2025 Volume 93 Issue 4 Pages 047001
There is considerable interest in all-solid-state sodium-ion batteries from the perspectives of safety and resource efficiency. However, it is known that the interfacial resistance of oxide-based solid electrolytes such as Na1+xZr2SixP3−xO12 (NASICON), which is used in all-solid-state sodium-ion batteries, is high because of poor contact at the solid/solid interface. In this study, to reduce the interfacial resistance, we conceived of introducing a ductile electrolyte, Na(CB9H10)0.7(CB11H12)0.3 (NaCBH), Na2.25Y0.25Zr0.75Cl6 (NYZC), and a polymer electrolyte (PE), into the NASICON/hard carbon interface. The ionic conductivity of NaCBH was the highest among three ductile electrolytes. In impedance measurements using a symmetric cell, the interfacial resistance was high for NYZC but low for NaCBH and the PE. This suggests that good contact was formed at the NaCBH/NASICON and NaCBH/PE interfaces and that NaCBH is an excellent electrolyte as an intermediate layer. In addition, when NaCBH and hard carbon were combined, charged, and discharged, a capacity comparable to that of organic electrolytes was obtained. Furthermore, the interfacial resistance between the solid electrolyte and hard carbon was reduced by introduction of NaCBH, proving that NaCBH is an effective intermediate layer on the anode side for use in all-solid-state sodium-ion batteries.
Numerous studies have been conducted on all-solid-state batteries, which have the advantages of high safety, wide operating temperature range, and resistance to degradation. Among these, all-solid-state sodium-ion batteries have advantages from the perspective of element strategy and have been widely researched.1–3 Many types of solid electrolytes with sodium ion conductivity are being investigated, and among these, oxide-based electrolytes are expected to be able to suppress cell short-circuiting and thermal runaway caused by the deposition of sodium metal, as they have high hardness and extremely low electronic conductivity.4 Especially the solid electrolyte called the sodium superionic conductor (NASICON), which is composed of Na1+xZr2SixP3−xO12, has a higher conductivity than other solid electrolytes such as Na-β′′-Al2O3 because it can diffuse Na+ ions in three dimensions.5–7 In actual all-solid-state cells, high conductivity on the order of 10−3 S cm−1 has been achieved at high temperatures,8 but many studies use sodium metal as the electrode, and there are still safety issues to be addressed. However, if a Na+ ion storage-type active material, such as hard carbon, which is generally used in batteries with liquid electrolytes,9,10 is employed, it is thought that the rigid particle nature of the solid electrolyte will result in a small contact area and large charge-transfer resistance at the interface. Therefore, in order to achieve the practical application of all-solid-state sodium-ion batteries while making use of the high sodium-ion conductivity and hardness that NASICON possesses, it is important to improve the interface contact between the hard carbon anode and the solid electrolyte to reduce the interface resistance.
One way to reduce the interfacial resistance is to insert an intermediate layer at the interface between the all-solid-state electrolyte and the electrode. One type of intermediate layer is a ductile electrolyte. Closo-boranes [BnHn]2− and closo-carboranes [CBn−1Hn]− are weakly coordinated anions with delocalized charge. Lithium and sodium salts of these compounds have high conductivity and, in particular, high superionic conductivity on the order of 10−2 S cm−1 has been reported for the sodium salt.11 Thus, attention was focused on the use of these compounds as all-solid-state electrolytes for sodium-ion batteries. It was reported that a mixture of NaCB9H10 and NaCB11H12 was effective as an electrolyte, considering the phase transition temperature at which the conductivity decreased and the conductivity.11 In addition, halogen ions generally have a large ionic radius and weak interaction with cations; thus, they are advantageous for cationic conduction, and in 2018, it was discovered that Li13YCl6 and Li3YBr6 have a conductivity of 0.5–0.7 mS cm−1 as Li+ conductive solid electrolytes.12 Furthermore, it has been reported that conductivity can be increased by introducing vacancies through Zr substitution.13,14 Na3−xY1−xZrxCl6, which is applied to sodium-ion batteries, has also been synthesized and has a relatively high conductivity.15 The use of a polymer as an intermediate layer is another way to reduce the interfacial resistance. Polymer electrolytes (PEs) have properties intermediate between those of liquid and solid electrolytes while offering superior safety and formability; therefore, they attract considerable research attention.16,17 In particular, polymers containing fluorine, such as polyvinylidene difluoride (PVDF), have high polarity and therefore exhibit a relatively high dielectric constant. In addition, high conductivity has been achieved by using anions with delocalized electrons, such as FSA− and TFSA−, as electrolytes within the polymer.18
In this study, we attempted to reduce the resistance at the electrode/electrolyte interface in all-solid-state batteries by introducing ductile electrolytes. The optimal electrolyte for the intermediate layer should have high ionic conductivity and form a good interface with the solid electrolyte and electrode. Therefore, we first searched for an electrolyte that forms a good interface with the solid electrolyte. In addition, to ensure adequate ion supply to the inside of the electrode, a composite electrode was used, in which the electrode and electrolyte were mixed. To search for an effective electrolyte as a composite electrode material, the ion-transfer reactions at the electrode/electrolyte interface were analyzed.
