Low‐Cost Preparation of High‐Performance Na‐B‐H‐S Electrolyte for All‐Solid‐State Sodium‐Ion Batteries

Abstract

All‐solid‐state sodium‐ion batteries have the potential to improve safety and mitigate the cost bottlenecks of the current lithium‐ion battery system if a high‐performance electrolyte with cost advantages can be easily synthesized. In this study, a one‐step dehydrogenation‐assisted strategy to synthesize the novel thio‐borohydride (Na‐B‐H‐S) electrolyte is proposed, in which both raw material cost and preparation temperature are significantly reduced. By using sodium borohydride (NaBH₄) instead of B as a starting material, B atoms can be readily released from NaBH₄ with much less energy and thus became more available to generate thio‐borohydride. The synthesized Na‐B‐H‐S (NaBH₄/Na‐B‐S) electrolyte exhibits excellent compatibility with current cathode materials, including FeF₃ (1.0–4.5 V), Na₃V₂(PO₄)₃ (2.0–4.0 V), and S (1.2–2.8 V). This novel Na‐B‐H‐S electrolyte will take a place in mainstream electrolytes because of its advantages in preparation, cost, and compatibility with various cathode materials.

The novel thio‐borohydride (Na‐B‐H‐S) electrolyte is prepared by a one‐step dehydrogenation‐assisted strategy. By using NaBH₄ instead of B as a starting material, B atoms are released from NaBH₄ with much less energy and thus became more available to generate thio‐borohydride. The synthesized Na‐B‐H‐S (NaBH₄/Na‐B‐S) electrolyte exhibits excellent compatibility with current cathode materials, including FeF₃, Na₃V₂(PO₄)₃, and S.

1. Introduction

To assuage recent concerns about the safety of lithium and lithium resource shortage issues, sodium‐ion batteries have emerged as promising alternative candidates.^{[ 1} , ^{2 ]} To mitigate recent concerns about the safety of lithium as well as lithium resource shortage issues, all‐solid‐state sodium‐ion batteries have emerged as promising alternative candidates, continuously promoting research on fast sodium ionic electrolytes.^{[ 3} , ⁴ , ⁵ , ⁶ , ⁷ , ^{8 ]} Among these electrolytes, NaBH₄‐based electrolytes have attracted growing research attention because of their light weight, low cost, and simple device integration.^{[ 9} , ^{10 ]} Matsuo et al.^{[ 11 ]} first investigated the Na⁺ conductivity of pure NaBH₄ and showed that it had a value of 10⁻¹⁰ S cm⁻¹ at room temperature, although it was still far from achieving the available target of 10⁻⁴ S cm⁻¹.

As an inexpensive industrial reducing agent, the significant cost advantage of NaBH₄ has inspired interest in improving its Na⁺ conductivity. The most straightforward way to improve conductivity is using a second‐phase incorporation strategy, such as NaBH₄‐C₆₀,^{[ 12 ]} NaBH₄‐Al₂O₃,^{[ 13 ]} or NaBH₄‐SiO₂.^{[ 13 ]} The NaBH₄‐C₆₀ nanocomposite has shown an increase of four orders of magnitude in ionic conductivity compared with pure NaBH₄. More recently, the NaBH₄‐Al₂O₃ nanocomposite exhibited an ionic conductivity of 7.2 × 10⁻⁶ S cm⁻¹ at 80 °C. These second phases were incorporated through the ball‐milling method or the melt infiltration process; in other words, two phases were mixed using a physical approach. Ngene et al.^{[ 13 ]} revealed that this second‐phase strategy could be attributed to the surface interaction between the oxides (Al₂O₃ and SiO₂) and NaBH₄. To strengthen the two‐phase interface, the Na₂B₁₂H₁₂ phase was generated in situ through a gas‐solid chemical reaction between NaBH₄ and B₂H₆. The end product of the NaBH₄‐Na₂B₁₂H₁₂ composite achieved a sodium ionic conductivity of 10⁻⁴ S cm⁻¹ at 115 °C.^{[ 9 ]}

The phase structure and incorporation method both played key roles in tuning the Na⁺ conductivity of NaBH₄. Enlighted by the finding that an increase of three orders of magnitude in ionic conductivity could be obtained by simple substitution of O in Li_3+x(P_1‐xSi_x)O₄ with S,^{[ 14} , ¹⁵ , ^{16 ]} we speculated that introducing Na‐B‐S (e.g., NaBS₃, Na₂B₂S₅, and Na₃B₃S₆) as the second phase to fabricate thio‐borohydride composites could lead to achievements of novel properties and extend new chemistry in the NaBH₄ electrolyte. The following consequential issues, however, are related to the sodium thioborate system (Na‐B‐S): i) high energy consumption, the raw material B₂S₃ is not commercially available, high temperatures ranged from 600–850 °C is needed to ensure the reaction between boron and sulfur powder;^{[ 17} , ¹⁸ , ^{19 ]} ii) the synthesized B₂S₃ is unstable upon multi‐step reactions due to its low chemical stability;^{[ 18 ]} iii) raw material costs also need to be considered.

