Li_xVS_y nanocomposite electrodes for high-energy carbon-additive-free all-solid-state lithium-sulfur batteries

Preparation of Li_xVS_y cathodes

Li_xVS_y (x = 5-9, y = 4-6) cathodes were prepared by the mechanochemical treatment of Li₂S (99%, Kojundo Chemical Lab. Co. Ltd.) and V₂S₃ (99%; Kojundo Chemical Lab. Co. Ltd.) according to a previously described method^[20]. A high-crystallinity sample of Li₈VS_5.5 (denoted as Li₈VS_5.5-HC) was synthesized under modified processing conditions. Stoichiometric mixtures of Li₂S and V₂S₃ were weighed - 6 g for Li₈VS_5.5-HC and 1 g for standard Li₈VS_5.5 - and placed into ZrO₂ milling pots (250 mL for Li₈VS_5.5-HC and 45 mL for Li₈VS_5.5) each loaded with ZrO₂ balls (30 balls of 15 mm diameter for Li₈VS_5.5-HC and 15 balls of 10 mm diameter for Li₈VS_5.5). All procedures were conducted under a dry Ar atmosphere. Mechanical milling was performed using planetary ball mills: a Pulverisette 5 classic line (Fritsch) for Li₈VS_5.5-HC and a Pulverisette 7 premium line (Fritsch) for Li₈VS_5.5. Both mixtures were milled for 160 h at rotational speeds of 300 rpm for Li₈VS_5.5-HC and 400 rpm for Li₈VS_5.5.

Construction of ASSBs

Composite cathodes were prepared by gently mixing Li_xVS_y and argyrodite-type sulfide SEs (Mitsui Mining & Smelting Co., Ltd.) for 1 h using a planetary ball mill (Pulverisette 7 premium line, Fritsch). To assemble ASS half-cells with a diameter of 10 mm, the sulfide SE was pressed at 30 MPa to fabricate an SE layer, and the composite cathode (10 mg) was placed on the SE layer and pressed at 720 MPa; subsequently, the Li-In alloy was attached to the other side of the SE layer. The resultant cell was sandwiched between two stainless steel (SUS) rods used as current collectors and confined at 160 MPa. Each 1-centimeter square ASS full cell was constructed using a Li₈VS_5.5 composite cathode and a composite Si anode; the loading mass of the positive and anode layers was 33 and 10.4 mg cm^-2, respectively. The fabricated cells were charged and discharged using a charge-discharge measuring device (TOSCAT-3100, Toyo System Co. Ltd.).

Electronic and ionic conductivity measurements

This study measured the electronic and ionic conductivity of composite cathodes by direct current (DC) polarization tests at 25, 45, and 60 °C. SUS/composite cathode/SUS and SUS/Li-In/SE/composite cathode/SE/Li-In/SUS systems were used for electronic and ionic conductivity measurements, respectively. The loading mass of the composite cathode layer in each system was > 30 mg cm^-2.

Synchrotron XRD measurements

Synchrotron XRD was used to analyze pristine Li_xVS_y and composite cathodes in charge and discharge states at the BL19B2 beamline of SPring-8 at room temperature. The samples were sealed in glass capillaries (~0.3 mm diameter) under dry Ar. The wavelength of the synchrotron X-ray was 0.5002 Å calibrated by CeO₂. The RIETAN-FP program was used for Rietveld refinement^[26], and multiphases were added one by one while peak identification was carefully performed to sufficiently reduce residual peaks.

Elemental analysis of Li_xVS_y

To investigate the amounts of Li and V in Li_xVS_y, inductively coupled plasma atomic emission spectroscopy (ICP-AES) was conducted using an inductively coupled plasma atomic emission spectrometer (Agilent 5110 VDV, Agilent Technologies, Inc.). Before measurements, the samples were thermally decomposed using acids, and their final volume was adjusted using ultrapure water. The amount of S in the samples was measured using an automatic combustion halogen and sulfur analysis system (HSU-35/SQ-10, Anatec Yanaco Co. Ltd. and ICS-2100, Dionex Corp.).