Na3Zr2Si2PO12 (NASICON) was synthesized from Na2CO3 (Nacalai Tesque), SiO2 (Nacalai Tesque), ZrO2 (Kojundo Chemical, 98 %), and NH4H2PO4 (Nacalai Tesque) using the solid-phase method, as previously reported.19 The resulting NASICON powder was pressed in a Newton press (NT-100H, NPa Systems, Inc.) at 45 MPa for 5 min and then in a cold hydrostatic isostatic press (CPA-50-300, NPa Systems, Inc.) at 300 MPa to produce NASICON pellets. Na(CB9H10)0.7(CB11H12)0.3 (NaCBH) was made by mixing each carborane (Katchem) in a ball mill. Na2.25Y0.25Zr0.75Cl6 was synthesized from NaCl (Nacalai Tesque), YCl3 (Kojundo Chemical, 99.9 %), and ZrCl4 (Kojundo Chemical, 99.9 %) in a glovebox under an Ar atmosphere using a solid-phase method, as previously reported.14 X-ray diffraction (XRD) measurements were performed on the solid electrolytes using a RINT-2200 instrument (Rigaku). The PVDF PEs were prepared using the casting method. Sodium bis(fluorosulfonyl)amide (NaFSA, Kishida Chemical) and PVDF were mixed at a salt : monomer unit molar ratio of 1 : 2 and dried in an Ar atmosphere, referring to a previous report.20 N,N-Dimethylformamide (DMF, Wako Pure Chemicals) was used as the solvent. The PEs were subsequently prepared via drying under vacuum conditions at 60 °C for 20 h. Thermogravimetry (TG) and XRD measurements were performed for characterization. TG measurements were performed using STA200RV (HITACHI).
Alternating-current (AC) impedance measurements were performed on these electrolytes using a two-electrode method to evaluate the conductivity. The frequency range was 100 kHz–1 Hz, and the applied AC voltage was 50 mV. Both sides of NASICON pellets were covered with gold to form blocking electrodes by a sputtering method (sputtering time: one minute). The pellets of NaCBH and NYZC were prepared in the measuring cell at the pressure of 350 MPa under Ar atmosphere, and the pellets were sandwiched between two sheets of Au foil (Nilaco, thickness: 20 µm).
A two-electrode symmetric cell was constructed to investigate the ion-transfer resistance at the NASICON/electrolyte interface. A NASICON/soft electrolyte/NASICON cell was constructed using NaCBH and NYZC sandwiched between NASICON layers. Polymer/NASICON/polymer cells were constructed using SUS as the blocking electrode because PEs are considered to have good adhesion at the interface. AC impedance measurements were performed on these cells in the frequency range of 1 MHz to 10 mHz.