Herein, thioborate (Na‐B‐S) was generated in situ within NaBH₄ through chemical reactions between NaBH₄ and S, which are both low‐cost raw materials. Additionally, with the aid of the partial dehydrogenation of NaBH₄, the synthesized temperature could be significantly decreased to 240 °C. Such a strategy of incorporating second‐phase Na‐B‐S through an in situ chemical reaction undoubtedly produces the effect of killing two birds with one stone, the cheapest stone. The synthesized NaBH₄/Na‐B‐S (Na‐B‐H‐S) electrolyte exhibited a Na⁺ conductivity of 1.7 × 10⁻⁴ S cm⁻¹ at 120 °C, which was an increase of three orders of magnitude compared with pristine NaBH₄. When assembled with the S cathode and Na₁₅Sn₄ anode, the S||Na₁₅Sn₄ device achieved a capacity of 364 mAh g⁻¹ after 50 cycles. Notably, a capacity of 78 mAh g⁻¹ was obtained in the Na₃V₂(PO₄)₃||Na₁₅Sn₄ device after 50 cycles. Such low‐cost and low‐energy‐consumption preparation of the novel Na‐B‐H‐S electrolyte, along with success in assembling Na₃V₂(PO₄)₃ cathode, will shed light on further commercialization.

2. Results and Discussion

We employed TG‐DSC to determine the thermodynamic properties between NaBH₄ and S, as shown in Figure  1a. The minor endotherm peak could be found ≈113 °C without weight loss because of the polymorphic transition and subsequent melting of S.^{[ 20} , ²¹ , ^{22 ]} The reaction between NaBH₄ and S occurred rapidly at 239 °C, as represented by the sharp exotherm peak. Corresponding to this peak, the TG curve shows a significant weight loss at temperatures between 200 and 260 °C. As schematized in Figure1b, S initially underwent a melting process at 113 °C to form the mixture of NaBH₄ and molten S. Once the temperature increased to 239 °C, the molten S was reacted with partially dehydrogenated NaBH₄. Note that the pure NaBH₄ released hydrogen higher than 400 °C, and the existence of molten S might have catalyzed the partial dehydrogenation of NaBH₄ at 239 °C.^{[ 23 ]} To further estimate the feasibility of this reaction, we employed a theoretical simulation, as shown in Figure 1c. First, obtained from Atomly, the open‐access density functional theory materials data infrastructure,^{[ 24 ]} the existing stable Na‐S phases were Na₂S, NaS, NaS₂, Na₂S₅, and NaS₃, as illustrated in Figure S1 (Supporting Information). The highest and lowest Na/S ratios (Na₂S and NaS₃) were assumed to be the end product of the reaction between NaBH₄ and S, as labeled in reactions one and two, respectively. The overall free energy of reactions one and two was calculated to be −2.67 eV and −1.88 eV, respectively. These two calculated negative free energy changes demonstrated the feasibility of the reaction between NaBH₄ and S, even with a varied and uncertain Na‐S end product.

Figure 1.

Synthesis of Na‐B‐H‐S electrolytes. a) TG‐DSC curve for as‐milled NaBH₄‐S. b) The schematic view of the reaction process between NaBH₄ and S during the annealing. c) The Gibbs energy of the Gibbs energy of two kinds of formation process by DFT simulation. d) The cost of raw materials of B (95%), Na₂S (95%), S (99.5%), and NaBH₄ (98%) (excerpted from Aladdin, https://www.aladdin‐e.com/zh_cn/). e) The Na‐B‐S reaction temperature comparison in references and this work. f). Comparison of the energy (E _a, in eV) required to remove one B atom from the surfaces of crystalline B and NaBH₄.

To highlight the cost advantage of the NaBH₄+S reaction route, we compared various raw materials (B, >95%; Na₂S, >95%; S, >99.5%; NaBH₄, >98%), as shown in Figure 1d (excerpted from the website of Aladdin). In the traditional synthesis route, Na₂S provided both Na and S sources, and B powder was the B source, whereas S powder was the additional S source.^{[ 17} , ¹⁸ , ^{19 ]} In contrast, raw materials were simplified in the NaBH₄+S reaction route. The NaBH₄ provided both Na and B sources, whereas S powder was the S source, in which the cost of raw material could be sharply decreased because both NaBH₄ and S were industrial materials. For the traditional route, the ultra‐high price of Na₂S could be attributed to the time‐ and energy‐intensive process necessary to obtain anhydrous Na₂S.^{[ 25 ]} Additionally, as compared with previously reported Na‐B‐S synthesis strategies, this route achieved low energy consumption, as shown in Figure 1e. Based on the previous preparation methods, high temperatures (>600 °C) were required to activate the reaction between Na₂S, S, and B within the sealed quartz ampoule under vacuum. Our synthesized temperature could be significantly decreased to 240 °C, as derived from the DSC result (Figure 1a). This significantly decreased synthesized temperature could be rationalized as the auxiliary effect of released H from NaBH₄. As calculated in Figure 1f, B atoms were more prone to escape from the NaBH₄ surface (ΔE_B = 6.88 eV) than from the B surface (ΔE_B = 79.74 eV).

