Fabrication and Lithium Intercalation Properties of Graphite-Li₁₀P₃S₁₂Br Anode Composites

Abstract

A Li₁₀GeP₂S₁₂-type lithium-ion conductor Li₁₀P₃S₁₂Br was investigated as a potential solid electrolyte for composite anodes in all-solid-state lithium-ion batteries. Cyclic voltammetry measurements using an Au/Li₁₀P₃S₁₂Br/Li cell revealed that Li₁₀P₃S₁₂Br possesses a sufficiently wide electrochemical window, making it suitable for anode reactions at low potentials. Spherical graphite-Li₁₀P₃S₁₂Br composite anodes, fabricated via rotary mixing, exhibited lithium (de)intercalation activity, demonstrating the feasibility of Li₁₀P₃S₁₂Br as an electrolyte for all-solid-state battery anodes. Composite anodes with a higher proportion of Li₁₀P₃S₁₂Br relative to graphite exhibited improved cycle retention of charge-discharge capacities. A composite comprising graphite (d₅₀: 8 µm) and Li₁₀P₃S₁₂Br (d₅₀: 0.3 µm) in a 20:80 wt.% ratio achieved a discharge capacity of 365 mAh g⁻¹ at the 30th cycle. In contrast, the 50:50 wt.% composite exhibited a notable decrease in lithium intercalation capacity in the stage 2 (LiC₁₂) and stage 1 (LiC₆) regions. These results suggest that reducing the lithium diffusion distance within graphite particles is crucial for enhancing the intercalation properties of graphite/Li₁₀P₃S₁₂Br composite anodes.

 

This Paper was Originally Published in Japanese in J. Jpn. Soc. Powder Powder Metallurgy 72 (2025) 129–136.

1. Introduction

All-solid-state lithium-ion batteries, which replace the organic electrolyte used in conventional lithium-ion batteries with a solid electrolyte, are attracting attention as next-generation energy storage devices owing to their excellent output and reliability over a wide temperature range [1]. These batteries typically consist of a cathode layer, an electrolyte (separator) layer, and an anode layer. In the cathode and anode layers, composites of solid electrolyte and active materials are employed to ensure ionic conductivity and provide sufficient reaction surface area. Because solid electrolytes do not flow easily, using electrolytes with appropriate ionic conductivity, electrochemical stability, and processability tailored to each layer is advantageous in battery design. Sulfide and halide electrolytes exhibit superior lithium-ion conductivity and processability compared with oxide electrolytes [2, 3]. Crystalline sulfides, such as Li₁₀GeP₂S₁₂ (LGPS) and argyrodites, exhibit ionic conductivities exceeding 10 mS cm⁻¹ at room temperature [1, 4, 5]. High-conductivity LGPS-type Li_9.54Si_1.74P_1.44S_11.7Cl_0.3 (28 mS cm⁻¹) and Li_9.54[Si_0.6Ge_0.4]_1.74P_1.44S_11.1Br_0.3O_0.6 (32 mS cm⁻¹), have demonstrated low ionic resistance in cathode and electrolyte layers, enabling high power output and the fabrication of thick high-capacity electrodes [6, 7]. Although incorporating electrolyte into the electrode layers enhances ionic conductivity and reaction surface area, it concurrently reduces the proportion of active material, thereby lowering energy density. Accordingly, efforts are underway to increase the active material ratio in the cathode layer by controlling the microstructure through particle morphology, mixing ratios, mixing methods, and pressing conditions [8, 9]. On the anode side, many sulfide electrolytes exhibit poor electrochemical stability at low potentials (approximately 0 V vs. Li/Li⁺), prompting research into improving interfacial stability with metal electrodes such as In–Li or Li [10]. Composite anodes incorporating graphite, considered the primary anode material, have been reported with electrolytes such as LGPS-type Li_9.6P₃S₁₂ [11], argyrodite-type Li₆PS₅Cl [12–14], and LiI–Li₂S–P₂S₅ glasses [9, 15–17], though most of these involve electrolytes with relatively low ionic conductivities. Furthermore, only limited studies have explored microstructural control and reaction analysis in these systems [18–20]. Therefore, from battery development and reaction analysis perspectives, expanding the options of sulfide electrolytes for use in anode layers is important.