XAFS measurements

The valence states and local structures of the S and V atoms in Li_xVS_y and composite cathodes after charge and discharge tests were examined by S and V K-edge XAFS measurements at the Synchrotron Radiation Center, Ritsumeikan University. The S and V K-edge spectra of the samples were recorded at the BL-10 and BL-3 beamlines, respectively^[27]. The incident X-ray beam was monochromatized with a Ge(111) crystal (2d = 6.532 Å) for S K-edge measurements by the total electron yield method. The photon energy was calibrated with the strong resonance of K₂SO₄ (S 1s → t₂) at 2481.7 eV^[28]. For V K-edge measurements in the conventional transmission mode, the incident X-ray beam was monochromatized with a Si(220) crystal (2d = 3.840 Å). All samples for S and Fe K-edge measurements were sealed in Ar-filled transfer vessels^[27] and Al-laminated bags, respectively.

TEM observations

The Li_xVS_y samples were analyzed by a high-resolution transmission electron microscope (Talos F200X, Thermo Fisher Scientific, Inc.). Bright-field TEM micrographs were acquired by a slow-scan charge-coupled device camera (Ceta, FEI Company) at 200 kV of accelerated voltage with 0.5 s exposure time.

Additionally, the composite cathodes before charging and after the second discharging process were analyzed using cryo-scanning TEM (cryo-STEM) on an atomic-resolution analytical electron microscope (JEM-ARM200F, JEOL Ltd.). Before STEM analysis, the samples were thinned to the desired shape using a focused ion beam milling apparatus (NanoDUE’ T NB5000, Hitachi High-Tech Corp.). To investigate the elemental mappings in the samples, STEM-energy dispersive X-ray spectroscopy (STEM-EDX) and STEM-electron energy loss spectroscopy (STEM-EELS) were conducted with EDX (JED-2300T, JEOL Ltd.) and EELS (GIF Quantum-ER, Gatan, Inc.) instruments.

Characterization of Li_xVS_y cathodes

Cathode materials with compositions Li₅VS₄, Li₆VS_4.5, Li₇VS₅, Li₈VS_5.5, and Li₉VS₆ were synthesized via ball milling of Li₂S and V₂S₃. As reported in our previous study^[20], the theoretical capacity of Li_xVS_y compounds increases with higher Li and S content (increasing x and y), whereas the electronic conductivity correspondingly decreases. The initial charge-discharge profiles of half-cells incorporating Li_xVS_y (x = 5-9, y = 4-6) cathodes were also investigated in that study^[20]. Among the compositions evaluated, Li₈VS_5.5 exhibited the highest initial discharge capacity, making it the most promising candidate. Therefore, Li₈VS_5.5 was selected for further investigation in the present study. Figure 1A and Supplementary Table 1 show the synchrotron XRD patterns and crystallographic data, respectively, of Li₈VS_5.5. The results of the other Li_xVS_y samples are shown in Supplementary Figure 1 and Supplementary Table 1. The synchrotron XRD patterns of all samples contain diffraction peaks corresponding to Li₂S (space group: Fm$$\overline{3}$$m) and LiVS₂ (space group: P$$\overline{3}$$m1), indicating the formation of LiVS₂ by the reaction of Li₂S and V₂S₃. The molar ratio of all samples from Li₂S to LiVS₂ was calculated from their synchrotron XRD patterns [Supplementary Table 1]. Except for Li₅VS₄, the molar and nominal ratios of all samples are similar. The spectrum of Li₅VS₄ contains diffraction peaks corresponding to V₂O₃; notably, the starting material V₂S₃ contains V₂O₃ as an impurity. Li₅VS₄, which requires the highest amount of V₂S₃ among all the Li_xVS_y samples, contained the highest amount of V₂O₃. For Li₅VS₄, the addition of vanadium atoms from V₂O₃ to the mol% of LiVS₂ results in Li₂S/LiVS₂ (V₂O₃) = 69.1/30.5, which is approximately the same as the nominal ratio of the system. The amounts of Li, V, and S in Li_xVS_y were determined using ICP-AES and an automatic combustion halogen and sulfur analysis system [Supplementary Table 2]. The measurement data for all samples were consistent with their nominal ratios, indicating that S volatilization did not occur during synthesis by ball milling.