A half-cell was constructed using NaCBH. The working electrode was a composite electrode composed of a mortar mixture of hard carbon (HC, Kuraray) and NaCBH at a weight ratio of 4 : 6. A Cu foil (Housen) was used as the current collector, which was pressed and molded at 200 MPa in the cell to form the composite electrode. The composite electrodes were prepared in Ar atmosphere. NASICON was used as the solid electrolyte, and Na metal (Sigma–Aldrich) was used as the counter electrode. A NaCBH layer was introduced between the composite electrode and NASICON pellets. Cells without NaCBH (HC/NASICON/Na) and without NASICON (HC + NaCBH composite electrode/NaCBH/Na) were constructed in the same manner. Furthermore, the 2032-type coin cell was assembled using HC composite electrode as the working electrode, Na metal as the counter electrolyte, and 1 mol dm−3 NaPF6 (Kishida Chemical)/ethylene carbonate (EC) + diethyl carbonate (DEC) (1 : 1 by vol.) (Kishida Chemical). Charge–discharge measurements were performed on these cells at a current density of 3 mA g−1, a cutoff voltage of 0 and 2.0 V, and a measurement temperature of 60 °C. After the charge–discharge measurements, impedance measurements were performed in the frequency range of 100 kHz–10 mHz. HJ1001SD8 (Hokuto Denko) was used for charge–discharge measurements, and Solartron 1470E+1255 (Solartron Analytical) was used for impedance measurements. All cells for electrochemical measurements were assembled under Ar atmosphere.
The density of NASICON synthesized via the solid-phase method was 3.17 g cm−3, which is close to the theoretical density (3.27 g cm−3). The XRD pattern of the obtained NASICON powder is shown in Fig. 1. Although some impurity ZrO2 peaks were observed,21 the desired material was obtained.22 AC impedance measurements were performed (Fig. S1a), and the conductivity was calculated from the overall resistance because the bulk and grain-boundary resistances could not be distinguished. The real-axis intercept obtained from the approximate straight line was used as the resistance value, and the conductivity was calculated. The obtained conductivity at 30 °C was 1.7 × 10−4 S cm−1, in agreement with the literature.6 The activation energy was 29 kJ mol−1 (Table 1, Fig. S2a).
| NASICON | NaCBH | NYZC | Polymer | |
|---|---|---|---|---|
| σ/S cm−1 | 1.7 × 10−4 | 3.0 × 10−2 | 8.3 × 10−6 | 1.5 × 10−4 |
| Ea/kJ mol−1 | 29 | 19 | 39 | 41 |
The XRD pattern of NaCBH powder is shown in Fig. 2a. Compared with the reported XRD pattern,23 there were no significant differences in peak position or intensity, indicating that no decomposition occurred. AC impedance measurements were performed using NaCBH (Fig. S1c), and the conductivity was calculated (Fig. S1b). The obtained value was 3.0 × 10−2 S cm−1 at 30 °C, and the activation energy of conductivity was 19 kJ mol−1 (Table 1, Fig. S2c), indicating that the conductivity was significantly higher than that of NASICON. This may have been due to the weak coordination of carborane ions, which favors cation transfer, and the fact that molecular crystals can be densified by compaction. The XRD pattern of the NYZC is shown in Fig. 2b. Although impurity peaks were observed, the results confirmed that target material was synthesized. The conductivity at 30 °C was 8.3 × 10−6 S cm−1, and the activation energy of conductivity was 39 kJ mol−1 (Table 1, Fig. S2c), which is close to the previously reported value.15
XRD patterns of (a) NaCBH and (b) NYZC.
TG curves of PE are shown in Fig. S3. For comparison, the results of PE with the NaFSA : monomer unit ratio of 1 : 3 was also shown. The weight loss at around 153 °C, which is the boiling point of DMF, indicated that only a small amount of DMF remained in PEs. In addition, the weight loss near 300 and 450 °C were due to the decomposition of NaFSA24 and carbonization of PVDF.25 In this study, a PE with the ratio of 1 : 2, which has a high sodium ion concentration, was used. The XRD analysis was performed using a PE. The polymers were analyzed using a sample cast on a Kapton film. For comparison, XRD measurements were also performed on the PVDF polymer and NaFSA electrolyte, and the corresponding patterns are shown in Fig. S4. The NaFSA peak disappeared from the XRD pattern of the PE, suggesting that the sodium salt was sufficiently dissolved in the polymer. The 18° and 20° peaks, which are characteristic of the PVDF polymer, also disappeared, suggesting that the PE became amorphous. The conductivity of the PE determined via impedance measurements (Fig. S1d) was 1.5 × 10−4 S cm−1, and the activation energy of conductivity was 41 kJ mol−1 (Table 1, Fig. S2d). This is attributed to the high relative permittivity of PVDF, which increases the concentration of sodium salt and thus the number of carrier ions.