Li₁₀P₃S₁₂Br is a halide-sulfide electrolyte with an LGPS-type structure [21]. In this structure, all Ge⁴⁺ ions in LGPS are replaced by P⁵⁺ ions, and Br⁻ is introduced for charge compensation, resulting in a unit cell with higher anion content than LGPS. Although it belongs to the same tetragonal system (space group P4₂/nmc) as LGPS, Br⁻ substitutes for the S2 (8g) and S3 (8g) positions among the three types of PS₄³⁻ tetrahedra, whereas S²⁻ partially occupies a new S4 (2a) site. Its synthesis temperature ranges from 463 to 493 K, which is lower than that of LGPS (823 K), and despite its poor sinterability, it exhibits a high ionic conductivity of 5.8 mS cm⁻¹ at 300 K. Although previous studies have reported that introducing Br⁻ into L–P–S electrolytes enhances electrochemical stability against lithium metal anodes [21, 22], Li₁₀P₃S₁₂Br has not yet been investigated as an electrolyte for anode layers. In this study, composite anodes composed of graphite and Li₁₀P₃S₁₂Br were fabricated to evaluate the applicability of this electrolyte for anode layers based on charge–discharge cycling characteristics. Furthermore, the effect of varying the composition ratio of graphite to Li₁₀P₃S₁₂Br on the microstructure and charge–discharge cycling characteristics was investigated.

2. Experimental Procedure

All measurements were conducted under an argon atmosphere or in a sealed environment to avoid air exposure. Li₁₀P₃S₁₂Br was synthesized via a solid-state reaction [21]. Li₂S (Mitsu Chemical, ≥99.9%), P₂S₅ (Sigma-Aldrich, ≥99%), and LiBr (Sigma-Aldrich, 99.999%) were weighed in stoichiometric amounts inside an argon glove box and mixed using an agate mortar for 30 min. The mixture was sealed in a ZrO₂ pot and milled in a planetary ball mill (Fritsch Japan, Pulverisette 7) at 380 rpm for 40 h. The obtained sample was pressed into pellets, vacuum-sealed in a quartz tube, and heated at 478 K for 4 h. After natural cooling, the pellets were ground in a mortar for 2 min to obtain Li₁₀P₃S₁₂Br powder. The crystalline phase was identified using X-ray diffraction (Rigaku, Smartlab, Cu Kα source). Particle morphology was examined by field-emission scanning electron microscopy (SEM; Hitachi High-Tech, Regulus 8230), and the particle size distribution was determined using a laser diffraction/scattering particle size analyzer (PSA; Horiba, Partica LA-960). For the PSA analysis, 5 mg of Li₁₀P₃S₁₂Br was dispersed in 20 mL of heptane with SPAN80 (Aldrich) as the dispersant, and the dispersion was achieved by ultrasonic treatment for 15 min.

The ionic conductivity of Li₁₀P₃S₁₂Br was evaluated via AC impedance spectroscopy using pellet-type symmetric cells with Au electrodes. 50 mg of Li₁₀P₃S₁₂Br powder was placed in a Polytetrafluoroethylene (PTFE) cylindrical die (5 mm diameter) and pressed into a pellet using a uniaxial press (Riken Seiki P-6) at 355 MPa. Gold powder was applied to both pellet surfaces and pressed at 710 MPa to ensure good electrode contact. The pellet was then installed in a stainless-steel sealed cell and pressed further with a torque equivalent to 10 N m. AC impedance measurements were performed at 298 K using a frequency response analyzer (Solartron Analytical, 1260A) with an applied voltage of 400 mV over a frequency range of 7 × 10⁶–10⁻¹ Hz. Electrochemical stability was determined by cyclic voltammetry using an Au/Li₁₀P₃S₁₂Br/Li cell. Measurements were carried out with an electrochemical measurement system (Bio-Logic, VSP-300) over a voltage range of −0.05 to 1.52 V at a scan rate of 0.1 mV s⁻¹.