Figure 1. (A) Rietveld analysis of the synchrotron XRD pattern of Li₈VS_5.5 with R_wp = 3.13%, R_e = 1.64%, and S = 1.91. (B) S K-edge and (C) V K-edge XANES spectra of Li₅VS₄, Li₆VS_4.5, Li₇VS₅, Li₈VS_5.5, and Li₉VS₆ with reference spectra of Li₂S, LiVS₂, and V₂S₃. The V K-edge XANES spectrum of Li₉VS₆ is not shown because of its low S/N ratio.

Figure 1B and C show the S and V K-edge X-ray absorption near-edge structure (XANES) spectra of the Li_xVS_y samples. The V K-edge spectrum of Li₉VS₆ is not shown in Figure 1C because of its low S/N ratio. In the S K-edge spectra, the peak intensities of Li₂S and LiVS₂ increase and decrease, respectively, with increasing x in Li_xVS_y, consistent with the results of synchrotron XRD. Moreover, the V K-edge spectra indicate that the electronic states of V in all the samples are almost the same as those in LiVS₂.

TEM was used to investigate the microstructures of Li₅VS₄, Li₆VS_4.5, Li₇VS₅, and Li₈VS_5.5 [Supplementary Figure 2]. The results indicate that most samples are in an amorphous state. In the TEM image of Li₈VS_5.5, LiVS₂ and Li₂S are mapped in red and green at each lattice spacing [Figure 2A]. The TEM image of Li₈VS_5.5 shows nanocomposites of Li₂S and LiVS₂ distributed in Li₈VS_5.5. Li_xVS_y cathodes show high electronic conductivity (~10^-1-10^-2 S cm^-1) because they comprise highly composited LiVS₂ with high electronic conductivity and insulating Li₂S. To compare the mixed states of Li₂S and LiVS₂, a Li₈VS_5.5 sample with high crystallinity (labeled Li₈VS_5.5-HC) was synthesized under different ball-milling conditions with relatively low mechanical energy. The TEM image of Li₈VS_5.5-HC [Figure 2B] indicates a higher crystallinity of Li₂S and LiVS₂ than that in Li₈VS_5.5. A comparison of Figure 2A and B indicates that Li₂S and LiVS₂ with a large crystallite size exist in Li₈VS_5.5-HC; in contrast, Li₂S and LiVS₂ with a smaller crystallite size are highly mixed in Li₈VS_5.5. Moreover, Li₈VS_5.5-HC exhibits a higher electronic conductivity (0.19 S cm^-1) than Li₈VS_5.5 (0.059 S cm^-1), possibly because Li₈VS_5.5-HC contains high-crystallinity LiVS₂ with high electronic conductivity. Figure 2C shows the initial charge-discharge curves of ASSBs containing Li₈VS_5.5 samples. Cells with Li₈VS_5.5-HC show large hysteresis and low capacity. Moreover, the mixture of Li₂S and LiVS₂ prepared by hand mixing did not operate in ASSBs. A comparison of Li₈VS_5.5 samples suggests that high electronic conductivity and a well-dispersed nanocomposite structure comprising Li₂S, which contributes to high capacity, and LiVS₂, which provides high electronic conductivity, is essential for optimal performance. As demonstrated in our previous study^[20], Li₈VS_5.5 exhibited the highest initial charge capacity among the Li_xVS_y series (x = 5-9, y = 4-6), indicating that its specific molar ratio of Li₂S to LiVS₂ is particularly favorable. Additionally, the presence of amorphous regions within the Li_xVS_y matrix is believed to facilitate efficient pathways for both electronic and ionic transport^[24].

Figure 2. High-resolution TEM images of (A) Li₈VS_5.5 and (B) Li₈VS_5.5-HC (high crystallinity) with LiVS₂ and Li₂S mapped in red and green, respectively. The color mapping is defined by a difference in lattice spacing. (C) Initial charge-discharge curves of ASSBs containing Li₈VS_5.5 and Li₈VS_5.5-HC under a current density of 0.13 mA cm^-2 at 25 °C.