We mixed and loaded S and NaBH₄ into the reactor with mass ratios of 2:8, 4:6, 6:4, denoted as NBHS‐2, NBHS‐4, and NBHS‐6, respectively. As shown in Figure  2a, the XRD patterns of the three samples could be assigned to the NaBH₄ phase (PDF#74‐1891). With increased S loading, NBHS‐6 showed the lowest intensity of the NaBH₄ pattern, which implied that the degree of the NaBH₄+S reaction was strongly related to the initial S loading. To evaluate the detailed chemical environment of end products, especially the amorphous Na‐B‐S phase as previously reported, we employed FTIR and Raman spectrums, as compared in Figure 2b,c. According to the FTIR results, typical B‐H stretching vibration bands were located around 2200–2400 cm⁻¹, and the B‐H bending mode of ≈1120 cm⁻¹ was seen in the NBHS‐2, NBHS‐4, and NBHS‐6 samples, which further verified the residual of NaBH₄. Among these samples, NBHS‐6 showed the weakest intensity of B‐H bands, which was in accordance with the previous XRD result (Figure 2a). Figure S2 (Supporting Information) shows a detailed comparison between pure NaBH₄ and NBHS‐6. Regarding pure NaBH₄, the B‐H bending mode of NBHS‐6 shifted slightly to a lower wavenumber, which could be attributed to the incorporation of S into the residual NaBH₄ phase.^{[ 22 ]} New peaks ranging from 800 and 1000 cm⁻¹ could be assigned to the B‐S bonds, in which the intensity gradually increased with the S amount. Similar observations could be made in the Raman analysis (Figure 2c). Three fingerprint peaks of NaBH₄ located at 2330, 2198, and 1278 cm⁻¹ gradually decreased from NBHS‐2 to NBHS‐6, which were accompanied by the appearance of new peaks of 2551 cm⁻¹, which could be assigned to the S─H bond.^{[ 26 ]} Moreover, for the NBHS‐4 and NBHS‐6 samples, two obvious new peaks of 295 and 609 cm⁻¹ could be indexed to S₆ ⁻ and S₇ ⁻, respectively, which implied the existence of sodium polysulfide (NaS_x) in Na‐B‐H‐S.^{[ 27 ]}

Figure 2.

Structure of Na‐B‐H‐S electrolytes. a) XRD patterns, b) FTIR spectra, and c) Raman spectra of the pure NaBH₄ and NBHS‐x (x = 2, 4, and 6). d) ¹¹B NMR, e) ²³Na NMR, and f) ¹H NMR spectra of NBHS‐x (x = 2, 4, and 6). g) B 1s, h) Na 1s, and i) S 2p XPS spectra of NBHS‐n (n = 2, 4, and 6).

We further evaluated the local chemical environment of Na‐B‐H‐S with NMR, as depicted in Figure 2d‐f. For ¹¹B NMR spectra (Figure 2d), a resonance peak ≈−41 ppm could be assigned to BH₄, which was clearly observed in all three samples. Because of their weak intensity, the signals ranging from −10 to 60 ppm were amplified by a factor of 25. All Na‐B‐H‐S samples showed a resonance peak ≈−2.5 ppm, representing B in the BS₄ tetrahedra with a bridging S environment.^{[ 28 ]} Among these samples, an additional peak split at 0.49 ppm was observed only in NBHS‐6, which was assigned to the isolated BS₄ tetrahedra.^{[ 28 ]} We identified the other two peaks, which were centered at 53 and 36 ppm, as BS₃. We observed these peaks in all of the Na‐B‐H‐S samples, but the highest intensity was in NBHS‐6. Another peak centered at 24 ppm in the NBHS‐6 sample was assigned to BO_xS_4‐x, which may have been introduced during sample preparation.^{[ 28 ]} Concerning the ²³Na NMR spectra (Figure 2e), a peak ≈−7 ppm verified the existence of NaBH₄, while a peak at 0.6 ppm was evident only in the NBHS‐6 sample. This result indicated that S preferentially reacted with BH₄ before affecting the local environment of Na. According to the ¹H NMR spectra (Figure 2f), an obvious bump ≈0 ppm could be assigned to the chemical environment change of BH₄. To quantitatively compare this variation, we used the full width at half maxima (FWHM) to estimate the activity of H, in which the smaller the FWHM value was, the more active the H was.^{[ 29 ]} The NBHS‐6 sample exhibited the smallest value, which demonstrated that more active H existed in NBHS‐6. This result was attributed to the intensified H─S bonds in NBHS‐6 because more active H could be found in the S─H bond as compared with the B─H bond.