Graphite–Li₁₀P₃S₁₂Br composite anodes were prepared by dry rotary mixing [23]. Spherical graphite (Nippon Graphite CGB-8R, average particle size 8 µm, or Fuji Graphite Works PTG-15, average particle size 15 µm) was dried overnight at 353 K and then weighed together with Li₁₀P₃S₁₂Br to achieve weight ratios of 50:50, 30:70, and 20:80, with a total weight of 30 mg. The estimated volume ratios based on bulk densities were 84:16, 69:31, and 57:43. The sample and three agate balls (φ 3 mm) were placed in a cylindrical glass tube and mixed on a rotary platform (Nitto Kohki, ANZ-10S0) at 140 rpm for 20 min. The electrochemical characteristics of the fabricated composite anodes were examined in half-cells with an In–Li counter electrode. 80 mg of Li₁₀P₃S₁₂Br electrolyte was pressed into pellets at 107 MPa for 1 min. The composite anode (8 mg) was spread on one face of the pellet, covered with a Cu mesh and Cu foil (φ 10 mm), and pressed at 533 MPa to form the working electrode. On the opposite face, In foil (φ 10 mm) and Li foil (φ 5 mm) were pressed with a Cu mesh and Cu foil at 210 MPa to form an In–Li alloy counter electrode. The assembled cell was then installed in a stainless-steel sealed cell and further pressed with a torque equivalent to 20 N m to form the all-solid-state cell. Room-temperature charge–discharge tests were performed using a charge–discharge apparatus (Toyo System, TOSCAT-3100). For the Li insertion process into the graphite anode (charging), the lower cut-off voltage was set to −0.65 V (−0.03 V vs. Li/Li⁺) [24] and the cut-off capacity was set at 372 mAh g⁻¹, assuming the lithium deintercalation reaction from graphite (6C + Li⁺ + e⁻ $ \rightleftarrows $ LiC₆). Constant-current and constant-current–constant-voltage methods were employed for this test, where the constant-current test was conducted at a current density of 0.05 C (18.6 mA g⁻¹) and the maximum duration for the constant-voltage step was 20 h. The Li deintercalation process from graphite (discharging) was conducted using only the constant-current method with an upper cut-off voltage of 0.9 V.

3. Results and Discussion

Figure 1 shows the X-ray diffraction pattern, SEM image (×3000), and particle size distribution of the synthesized Li₁₀P₃S₁₂Br. Most of the diffraction patterns can be indexed to the LGPS-type Li₁₀P₃S₁₂Br [21], although all reflections exhibited broad peak widths, suggesting poor crystallinity. Because Li₁₀P₃S₁₂Br decomposes into β-Li₃PS₄ above approximately 500 K [25], it was synthesized at a relatively low temperature (478 K) for an LGPS-type material, which is believed to be the cause of its poor crystallinity. The lattice parameters were estimated as a = 8.7721(13) and c = 12.510(2) Å, slightly differing from the reported values (a = b = 8.76184 Å, c = 12.5332 Å); the a and b axes are slightly larger, whereas the c axis is slightly smaller. Reflections at 28° and 32.5°, originating from the raw material LiBr, were observed, suggesting a minor compositional deviation. The SEM image revealed that particles smaller than 1 µm tended to aggregate, forming clusters ranging from several micrometers to approximately 10 µm. In the particle size distribution profile obtained after solvent dispersion, the d₅₀ particle size was 0.34 µm, indicating that the submicron particle aggregates observed in the SEM image readily disaggregated. Generally, when submicron sulfide electrolyte particles are mixed with micrometer-sized active materials, the smaller electrolyte particles can surround the larger ones, thereby increasing interfacial contact area and facilitating the formation of ion conduction pathways [8]. As the graphite particles used in this study were approximately 10 µm in size, the synthesized Li₁₀P₃S₁₂Br particles were expected to be suitably sized for use as an electrolyte in composite anodes. Furthermore, the particle size distribution profile showed a small peak centered around 1.2 µm, indicating that particles in the range of 1–10 µm constitute about 20% of the total, which corresponds to the several-µm particles partially observed in the SEM image.