Electrochemical properties of Li₈VS_5.5 cathodes

To determine an optimal composition for electrode fabrication, composite cathodes were prepared by mixing Li₈VS_5.5 with a sulfide SE at various weight ratios (70/30, 75/25, 80/20, and 85/15). The electronic and ionic conductivities and electrochemical capacities of these composites were systematically analyzed and compared. In our previous study^[20], the electronic and ionic conductivities of pure Li₈VS_5.5 were reported to be 5.9 × 10^-2 and 5.6 × 10^-6 S cm^-2 at room temperature, respectively.

Figure 3A shows the electronic conductivities of the fabricated composite cathodes measured by DC polarization at 25 °C. The electronic conductivity is enhanced by increasing the weight ratio of Li₈VS_5.5. The electronic conductivity of electrodes containing 70 wt.% of Li₈VS_5.5 is higher than the ionic conductivity of the SE (ca. 2 × 10^-3 S cm^-1); therefore, the Li₈VS_5.5 composite cathodes exhibit sufficient electronic conductivity. The ionic conductivity values for each composite electrode measured by DC polarization at 25, 45, and 60 °C are shown in Figure 3B. The ionic conductivities of the composite electrodes decrease with increasing content of Li₈VS_5.5. Notably, the ionic conductivities of the composite electrodes are lower than their electronic conductivities, even at 60 °C, indicating that ionic conduction is a rate-limiting process in the composite electrodes.

Figure 3. (A) Electronic and (B) ionic conductivities of composite cathodes comprising Li₈VS_5.5 and the SE in weight ratios of 70/30, 75/25, 80/20, and 85/15. The electronic conductivities are measured at 25 °C, and the ionic conductivities are measured at 25, 45, and 60 °C. (C) Discharge capacities per composite cathode for ASS cells with different weight ratios of Li₈VS_5.5 and the SE and the same amount of Li₈VS_5.5. (D) Charge-discharge curves of ASS full cells containing Li₈VS_5.5 and Si electrodes. The mass loadings of the positive and anodes are 33 and 10.4 mg cm^-2, respectively. The cell is charged and discharged with a constant current and voltage at a rate of 0.01 C (until a rate of 0.001 C) over cut-off voltages of 1.22-3.62 V at 60 °C.

Figure 3C shows the discharge capacities of cells containing composite cathodes with various weight ratios of Li₈VS_5.5 and the SE at 60 °C. Each composite cathode was fabricated with the same amount of Li₈VS_5.5, and the discharge capacities of the fabricated cells were calculated using the amount of composite cathode. The ionic conductivity of the cells decreases with increasing weight ratio of Li₈VS_5.5. Moreover, the amount of composite cathode increases, and the electronic conductivity decreases with decreasing weight ratio of Li₈VS_5.5. The cell with a Li₈VS_5.5 to SE weight ratio of 80/20 shows the highest discharge capacity owing to an optimal balance of ionic and electronic conductivities.

Figure 3D shows the charge-discharge curves of ASS full cell with Li₈VS_5.5 and Si electrodes. The cathode contains 33 mg cm^-2 of the composite cathode consisting of Li₈VS_5.5 and SE in a weight ratio of 80/20, whereas the anode contains 10.4 mg cm^-2 of the Si composite electrode. The size of the full cell is a 1 × 1 centimeter square. Considering a discharge capacity of ca. 15 mAh cm^-2 and an average operating voltage of 1.646 V, 20 stacked A4 size (210 mm × 297 mm) ASSBs are expected to show high volumetric and gravimetric energy densities of 853 Wh L^-1 and 515 Wh kg^-1, respectively, in a calculation. The cycling performance of the ASS full cell is shown in Supplementary Figure 3A. The capacity retention rate decreases to ~80% after the 10th cycle. Because Li₈VS_5.5 shows large volume changes during the charge-discharge process, the cell with a high mass loading shows capacity degradation owing to the disconnection of conduction paths between the solid and solid interfaces. Supplementary Figure 3B demonstrates that a half-cell containing Li₈VS_5.5 with a low mass loading of 6.4 mg cm^-2 exhibits excellent cycling performance. Supplementary Figure 3C presents the charge-discharge profiles for the first five cycles of the same cell. With increasing cycle numbers, electrode polarization decreased while capacity improved, indicating progressive activation of the electrode. This activation behavior in the initial cycles is consistent with that observed in previously reported ASS lithium-sulfur batteries^[29-31].