The XPS analysis provided more detailed information, as depicted in Figure g‐i. For the B 1s spectra (Figure 2g), the peak at 187.7 eV was assigned to the B─H bond, in which the intensity decreased as the S mount increased. Two other peaks at 189 and 192 eV were assigned to the isolated BS₄ and BS₄ with bridging S, respectively. We noted more obvious thioborates in NBSH‐6, which suggested the possible significant structural differences in NBSH‐6 compared with its counterparts. As displayed in Figure 2h, the Na 1s peaks slightly shifted from 1071.6 to 1071.2 eV with gradually increased S loading, which suggested that the Na⁺ migration became more flexible in NBHS‐6. Concerning S 2p (Figure 2i), the doublet peaks at 163.8 eV (S2p_1/2) and 162.8 eV (S2p_3/2) could be assigned to BS₄ with bridging S, while the doublet peaks at 162.6 eV (S2p_1/2) and 161.5 eV (S2p_3/2) could be attributed to terminal S (NaS_x).^{[ 30} , ³¹ , ^{32 ]} With increased S loading, especially for NBHS‐6, the terminal S (NaS_x) significantly increased, which was in accordance with the previous Raman analysis. The existence of NaS_x in NBSH‐6 was supported by the result that a capacity of 100 mAh g⁻¹ could be obtained by mixing NBSH‐6 with carbon as the electrode in KB/NBSH‐6||NBSH‐6||Na₁₅Sn₄ (Figure S3, Supporting Information). The morphology of NBHS‐6 is shown in Figures S4 and S5 (Supporting Information). The size of a single particle was at the micrometer scale. Further Energy Dispersive Spectroscopy (EDS) mapping revealed that the Na and S elements were evenly distributed.^{[ 33 ]}

We also found that all of the Na‐B‐H‐S samples included a NaBH₄ crystal structure and amorphous thioborate (Na‐B‐S). With increased S loading, especially for NBHS‐6, NaBH₄ and thioborate were tightly bonded to each other with newly formed S─H bonds. Compared with the NBHS‐2 and NBHS‐4 samples, in addition to the higher intensity, NBHS‐6 exhibited a slightly different thioborate structure, with more isolated BS₄ and NaS_x. We theoretically evaluated the thermodynamic stability of NaBH₄‐NaS_x and thioborate‐NaS_x‐NaBH₄, as shown in Figures S6 and S7 (Supporting Information), respectively. The possibility of spontaneous reactions at various ratios between thioborate‐NaS_x‐NaBH₄ interfaces was extremely low, ensuring the stability of the Na‐B‐H‐S electrolyte.

The stability with water or oxygen has been evaluated, as shown in Figure S8 (Supporting Information). The freshly synthesized NBHS‐6 electrolyte was subjected to ambient conditions within a fume hood with a prolonged time. During the initial 1 h, the NBHS‐6 electrolyte exhibited a color transition from off‐white to red (Figure S8a,b, Supporting Information). Additionally, the FTIR (Figure S8c, Supporting Information) comparison revealed that the B─S bonds ranging from 800 and 1000 cm⁻¹ disappeared, which could be assigned to the reaction of the NBHS‐6 with water and oxygen. Prolonging the exposure time to 24 h, the NBHS‐6 was deliquescent, and could no longer used as a solid electrolyte (Figure S8d, Supporting Information). This unstable nature with water or oxygen has been reported in most of the hydride‐ and sulfide‐based electrolyte.^{[ 34} , ^{35 ]}