Fig. 1

(a) XRD pattern, (b) SEM image, and (c) particle size distribution of Li₁₀P₃S₁₂Br. The XRD pattern of Li₁₀P₃S₁₂Br (ICSD #54662) [23] is depicted in (a). (online color)

Figure 2(a) presents the Nyquist plot of the impedance measured for the Au/Li₁₀P₃S₁₂Br/Au cell at 298 K. A semicircle with a peak frequency (f) of 5.0 MHz and a diameter (R) of 170 Ω was observed. In the equivalent circuit model for the pressed powder of the solid electrolyte, two sets of parallel RC circuits, corresponding to ionic conduction within the grains and along the grain boundaries, are included. The grain and grain boundary capacitances were estimated to be on the order of 10⁻¹¹ F cm⁻² and 10⁻¹⁰–10⁻⁷ F cm⁻², respectively [26]. Calculating the capacitance C from the relation RC = 1/(2πf) yields C = 6.1 × 10⁻⁹ F cm⁻². The similarity in time constants of these components suggests that their semicircles overlapped. In this case, the diameter of the semicircle represents the sum of the ionic conduction resistances within the grains and at the grain boundaries, resulting in a total ionic conductivity of 4.0 mS cm⁻¹, which is slightly lower than that reported for hot-pressed Li₁₀P₃S₁₂Br pellets (5.8 mS cm⁻¹). This difference is likely due to the pressed powder samples having a lower interparticle contact area and higher grain boundary ionic resistance [25]. Figure 2(b) shows the cyclic voltammogram of the Au/Li₁₀P₃S₁₂Br/Li cell. In the initial cycles, a reduction peak was observed at approximately 0.18 V, along with two oxidation peaks at around 0.38 and 0.43 V, corresponding to the lithium alloying/dealloying reaction of the Au electrode (xLi⁺ + Au + xe⁻ $ \rightleftarrows $ Li_xAu) [27]. This reaction exhibited poor reversibility, as evidenced by the decrease in the reduction and oxidation currents, with the peaks disappearing by the 25th cycle. In contrast, reversible redox peaks were observed at around 0.35 and 1.00 V, suggesting a redox reaction distinct from the Au electrode’s alloying/dealloying behavior. Electrolyte-derived redox activity due to lithium (de)intercalation has been reported for Li₆PS₅Cl [28] and Li₃PS₄ [29] electrolytes, and may correspond to the reversible redox peaks observed in Li₁₀P₃S₁₂Br. The redox peak potentials and current values remained stable over repeated cycles, and no increase in interfacial lithium-ion conduction resistance was observed. These findings suggest that Li₁₀P₃S₁₂Br offers a sufficiently wide electrochemical window for application as a solid electrolyte with anode materials operating at low potentials.

Fig. 2

(a) Nyquist plot of the EIS result for Au/Li₁₀P₃S₁₂Br/Au and (b) cyclic voltammograms of Au/Li₁₀P₃S₁₂Br/Li. (online color)

