under the terms of the designated Creative Commons License in J-STAGE Data at https://doi.org/10.50892/data.electrochemistry.22730429.

• 1) We evaluated and compared the physical properties of electrolyte solutions consisting of lithium bis(fluorosulfonyl)amide and sulfone solvents (molar ratio = 1 : 3) and their mixtures containing aluminum oxide (α-Al₂O₃) or lithium lanthanum zirconate (LLZ) particles of various particle sizes. Sulfolane (SL), 3-methylsulfolane, and ethyl isopropyl sulfone were evaluated as sulfone solvents for the electrolyte solutions. The phase-change heat, phase-change temperature, and spin–spin relaxation time in nuclear magnetic resonance (NMR) measurements decrease in a mixture of SL electrolyte with a metal oxide, with an apparent average liquid thickness in the order of nm resulting from the SL electrolyte solution. This indicates a decrease in molecular mobility around the particle surface. For the α-Al₂O₃ system, no substantial changes are observed in the activation energy of ionic conductivity, self-diffusion coefficient of Li⁺ (determined via pulsed-field gradient NMR), or relative cross-peak intensities of Li⁺ and ¹H of SL in the two-dimensional NMR of the mixture. Therefore, despite its low molecular mobility, the SL electrolyte solution at the solid–liquid interface is considered to exhibit an ionic conductivity mechanism similar to that of the bulk electrolyte. It was suggested that LLZ system has a different ionic conduction mechanism than α-Al₂O₃ system.

CRediT Authorship Contribution Statement

Yuu Watanabe: Data curation (Lead), Investigation (Lead), Writing – original draft (Lead), Writing – review & editing (Lead)

Yuki Takeuchi: Investigation (Equal), Project administration (Lead)

Masato Taki: Writing – review & editing (Equal), NMR measurement (Lead)

Hidetoshi Mizutani: Supervision (Equal)

Masato Iwasaki: Supervision (Equal)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Footnotes