The volume-change rate of Li₈VS_5.5 during the charging process was calculated by estimating the molar volumes of Li₈VS_5.5 and VS_5.5 listed in Supplementary Table 3. The molar volume of Li₈VS_5.5 was calculated from its molar mass and measured density value (2.06 g cm^-3). The molar volume of VS_5.5 was calculated from the molar ratio, molar mass, and density values of S and VS₂. After charging, Li₈VS_5.5 shows a lower shrinking rate (36.5%) than Li₂S (42.1%). Moreover, compared with Li₂S, Li₈VS_5.5 shows a better electronic conductivity and volume change rate. To improve the cycle performance of ASSBs containing Li₈VS_5.5, the volume changes of Li₈VS_5.5 during charge-discharge tests should be investigated by operando measurements. Several approaches have been proposed for reducing crack formation in cells, such as selecting anodes with large volume changes to adjust the total volume change of cells. Details regarding the volume changes of Li₈VS_5.5 will be discussed in our next paper.

Reaction mechanism of Li₈VS_5.5 cathodes

Li₈VS_5.5 enables the construction of high-energy-density ASSBs. To investigate the structural changes during charging and discharging, synchrotron XRD measurements were conducted on composite Li₈VS_5.5 cathodes before and after charge and discharge [Figure 4]. For the XRD measurements, the ASSBs with a Li₈VS_5.5 to SE weight ratio of 85/15 were fabricated to detect structural changes of Li₈VS_5.5 clearly by increasing amount of Li₈VS_5.5 in composite cathodes. The XRD patterns of Li₈VS_5.5 and the SE are shown in Figure 4. The XRD pattern before charging [Figure 4A] contains diffraction peaks attributable to Li₂S, LiVS₂, and the SE, whereas that after charging [Figure 4B] mainly contains diffraction peaks attributable to the SE, possibly owing to the formation of amorphous S by the delithiation of Li₂S. In addition, the diffraction peaks attributed to LiVS₂ are shifted to higher angles because of delithiation from LiVS₂. Moreover, the XRD pattern after charging indicates the partial decomposition of the SE to LiCl. The XRD pattern recorded after the discharging process [Figure 4C] contains diffraction peaks attributed to Li₂S resulting from lithiation to S. Furthermore, the diffraction peaks corresponding to LiVS₂ shift to lower angles after discharge, suggesting lithiation to form VS₂. Notably, elemental S and LiCl remain in the composite cathodes post-discharge. Modifications such as O-doping have been proposed as effective strategies to mitigate the decomposition of the SE^[32]. Table 1 shows the crystallographic data of LiVS₂ in the Li₈VS_5.5 composite cathodes and reference data for LiVS₂ and VS₂^[33]. The a and c axes shrink after charging, corresponding to structural changes from LiVS₂ to VS₂. After discharging, the a and c axes change to their pristine states, indicating the reversible formation of LiVS₂.

Figure 4. Rietveld analysis of the synchrotron XRD patterns of Li₈VS_5.5 composite cathodes (A) before charging, (B) after charging, and (C) after discharging.

Figure 5 shows the S K-edge spectra of the composite cathodes after the first and third charge and discharge. The spectra of the samples after the first charge and discharge are similar to the patterns after the third charge and discharge, respectively, suggesting that the structure changes reversibly over a few cycles.

Figure 5. S K-edge XANES spectra of Li₈VS_5.5 composite cathodes before and after the first and third charge and discharge tests.