We evaluated the electrochemical performance of the Na‐B‐H‐S samples, as shown in Figure  3 . The ionic conductivity of the Na‐B‐H‐S and NaBH₄ samples were determined using the EIS, which ranged from 30 to 150 °C, as illustrated in Figure 3a. Compared with pristine NaBH₄, all of the Na‐B‐H‐S samples exhibited enhanced Na⁺ conductivity. At 90 °C, this value could be increased by three orders of magnitude, and at 120 °C, this value could be increased by at least two orders of magnitude. Among the samples, the NBHS‐6 sample manifested the optimal Na⁺ conductivity, by virtue of more thioborates (BS₄ and BS₃) as well as more flexible Na⁺ migration, which was evidenced by its structure characterization. The bulk resistance (R _bulk) and grain boundary resistance (R _grain) were analyzed with EIS, as shown in Figures S9 and S10 (Supporting Information), one semicircle in pristine splits into two semicircles. The higher frequency semicircle represents R _bulk, while the lower frequency semicircle is related to R _grain. The R _bulk and R _grain of the NBHS‐6 were reduced to 296 and 373 Ω cm², respectively, at 120 °C. This is attributed to the diminishing interface between NaBH₄ and amorphous Na‐B‐S within it, as well as the lower impedance in the amorphous Na‐B‐S. In addition, the presence of the amorphous Na‐B‐S phase contributes to good interfacial contact and reduces interfacial impedance.^{[ 36 ]} A further increase in the S amount and annealing temperature (260 °C) failed to continuously enhance Na⁺ conductivity, as illustrated in Figures S11 and S12 (Supporting Information), respectively. The electronic conductivity of NBHS‐6 was evaluated to 5.53 × 10⁻⁹ S cm⁻¹, which met the basic demand for a solid‐state electrolyte (Figure 3b). We compared the activation energy of Na‐B‐H‐S samples with different NaBH₄/S mass ratios, as shown in Figure 3c. Interestingly, the activation energy of NaBH₄ initially decreased to 0.40 in NBHS‐2 and remained relatively stable even within high S loading in the NBHS‐6 sample. With the introduction of thioborates (BS₄ and BS₃), the activation barriers could be lowered considerably because of the larger size of thioborates (BS₄ and BS₃) than the BH₄ unit, as shown in Figure 3d. Yu et al.^{[ 37 ]} indicated that the introduction of a second phase with a large anionic radius could bring significant changes to the local environment of alkali ion in pristine metal hydrides, further facilitating fast ion diffusion. The introduction of a second phase with a large anionic radius is often accompanied by an increase in lattice volume, which could augment the rotational freedom of the anions.^{[ 38 ]} The size of BS₄/BS₃ was larger than the BH₄ unit as shown in Figure 3d. The larger BS₄/BS₃ anion embedded in the anionic frameworks may contribute to the reconstruction of wider Na⁺ diffusion channels, which result in reduced electrostatic interactions between the Na⁺ and the anionic skeleton, thus facilitating faster Na⁺ mobility. A rational explanation was provided for NBHS‐6, in which although the ionic conductivity increased, the activation energy was relatively stable without obvious change (Figure 3e). The Na⁺ conductivity of initial NaBH₄ was rather low, which was like Na⁺ walking on a country road with limited speed, which was equivalent to active energy barrier (E _a). In the case of NBHS‐2 and NBHS‐6, the interaction between BH₄ anions was reduced while also broadening the transportation space for Na⁺, like the construction team (BS₄ and BS₃) widened the original country road and turned it into a highway, further dramatically increasing the moving speed. With the further accumulation of BS₄/BS₃ in NBHS‐6, the speed limit of the highway, which was similar to E _a, no longer changed; however, BS₄/BS₃ turned from single‐lane to multi‐lanes to ensure greater Na⁺ transportation, as reflected by the enhanced Na⁺ conductivity.

Figure 3.

Electrochemical properties of the Na‐B‐H‐S electrolytes. a) Conductivity–temperature curves of pure NaBH₄ and NBHS‐x (x = 2, 4, and 6) in the temperature range of 30–150 °C. b) Time‐dependent current density curves during polarization at 1 V c) The ionic conductivity at 90 and 120 °C, and the activation energies of NBHS‐x as a function of x. d) Illustrations of the geometric shapes on an identical scale of BH₄, BS₃, and BS₄ anions, BH₄ and BS₄ anions are shown in top and side views and BS₃ is only shown in top view. Green balls: B atoms; pink balls: H atoms; Yellow balls: S atoms. (e) Schematic illustration of the Li‐ion diffusion with S content increasing. Galvanostatic cycling of the Na₁₅Sn₄lNBHS‐6lNa₁₅Sn₄ symmetric cells under (f) constant current density and (g) step‐increased current densities at 125°C.

The synthesized NBHS‐6 electrolyte exhibited a Na⁺ conductivity of 1.7 × 10⁻⁴ S cm⁻¹ at 120 °C. Given the low melting point of Na metal (97 °C), the Na₁₅Sn₄ alloy was employed to assemble devices (Figure S13, Supporting Information). Figure 3f shows the constant current cycling of the NBHS‐6 electrolyte in a symmetrical cell with Na₁₅Sn₄ electrodes at 125 °C. It is demonstrated that NBHS‐6 electrolyte can reversibly transfer Na⁺ for 250 h. Detailed information has been magnified and is shown in the inset in Figure 3f, in which the overpotential for the initial 20 h was 46 mV, and a value of 52 mV was obtained after 200 h, which demonstrated the stability and compatibility between NBHS‐6 electrolyte and Na₁₅Sn₄ anode. The smaller overpotential variation indicated that no significant side reactions occurred with the critical current density of 1.25 mA cm⁻² (Figure 3g).