Graphite (8 µm)-Li₁₀P₃S₁₂Br composite pellets (50:50 wt.%) were used to fabricate all-solid-state cells (graphite/Li₁₀P₃S₁₂Br/In–Li), which were tested at room temperature under a constant-current charge–discharge protocol at 0.05 C. Figure 3 presents the obtained charge–discharge curves, the evolution of capacity and Coulombic efficiency over cycles, and the cycle-dependent changes in the charging capacity across various voltage regions. Lithium alloying and dealloying of In proceed via a two-phase coexistence reaction between In and In–Li at 0.62 V vs. Li/Li⁺ [24, 30]. The electrode potential of the graphite–Li₁₀P₃S₁₂Br composite, calculated using the In–Li counter electrode as a pseudo-reference, is plotted on the right axis. In the initial charge, a capacity of approximately 130 mAh g⁻¹ was observed as the cell voltage decreased from 0.4 to −0.4 V. This capacity is mainly attributed to the reduction in side reactions, likely due to impurities on the graphite surface or from the electrolyte [31]. A voltage plateau was observed near −0.45 V (0.17 V vs. Li/Li⁺), after which the voltage gradually decreased until the cutoff capacity of 372 mAh g⁻¹ was reached at −0.56 V (0.06 V vs. Li/Li⁺) at the end of the charge. The initial discharge capacity was 190 mAh g⁻¹, corresponding to a Coulombic efficiency of 51.0%. During lithium intercalation into graphite, lithium is arranged regularly between the graphene layers, which are stacked along the c-axis, thereby forming a staged structure [32]. Depending on the stage number (n) and specific lithium arrangement within the layers, graphite exhibits various crystal structures [33]. In liquid electrolyte lithium-ion batteries, voltage plateaus corresponding to transitions to LiC₂₇ (stage 3), LiC₁₂ (stage 2), and LiC₆ (stage 1) occur at approximately 0.2, 0.12, and 0.08 V vs. Li/Li⁺, respectively. Assuming an In–Li electrode potential of 0.62 V vs. Li/Li⁺ [24, 33], the transitions to LiC₂₇, LiC₁₂, and LiC₆ in the all-solid-state cells used here would correspond to potentials below −0.42 V, −0.50 V, and −0.54 V, respectively. During the initial charge, side reactions appeared to inhibit significant lithium intercalation in the stage 1 region, thereby causing the cutoff capacity to be reached prematurely. Although the cutoff capacity was reached during charging in cycles 2 through 4, the charge-end voltage further decreased to −0.64 V, reaching the reaction region of stage 1. The discharge capacity gradually increased, reaching 334 mAh g⁻¹ with a Coulombic efficiency of 90% in the 4th cycle, indicating an improved ratio of lithium intercalation relative to side reactions during charging. However, by the 5th cycle, charging terminated upon reaching the cutoff voltage, with a charging capacity of 316 mAh g⁻¹ and a discharge capacity of 298 mAh g⁻¹, indicating a reduction in the lithium intercalation amount. Thereafter, the charge and discharge capacities decreased, with the discharge capacity at the 30th cycle being 190 mAh g⁻¹.

Fig. 3

(a) Charge-discharge curves and (b) variations of charge-discharge capacities and efficiency of the graphite (8 µm)-Li₁₀P₃S₁₂Br anode composite with a weight ratio of 50:50. (c) Capacity variation in different charging voltage regions of >−0.45 V (gray), −0.45 to −0.53 V (red), and <−0.53 V (blue). The digits in (a) represent the cycle number. (online color)

Figure 3(c) shows the decomposition of charging capacity by voltage region. The voltage region above −0.45 V corresponds to stage 4 and higher, the region between −0.45 and −0.53 V corresponds to stages 3 and 2, and the region below −0.53 V corresponds to lithium intercalation in stage 1. Additionally, the region below −0.53 V includes contributions from side reactions. From the 2nd to the 30th cycle, the charging capacity in stage 3 and the higher voltage region was maintained. The overlapping of the charge curves above −0.53 V (Fig. 3(a)) indicated that the overpotentials did not result in a further decrease in the charging voltage. Similar to the interfacial behavior observed for the Au and Li electrodes (Fig. 2(b)), no highly resistive interphase (e.g., due to Li₁₀P₃S₁₂Br decomposition) was formed. In contrast, the capacity below −0.53 V decreased after the 5th cycle. Compared with the discharge capacity indicated by the filled circles, the irreversible capacity also decreased, suggesting that the amount of lithium intercalation in the stage 1 region diminished. Lithium intercalation in the graphite composite involves several processes, including lithium diffusion within graphite, electronic conduction between graphite particles, and lithium-ion transport across Li₁₀P₃S₁₂Br particles and the graphite–Li₁₀P₃S₁₂Br interfaces. If either the electronic conduction between graphite particles or the lithium-ion transport across Li₁₀P₃S₁₂Br particles and the graphite–Li₁₀P₃S₁₂Br interfaces degrades during cycling, it would lead to a decrease in charging voltage and an increase in discharging voltage, independent of the intercalation reaction potential. Novák et al. [34] reported that lithium intercalation from stage 2 to stage 1 was slower than that in other stages. They proposed that, in the stage 1 phase, where lithium is densely accumulated between the graphene layers, hindering further lithium movement, the migration of the phase boundary from the surface of stage 2 to the interior of the newly formed stage 1 particles becomes the rate-limiting step. Additionally, during charge–discharge, graphite particles undergo approximately 13% volumetric expansion and contraction owing to lithium (de)intercalation [35]. When graphite contracts upon discharge, the electrolyte may fail to maintain the interfacial contact. In such cases, during the subsequent charging, lithium ions must traverse across the entire particle to access isolated regions. In contrast to liquid electrolytes, which can infiltrate the voids within the spherical graphite particles and reduce diffusion distance from the interface, solid electrolytes cannot penetrate the particles. These factors likely increased lithium diffusion resistance, causing the charging voltage to drop and the cutoff voltage to be reached before the phase transition was completed.