Y. Watanabe, Y. Takeuchi, H. Mizutani, and M. Iwasaki: ECSJ Active Members

References

• 1)  S. Deki, M. Mizuhata, K. Nakamura, A. Kajinami, and Y. Kanaji, J. Electrochem. Soc., 139, 1544 (1992).
• 2)  M. Mizuhata, M. Kitamura, and S. Deki, Electrochemistry, 71, 1093 (2003).
• 3)  H. Ogawa, A. Unemoto, and I. Honma, Electrochemistry, 80, 765 (2012).
• 4)  Y. Gambe, Y. Sun, and I. Honma, Sci. Rep., 5, 8869 (2015).
• 5)  Y. Ji, C. Zhou, F. Lin, B. Li, F. Yang, H. Zhu, J. Duan, and Z. Chen, Materials, 13, 560 (2020).
• 6)  A. Sagara, H. Yabe, X. Chen, B. Put, T. Hantschel, M. Mees, H. Arase, Y. Kaneko, A. Uedono, and P. M. Vereecken, ACS Appl. Mater. Interfaces, 13, 40543 (2021).
• 7)  H. Munakata and K. Kanamura, Jisedai-nijidenchi no Kaihatsu-doukou, Kadai, Shourai-tenbou, AndTech, Kanagawa, Chap. 1 (2022). [in Japanese]
• 8)  H. W. Kim, P. Manikandan, Y. J. Lim, J. H. Kim, S. C. Nam, and Y. S. Kim, J. Mater. Chem. A, 4, 17025 (2016).
• 9)  E. J. Cheng, T. Kimura, M. Shoji, H. Ueda, H. Munakata, and K. Kanamura, ACS Appl. Mater. Interfaces, 12, 10382 (2020).
• 10)  A. Paolella, G. Bertoni, W. Zhu, D. Campanella, A. L. Monaca, G. Girard, H. Demers, A. Cristina, G. Nita, Z. Feng, A. Vijh, A. Guerfi, M. Trudeau, M. Armand, and S. A. Krachkovskiy, J. Am. Chem. Soc., 144, 3442 (2022).
• 11)  A. Tsurumaki, R. Rettaroli, L. Mazzapioda, and M. A. Navarra, Appl. Sci., 12, 7318 (2022).
• 12)  K. Xu and C. A. Angell, J. Electrochem. Soc., 149, A920 (2002).
• 13)  V. S. Kolosnitsyn, E. V. Karaseva, D. Y. Seung, and M. D. Cho, Russ. J. Electrochem., 39, 1089 (2003). Translated from Elektrokhimiya, 39, 1218 (2003).
• 14)  J. Alvarado, M. A. Schroeder, M. Zhang, O. Borodin, E. Gobrogge, M. Olguin, M. S. Ding, M. Gobet, S. Greenbaum, Y. S. Meng, and K. Xu, Mater. Today, 21, 341 (2018).
• 15)  B. Kurc, Int. J. Electrochem. Sci., 13, 5938 (2018).
• 16)  J. Fu, X. Ji, J. Chen, L. Chen, X. Fan, D. Mu, and C. Wang, Angew. Chem., Int. Ed., 59, 22194 (2020).
• 17)  K. Yoshida, M. Nakamura, Y. Kazue, N. Tachikawa, S. Tsuzuki, S. Seki, K. Dokko, and M. Watanabe, J. Am. Chem. Soc., 133, 13121 (2011).
• 18)  C. A. Angell, Y. Ansari, and Z. Zhao, Faraday Discuss., 154, 9 (2012).
• 19)  Y. Yamada and A. Yamada, Electrochemistry, 82, 1085 (2014).
• 20)  K. Dokko, D. Watanabe, Y. Ugata, M. L. Thomas, S. Tsuzuki, W. Shinoda, K. Hashimoto, K. Ueno, Y. Umebayashi, and M. Watanabe, J. Phys. Chem. B, 122, 10736 (2018).
• 21)  A. Unemoto, E. Seki, K. Ueno, K. Dokko, and D. Watanabe, Kokai Tokkyo Koho, JP 2021-89875 (2021). [in Japanese]
• 22)  M. Mizuhata, K. Yaso, Y. Ito, and S. Deki, Electrochemistry, 73, 606 (2005).
• 23)  M. Mizuhata, H. Kubo, Y. Ichikawa, H. Maki, and M. Matsui, J. Electrochem. Soc., 168, 046516 (2021).
• 24)  K. Nishikawa, Toyota Kenkyū Houkoku, 73, 103 (2020). [in Japanese]
• 25)  Y. Inaguma, Denchi Handbook (Ed. The Electrochemical Society of Japan, The Committee of Battery Technology), Ohmsha, Tokyo, Chap. 8.4.8 (2010). [in Japanese]
• 26)  Y. Takeuchi, H. Hikosaka, D. Shishihara, and H. Mizutani, The 13th Pacific Rim Conference of Ceramic Societies, Program Book, 30-Theater-S20-02 (2019).
• 27)  Y. Takeuchi, H. Ando, D. Shishihara, and H. Mizutani, JP Patent, 6632240 (2020). [in Japanese]
• 28)  Y. Watanabe, Y. Takeuchi, M. Taki, H. Mizutani, and M. Iwasaki, Electrochemistry, 90, 097002 (2022).
• 29)  M. Ue, Y. Sasaki, Y. Tanaka, and M. Morita, Electrolytes for Lithium and Lithium-Ion Batteries. Modern Aspects of Electrochemistry (Eds. T. R. Jow, K. Xu, O. Borodin, and M. Ue), Vol. 58, Springer, New York, Chap. 2.5.2, p. 146 (2014), referred on Jan 30, 2023.
• 30)  H. Senoh, H. Sakaebe, H. Sano, M. Yao, K. Kuratani, N. Takeichi, and T. Kiyobayashi, J. Electrochem. Soc., 161, A1315 (2014).
• 31)  R. L. Vold, J. S. Waugh, M. P. Klein, and D. E. Phelps, J. Chem. Phys., 48, 3831 (1968).
• 32)  E. J. Cheng, M. Shoji, T. Abe, and K. Kanamura, iScience, 25, 103896 (2022).
• 33)  H. Y. Carr and E. M. Purcell, Phys. Rev., 94, 630 (1954).
• 34)  P. C. M. V. Zijl and C. T. W. Moonen, J. Magn. Reson., 87, 18 (1990).
• 35)  W. Bauer, Magn. Reson. Chem., 34, 532 (1996).
• 36)  V. Ladizhansky and S. Vega, J. Chem. Phys., 112, 7158 (2000).
• 37)  Y. Inaguma, J. Yu, T. Katsumata, and M. Itoh, J. Ceram. Soc. Jpn., 105, 548 (1997).
• 38)  S. L. Zelinka, S. V. Glass, and D. S. Stone, Wood Fiber Sci., 40, 544 (2008).
• 39)  N. Yabuuchi, M. Nakayama, M. Takeuchi, S. Komaba, Y. Hashimoto, T. Mukai, H. Shiiba, K. Sato, Y. Kobayashi, A. Nakao, M. Yonemura, K. Yamanaka, K. Mitsuhara, and T. Ohta, Nat. Commun., 7, 13814 (2016).
• 40)  J. Li and R. Wang, Ceram. Int., 47, 13280 (2021).
• 41)  J. Zheng and Y. Hu, ACS Appl. Mater. Interfaces, 10, 4113 (2018).
• 42)  Y. Ugata, S. Sasagawa, R. Tatara, K. Ueno, M. Watanabe, and K. Dokko, J. Phys. Chem. B, 125, 6600 (2021).
• 43)  L. Xing, J. Vatamanu, O. Borodin, G. D. Smith, and D. Bedrov, J. Phys. Chem. C, 116, 23871 (2012).
• 44)  Y. Mizuno and N. Hanafusa, Low Temperature Science. Series A, Physical Sciences, 45, 1 (1987). [in Japanese]
• 45)  K. Hayamizu and Y. Aihara, Mater. Jpn., 38, 555 (1999). [in Japanese]
• 46)  L. O. Valøen and J. N. Reimers, J. Electrochem. Soc., 152, A882 (2005).
• 47)  K. Damodaran, Annu. Rep. NMR Spectrosc., 88, 215 (2016).
• 48)  N. Bloembergen, E. M. Purcell, and R. V. Pound, Phys. Rev., 73, 679 (1948).
• 49)  R. Nanda and K. Damodaran, Magn. Reson. Chem., 56, 62 (2018).
• 50)  V. Kumar, R. R. Reddy, B. V. N. P. Kumar, C. V. Avadhani, S. Ganapathy, N. Chandrakumar, and S. Sivaram, J. Phys. Chem. C, 123, 9661 (2019).
• 51)  E. Zhang, N. Fulik, S. Paasch, L. Borchardt, S. Kaskel, and E. Brunner, Energy Storage Mater., 19, 432 (2019).
• 52)  N. R. Babij, El. O. McCusker, G. T. Whiteker, B. Canturk, N. Choy, L. C. Creemer, C. V. D. Amicis, N. M. Hewlett, P. L. Johnson, J. A. Knobelsdorf, F. Li, B. A. Lorsbach, B. M. Nugent, S. J. Ryan, M. R. Smith, and Q. Yang, Org. Process Res. Dev., 20, 661 (2016).