We conducted a CV test on S||Na₁₅Sn₄ device within the voltage range of 0.1–5 V at a scanning rate of 0.5 mV s⁻¹. In the CV curve (Figure S14, Supporting Information), a broad oxidation peak is observed at 1.8 V, accompanied by a corresponding reduction peak positioned at 1 V. These peaks are attributed to the reversible conversion between S and Na₂S, respectively. No other oxidation peak was observed, indicating the stability of the NBHS‐6 electrolyte up to 5 V.

To evaluate the universality of the NBHS‐6 electrolyte, we used various cathode materials, including FeF₃ (1.0–4.5 V), Na₃V₂(PO₄)₃ (2.0–4.0 V), and S (1.2–2.8 V), as shown in Figure  4 . The FeF₃||Na₁₅Sn₄ device exhibited a reversible capacity of 290 mAh g⁻¹ at 50 mA g⁻¹ after 15 cycles (Figure 4a). It should be noted here the cycling stability of FeF₃ cathode is still unfavorable, due to multiphase redox steps involved in metal fluorides, with extensive structural rearrangements.^{[ 39} , ^{40 ]} Figure 4b shows the corresponding galvanostatic charge–discharge (GDC) curves for the different cycles. Figure 4c shows the differential capacity versus voltage (dQ/dV) curves within the voltage window of 1.0–4.5 V. The dQ/dV curves revealed that the oxidation/reduction peaks were slightly lower than the theoretical reaction potential between FeF₃ and Na⁺, which further confirmed that the NBHS‐6 electrolyte participated in the reaction. The oxidation peaks located between 2.5 and 4.5 V and the reduction peak at 2 V were consistent with Na⁺ insertion/extraction reaction. The oxidation peaks located at 2 and 1.7 V and the reduction peaks at 1.6 and 1.2 V, however, were consistent with the conversion reaction between NaFeF₃ and sodium fluoride (NaF).^{[ 41 ]}

Figure 4.

Performance of All‐solid‐state sodium‐ion batteries with NBHS60 at 125 °C. a) Cycling performance of FeF₃||Na₁₅Sn₄ cell at 0.05 A g⁻¹ and the b) corresponding GDC curves. c) Differential capacity curves (dQ/dV) of the FeF₃ electrode were obtained from the galvanostatic discharge–charge curves. d) Cycling performance of Na₃V₂(PO₄)₃||Na₁₅Sn₄ cell at 0.1 C and the e) corresponding GDC curves. f) Differential capacity curves (dQ/dV) of the Na₃V₂(PO₄)₃ electrode obtained from the galvanostatic discharge‐charge curves. g) Cycling performance of S||Na₁₅Sn₄ cell at 0. 1 A g⁻¹ and the h) corresponding GDC curves. i) Differential capacity curves (dQ/dV) of the S electrode obtained from the galvanostatic discharge–charge curves.

To evaluate the compatibility of NBHS‐6 electrolyte for commercial electrode, we prepared the Na₃V₂(PO₄)₃||Na₁₅Sn₄ device. As illustrated in Figure 4d, the Na₃V₂(PO₄)₃||Na₁₅Sn₄ showed an exciting capacity of 78 mAh g⁻¹ at a current density of 0.1 C after 50 cycles. The corresponding GDC curves (Figure 4e) demonstrated that a large deviation for the first discharge process could generate a stable interface between the NBHS‐6 electrolyte and Na₃V₂(PO₄)₃ cathode. From the second cycle onward, the GDC curves overlapped, which indicated the reversibility of the Na₃V₂(PO₄)₃||Na₁₅Sn₄ device. Figure 4f shows the dQ/dV curves within the voltage window of 2.0–4.0 V. The oxidation peak located at 3.3 and the reduction peak at 3.2 V were consistent with Na₃V₂(PO₄)₃||Na₁₅Sn₄.^{[ 42 ]} The oxidation peak located at 2.8 and the reduction peak at 2.1 V were consistent with Na₂S_x in NBHS‐6.