The resistance at the electrolyte interface was subsequently investigated using a symmetric cell. Figure 3a shows the Nyquist plot of the NASICON/NaCBH/NASICON symmetric cell. No semicircular arcs appeared in the resulting Nyquist plot, only straight lines. The resistance was determined as the intercept of the real axis of the straight line, and the temperature dependence was investigated. The Arrhenius plots are shown in Fig. S5a. From the slope of the approximate straight line, the activation energy was found to be 30 kJ mol−1 (Table 2), which is close to the activation energy of NASICON conductivity. Because the conductivity of NaCBH is two orders of magnitude higher than that of NASICON, NASICON contributes a large fraction of the bulk resistance. The ion-transfer resistance at the NASICON/NaCBH interface was not confirmed by the Nyquist plots. This may be because the ion-transfer resistance at the interface was far smaller, as the interface contact was well secured by the soft NaCBH. In conclusion, the cell resistance at the NASICON/NaCBH/NASICON is mainly attributed to the bulk and grain-boundary resistances of NASICON, and the ion-transfer resistance at the NASICON/NaCBH interface was small.
Nyquist plots of symmetric cells: (a) NASICON/NaCBH/NASICON, (b) NASICON/NYZC/NASICON, and (c) PE/NASICON/PE.
| NaCBH | NYZC | Polymer | |
|---|---|---|---|
| Ea/kJ mol−1 | 30 | 40 | 37 |
The high-frequency region of the Nyquist plot of the NASICON/NYZC/NASICON symmetric cell is shown in Fig. 3b. The activation energies obtained from the slopes of the Arrhenius plots (Fig. S5b), where the real-axis intercept corresponds to bulk intergranular resistance, are close to the value of the conductivity of NYZC at 40 kJ mol−1 (Table 2). The Nyquist plots of the bulk grain-boundary resistivity of NYZC are similar to those of NASICON. A semicircular arc is also observed on the low-frequency side of the Nyquist plot (Fig. S6a). This is attributed to the ion-transfer resistance of NASICON/NYZC. The activation energy of this arc was derived from an Arrhenius plot (Fig. S6b) and was 54 kJ mol−1. These results indicate that NYZC is more ductile and easier to form than solid oxide electrolytes because of the halogen ions but harder to form than NaCBH. Thus, NYZC has a smaller contact area with NASICON, resulting in a larger interfacial resistance.
The Nyquist plot for PE/NASICON/PE is shown in Fig. 3c. The temperature dependence was assessed using the real-axis intercept obtained from the approximately straight line representing the blocking electrode behavior, and the activation energy was derived from the Arrhenius plot (Fig. S5c). The obtained value was 37 kJ mol−1 (Table 2). As there was no significant difference between the conductivities of NASICON and the PE, this resistance was considered to include the bulk and grain-boundary resistances of both NASICON and the PE. However, as with NaCBH, there was no semicircular arc on the low-frequency side, suggesting that the improved interfacial adhesion in the PE also increased the contact area, reducing the ion-transfer resistance at the interface.
The conditions required for an intermediate layer between the solid electrolyte and electrodes must have both high conductivity and the ability to form a good interface with the solid electrolyte and electrodes. Among the three types of intermediate-layer candidates used in this study, NaCBH and the PE formed a good interface with the solid electrolyte and reduced the interfacial resistance. In addition, NaCBH has high conductivity, and when introduced as an intermediate layer, it does not hinder the movement of sodium.
Finally, hard carbon composite electrodes were prepared using NaCBH, which had the lowest interfacial resistance to NASICON among the intermediate layers investigated in this study, and a half-cell (Na/NASICON/NaCBH/(HC + NaCBH)) was constructed, and electrochemical measurements were performed. The results of the constant-current charge–discharge measurement at a current density of 3 mA g−1 are shown in Fig. 4a. The capacity obtained was 200 mAh g−1. The reversible capacity obtained from the charge–discharge measurements of the hard carbon composite electrode using the organic electrolyte (1 mol dm−3 NaPF6/EC + DEC) was approximately 230 mAh g−1 (Fig. S7), and a capacity comparable to that of the liquid was obtained. The irreversible capacity in the 1st cycle is thought to be due to decomposition of the surface oxygen functional groups of HC. For comparison, we measured the charge–discharge properties of a cell without NaCBH (Na/NASICON/HC) (Fig. 4b). When NaCBH was not used, the capacity was limited to approximately 50 mAh g−1. The composite electrode layer was approximately 20 µm thick, and the active material layer was approximately two layers thick, and it is thought that the thickness of this layer limited the utilization of the active material. However, through mixing with NaCBH, the high ionic conductivity of NaCBH allows ion transport deep into the electrode, increasing the utilization of the active material.