The XRD patterns and XANES spectra before and after charge-discharge tests indicate reversible delithiation and lithiation. Cryo-STEM measurements were conducted on the Li₈VS_5.5 composite cathode to investigate nanoscale structural changes after the second discharge test. Figure 6A and Supplementary Figure 4 show the bright-field STEM (BF-STEM) and high-angle annular dark-field STEM (HAADF-STEM) images of the composite electrodes. A red square highlights the EDX measurement area. Figures 6B-G show the annular dark field STEM (ADF-STEM) images and count mappings of the P, Cl, V, S, and O K-edges of the measurement area. In the mapping results of P, Cl, and V, the bright contrast areas highlighted by yellow circles in the ADF-STEM image and the other areas indicate the SE and Li₈VS_5.5, respectively. Figure 7A-D shows the STEM-EELS mapping results of the measurement area; the yellow circles and other areas in the HAADF-STEM image indicate SE and Li₈VS_5.5, respectively. The intensity of Li and S in the Li₈VS_5.5 area decreases toward the interface between Li₈VS_5.5 and the SE, which is highlighted with orange dotted circles. The thickness of the low-intensity area is ~200 nm. No low-intensity area was observed before the charge-discharge test (not shown here); therefore, the interface was assumed to change during charge-discharge testing. Moreover, Li in the SE area showed low intensity, suggesting that the SE near Li₈VS_5.5 was delithiated.

Figure 6. (A) BF-STEM images of composite cathodes consisting of Li₈VS_5.5 and the SE after the second discharge test. The EDX measurement area is highlighted by a red square. (B) ADF-STEM image and EDX mappings of (C) P, (D) Cl, (E) V, (F) S, and (G) O K-edges. Yellow circles indicate the SE areas.

Figure 7. (A) HAADF-STEM images of composite cathodes consisting of Li₈VS_5.5 and the SE after the second discharge test. Yellow and orange dotted circles indicate the SE area and interface between Li₈VS_5.5 and the SE, respectively. STEM-EELS mappings of (B) Li, (C) S, and (D) V. (E) Sets of decomposed spectra of components 1-3 and (F-H) their distribution maps.

Figure 7E-H shows decomposed spectra and their distribution mappings, respectively. Component 1 is distributed in the bulk of Li₈VS_5.5, whereas component 2 is observed at the interface of Li₈VS_5.5 and the SE. The V L-edge peak of component 2 is shifted to a higher position than that of component 1, indicating an increase in the valence of V. Moreover, the peak intensity of the O K-edge in component 2 is higher than that in component 1. Component 3, which shows a low intensity for the V L-edge and a high intensity for the O K-edge, is distributed mainly in the SE areas. Figure 7 indicates a deficit of Li and S and an increase in the valence of V in the Li₈VS_5.5 region near the interface of Li₈VS_5.5 and the SE, suggesting the formation of an interphase between Li₈VS_5.5 and the SE after a few charge-discharge cycles. In addition, the SE undergoes delithiation and partial decomposition by oxygen. Maintaining a low oxygen content can improve the cycle performance of ASSBs with Li_xVS_y. Moreover, investigating the behavior of the Li_xVS_y-SE interphase is critical for an accurate battery performance analysis.

Comparison with other lithium-containing vanadium sulfide electrodes

A comparison of Li₈VS_5.5, 90(0.75Li₂S·0.25V₂S₃)·10LiI^[24], and core-shell Li₂S/LiVS₂^[25] is provided in Table 2. Li₈VS_5.5 contains the highest amount of Li₂S, which is an insulator; however, because of the structure of the nanocomposite in the amorphous matrix, high electronic conductivity was obtained. Despite containing a high amount of Li₂S, ASSBs with Li₈VS_5.5 (even those with high loading) show the highest capacity among all the ASSBs with lithium-containing vanadium sulfide materials because nanosized LiVS₂ in the amorphous matrix forms sufficient electron conduction paths. Our previous study demonstrated that Li₅VS₄ and Li₆VS_4.5 exhibit superior rate performance than Li₈VS_5.5. In particular, the Li₆VS_4.5 half-cell delivered a higher areal capacity at a relatively high current density of 1.3 mA cm^-2 compared with 90(0.75Li₂S·0.25V₂S₃)·10LiI^[24] and core-shell Li₂S/LiVS₂^[25] [Supplementary Table 4]. These results indicate that tuning the composition of Li_xVS_y is an effective strategy for improving rate performance.