When assembled with the S cathode and Na₁₅Sn₄ anode, the S||Na₁₅Sn₄ device could achieve a capacity of 364 mAh g⁻¹ after 50 cycles, as shown in Figure 4g. Apart from the first discharge curve, the GDC curves of the S||Na₁₅Sn₄ device overlapped from the second cycle onward. Figure 4i shows the dQ/dV curves within the voltage window of 1.0–2.6 V. The dQ/dV curves revealed two oxidation peaks located at 1.6 and 1.8 V and two reduction peaks at 1.3 and 1.7 V. The oxidation peak located at 1.6 V and the reduction peak at 1.3 V was consistent with the reversible conversion between Na₂S₄ and Na₂S, whereas the oxidation peak located at 1.8 V and the reduction peak at 1.7 V were consistent with the reversible conversion between S and Na₂S₄. We further investigated the rate performance of the S||Na₁₅Sn₄ at different current densities (Figure S15, Supporting Information). At current densities of 0.05 0.1, 0.2, 0.4, and 0.8 A g⁻¹, the average discharge capacities of the S||Na₁₅Sn₄ electrode were 1003, 507, 381, 272, and 152 mAh g⁻¹, respectively. The capacity could recover to 331 mAh g⁻¹ when the current density returned to 0.2 A g⁻¹. The corresponding GDC curves at various current densities are illustrated in Figure S16 (Supporting Information). Even at a high current density of 0.8 g⁻¹, the voltage plateau was well preserved. Once assembled with the adapted processes, these three typical cathode materials, Na‐free cathode (S and FeF₃), high‐voltage (Na₃V₂(PO₄)₃ and FeF₃), and high‐capacity (FeF₃ and S) materials, were all compatible with the NBHS‐6 electrolyte.

3. Conclusion

In conclusion, we proposed a one‐step dehydrogenation‐assisted strategy to synthesize the novel thio‐borohydride Na‐B‐H‐S (NaBH₄/Na‐B‐S) electrolyte. By using NaBH₄ instead of B as a starting material, both raw material cost and preparation temperature were significantly reduced. Theoretical modeling demonstrated that B atoms were readily released from NaBH₄ with much less energy of 6.88 eV than 79.74 eV of elemental B. The synthesized Na‐B‐H‐S (NaBH₄/Na‐B‐S) electrolyte could exhibit a Na⁺ conductivity of 1.7 × 10⁻⁴ S cm⁻¹ at 120 °C, which was an increase of three orders of magnitude compared with pristine NaBH₄, by virtue of larger thioborates (BS₄ and BS₃) as well as more flexible Na⁺ migration. When assembled with the S cathode and Na₁₅Sn₄ anode, the S||Na₁₅Sn₄ device could manifest a capacity of 364 mAh g⁻¹ after 50 cycles. Excitingly, the Na₃V₂(PO₄)₃||Na₁₅Sn₄ and FeF₃||Na₁₅Sn₄ devices achieved a promising performance that demonstrated the potential for further commercialization.

4. Experimental Section

Material Synthesis

The raw materials applied for the preparation of Na‐B‐H‐S solid electrolytes (SEs) were NaBH₄ (98% purity, Aladdin, Shanghai, China) and S (99.5% purity, Sinopharm, Shanghai, China). For the synthesis of NBHS SEs, NaBH₄ and S is added to argon (Ar)‐filled agate tank and then ball‐milled the mixture at 400 rpm for 4 h in the mass ratio of 8:2, 6:4, and 4:6. The ratio of grinding media to powder was 10:1. It then placed the mixture in a glass bottle, heated it to 240 °C at the heating rate of 1 °C min⁻¹ with a muffle furnace (KSL‐1100X‐S, Hefei Kejing, Anhui, China) under an Ar atmosphere, and held for 1 h. After cooling to room temperature, the SEs were ground using mortar. Na₁₅Sn₄ alloy was synthesized by heating the mixture of tin powder and sodium foil in a stoichiometric ratio at 150 °C with mechanical stirring at 200 rpm for three days under an argon atmosphere. All synthesis procedures were conducted inside an Ar‐filled glovebox (O₂ < 0.01 ppm; H₂O < 0.01 ppm).

Characterization Methods

It performed differential scanning calorimetry (DSC) tests using a TGA‐DSC, Netzsch, STA 409 PC (Selb, Germany). It conducted X‐ray diffraction (XRD) analysis on a Bruker D8 Advance X‐ray diffractometer (Karlsruhe, Germany) with Cu Kα radiation (λ = 1.5406 Å). It used a field‐emission scanning electron microscope (TESCAN VEGA 3 XMU, Brno, Czech Republic) equipped with energy‐dispersive X‐ray photoelectron spectroscopy (EDS) to analyze the sample morphology. It recorded the Fourier‐transformed infrared (FTIR) spectra of the SEs on an infrared spectrometer (NICOLET iS50 FT‐IR, ThermoFisher Scientific, Waltham, MA, USA). It recorded the Raman spectra of the SEs on an Integrated FTIR‐Raman spectrometer (Bruker). It performed the ²³Na, ¹¹B, and ¹H nuclear magnetic resonance (NMR) tests on a Bruker NEO 400 m solid‐state NMR spectrometer and conducted S 2p, Na 1s, and B 1s XPS spectra on an ESCALAB 250Xi spectrometer (ThermoFisher Scientific). All samples were prepared in a glove box.