Charge–discharge curves of the cells: (a) Na/NASICON/NaCBH/(NaCBH + HC) and (b) Na/NASICON/HC. Current density: 3 mA g−1. Temperature: 60 °C.
In addition, impedance measurements were performed on a Na/NaCBH/(HC + NaCBH) cell after charging and discharging. A capacity of approximately 200 mAh g−1 was obtained in this cell (Fig. S8), similar to the Na/NASICON/NaCBH/(HC + NaCBH) cell. The results of the impedance measurements performed at a voltage of 0.1 V are shown in Fig. 5a. There was one semicircular arc in the Nyquist plot. This arc exhibits potential dependence (Fig. S9), which is attributed to the charge-transfer resistance at the NaCBH/HC interface. The arc also contains the charge-transfer resistance of Na/NaCBH, but its value is so small that it can be ignored. The intercept of the real axis on the high-frequency side is assigned to the bulk and grain-boundary resistances. The activation energies for these resistances were calculated from the slopes of the Arrhenius plots (Fig. 5b). The activation energy of the bulk and grain-boundary resistances was 19 kJ mol−1, which agreed well with the results for the conductivity in NaCBH alone. The activation energy of the charge-transfer resistance at the HC/NaCBH interface was 47 kJ mol−1, which is consistent with a previously reported value.23 This value was also lower than the activation energy for the charge-transfer reaction at the HC/NASICON interface (68 kJ mol−1, Fig. S10). Thus, the composite electrode of hard carbon and NaCBH not only increased the utilization of the active material but also accelerated the ion-transfer reaction at the interface.
(a) Nyquist plots of Na/NaCBH/(NaCBH + HC) cell at 0.1 V. (b) Arrhenius plots of the bulk and grain boundary resistance and the charge transfer resistances of Na/NaCBH/(NaCBH + HC) cell.
From the above results, it was found that introducing NaCBH as an intermediate layer between the solid electrolyte and HC electrode reduced the resistances at both the solid electrolyte/NaCBH and NaCBH/HC interfaces. In addition, the high conductivity of NaCBH indicates that the HC + NaCBH composite electrode has a high utilization rate and that NaCBH is an effective intermediate layer.
The intermediate layers introduced at the NASICON–anode interface were examined. Among NaCBH, NYZC, and the PEs, NaCBH exhibited the highest conductivity. In addition, the resistance at the NASICON/soft electrolyte interface was low for NaCBH and the PE, suggesting that these electrolytes form good interfaces with NASICON. In experiments using NaCBH and HC composite electrodes, the introduction of NaCBH significantly increased the charge/discharge capacity compared with the case where NaCBH was not present. This value was close to the charge/discharge capacity of the liquid, indicating that the utilization of HC was increased by the composite formation. In addition, the low interfacial resistance of NaCBH/HC suggests that it not only improves the contact ability of the interface but also promotes ion-transfer reactions. Our findings indicate that NaCBH is an excellent intermediate layer between solid electrolytes and HC.
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.28226408.
Shota Tsujimoto: Investigation (Lead), Writing – original draft (Lead)
Ryoto Nogami: Data curation (Equal)
Kyosuke Yoshida: Data curation (Equal)
Changhee Lee: Supervision (Equal)
Yuto Miyahara: Supervision (Equal)
Takeshi Abe: Supervision (Equal)
Kohei Miyazaki: Supervision (Equal)
The authors declare no conflict of interest in the manuscript.
S. Tsujimoto, C. Lee, Y. Miyahara, T. Abe, and K. Miyazaki: ECSJ Active Members
T. Abe: ECSJ Fellow