Density Functional Theory Calculations

Theoretical calculations based on density functional theory (DFT) were performed using a spin‐polarized generalized‐gradient approximation with a PW91 functional and a double numerical basis set, including a polarization function as implemented in the DMol3 package. The Brillouin zone was sampled by 2 × 2 × 1 special k‐points using the Monkhorst‐Pack scheme. The (01−1) surfaces of pure B and (110) surface of NaBH₄ crystals were represented by 3‐D slab models in a periodic boundary condition. The slab models were constructed initially as 2 × 3 × 1 super cells for B and 3 × 3 × 4 super cells for NaBH₄, respectively. The exposed surface was given a vacuum layer of >10 Å to avoid interactions with its periodic image in the c direction. The energy (ΔE _B) required to remove one B atom from the surfaces is defined as

$${\mathrm{\Delta }\mathrm{E}}_{\mathrm{B}}={\mathrm{E}}_{\mathrm{total}}({\mathrm{N}}_{\mathrm{B}}-1)+{\mathrm{E}}_{\mathrm{B}}-{\mathrm{E}}_{\mathrm{total}}\left({\mathrm{N}}_{\mathrm{B}}\right)$$ (1)

where E _total (N _B−1) is the total energy of the surface model with the (N _B−1) number of B atoms, E _B is the energy of one isolated B atom, and E _total (N _B) is the total energy of the surface model with the N _B number of B atoms.

Electrochemical Measurements

It determined the bulk resistance of the samples from electrochemical impedance spectroscopy (EIS) using a High‐Frequency Impedance Test System (HT‐Z2‐HF, TOYO Corporation, Tokyo, Japan). It completed the measurements at a temperature ranging from 30 to 150 °C. The powder was pelletized in a 10 mm diameter at room temperature under the pressure of 200 MPa.

It measured electronic conductivity via DC polarization using a 13 mm diameter Polyetheretherketone (PEEK) mold, in which powder was pressed between two stainless steel pistons. It applied a voltage of 1 V for 5 h on an electrochemical workstation (Gamry INTERFACE1010E, Warminster, PA, USA).

It pressed NBHS‐6 SE (100 mg) in a 13 mm diameter PEEK mold to form a pellet. The powder Na_xSn (20 mg) was uniformly spread on either side and kept in contact with the SE. The cell was then sandwiched between two stainless steel rods and held under pressure using a custom‐made cell. It performed galvanostatic cycling of the NBHS‐6 cells at 0.05 mA cm⁻². It tested step‐increased current densities from 0.05 to 1.35 mA cm⁻².

For the fabrication of Na₁₅Sn₄|NBHS‐6|Na₁₅Sn₄ symmetric cells, ≈100 mg NBHS‐6 powder was pressed into a 13 mm pellet at 200 MPa for 1 min, and then 10 mg Na₁₅Sn₄ powder was added to both sides of the pellet and pressed at 200 MPa for another 1 min. The S–graphene oxide (GO) composite was prepared by mixing the S with GO (sheet diameter: >5 um, number of layers: 1–6 layers, XFNANO) at a weight ratio of 3:7. The Na₃V₂(PO₄)₃/NBHS‐6/graphene composite was prepared by mixing the Na₃V₂(PO₄)₃ (20 µm, Shenzhen Kejing), NBHS‐6 with graphene (sheet diameter: 0.5–5 µm, thickness: 0.8 nm, monolayer rate: 80%, XFNANO) at a weight ratio of 4:3:3. The FeF₃/NBHS‐6/ Ketjen black (KB) composite was prepared by mixing the FeF₃, NBHS‐6 with KB at a weight ratio of 7:3:10. First, 100 mg NBHS‐6 powder was placed into a PEEK cylinder and pressed at 200 MPa for 1 min (13 mm diameter). On one side of the NBHS‐6 pellet, the cathode composite was spread. Finally, 20 mg Na₁₅Sn₄ powder was spread over the other side of the NBHS‐6 pellet. The S||Na₁₅Sn₄ cell was cycled at a current density of 100 mA g⁻¹ and voltage range between 1.0–2.8 V versus Na₁₅Sn₄. The FeF₃||Na₁₅Sn₄ cell was cycled at a current density of 50 mA g⁻¹ and voltage range between 1.0–4.5 V versus Na₁₅Sn₄. The Na₃V₂(PO₄)₃||Na₁₅Sn₄ cell was cycled at a current density of 11.8 mA g⁻¹ and voltage range between 2.0–4.0 V versus Na₁₅Sn₄. All cells were maintained at 125 °C for 24 h and then used for electrochemical tests. It assembled all cells in an Ar‐filled glove box with O₂ and H₂O < 0.01 ppm.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Zhou W., Song C., Li S., Liu M., He H., Yang S., Xie J., Wang F., Fang F., Sun D., Zhao J., Song Y., Low‐Cost Preparation of High‐Performance Na‐B‐H‐S Electrolyte for All‐Solid‐State Sodium‐Ion Batteries. Adv. Sci. 2023, 10, 2302618. 10.1002/advs.202302618

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.