To evaluate the effect of the increased lithium diffusion distance within graphite, a composite using graphite particles with a mean diameter of 15 µm was prepared, and its charge–discharge behavior was investigated. Figure 4 shows the charge–discharge curves, cycle-to-cycle changes in capacity and Coulombic efficiency, and cycle-dependent changes in charging capacity over various voltage regions for all-solid-state cells fabricated using graphite (15 µm)–Li₁₀P₃S₁₂Br composite pellets (50:50 wt.%). After 26 cycles of constant-current charge–discharge testing (Fig. 4(a)), the cell was allowed to rest, then charging was resumed in constant current–constant voltage mode (Fig. 4(b)). In the 2nd cycle, a discharge capacity exceeding 300 mAh g⁻¹ was observed, demonstrating that Li₁₀P₃S₁₂Br effectively functioned as an electrolyte, even with 15 µm graphite. However, from the 3rd cycle onward, the capacity decreased, reaching 113 mAh g⁻¹ in the 25th cycle (Fig. 4(c)). Analysis of the charging capacity in different voltage regions revealed a significant decrease below −0.53 V (Fig. 4(d)), suggesting that, similar to the graphite (8 µm)–Li₁₀P₃S₁₂Br system, lithium intercalation from stage 2 to stage 1 is hindered. In cycles 2 through 25, the charging capacity below −0.53 V was maintained at only 19.7% (254 → 50 mAh g⁻¹), which is lower than that observed for 8 µm graphite (31.8%, 251 → 80 mAh g⁻¹). The reduced cycling stability with larger graphite particles supports the model in which increased lithium diffusion distances lead to a higher resistance. When the capacity-depleted cell was switched to constant-current constant-voltage charging, the charge and discharge capacities increased to approximately 240 mAh g⁻¹ (Figs. 4(b) and 4(c)). Furthermore, Fig. 4(d) shows that the charging capacity in the −0.53 to −0.65 V region recovered, indicating that extending the reaction time during constant-voltage charging promoted lithium-ion diffusion within the graphite particles. In addition, the charging capacity measured in the constant-current region increased from 113 (at the 25th cycle under constant-current charging) to 238 mAh g⁻¹ (at the 30th cycle) after constant-voltage charging, suggesting a reduction in diffusion resistance within the graphite particles. During constant-voltage charging, the volumetric expansion of graphite particles due to lithium intercalation is believed to improve the interfacial contact with the electrolyte and the contact between the graphite particles, thereby shortening the lithium diffusion pathway through graphite. However, in subsequent constant-current–constant-voltage charging and constant-current discharging cycles, the capacity gradually decreased, with the discharge capacity at the 50th cycle being 144 mAh g⁻¹. Further, with increasing cycle number, resistance increased to the extent that even constant-voltage charging at −0.65 V could not sustain further lithium intercalation.

Fig. 4

(a), (b) Charge-discharge curves and (c) variations of charge-discharge capacities and efficiency of the graphite (15 µm)-Li₁₀P₃S₁₂Br anode composite with a weight ratio of 50:50. (d) Capacity variation in different charging voltage regions of >−0.45 V (gray), −0.45 to −0.53 V (red), and <−0.53 V (blue). The digits in (a), (b) represent the cycle number. The cell was charged (a) in CC mode for the initial 25 cycles and (b) in CC-CV mode for subsequent cycles. (online color)

Increasing the sulfide electrolyte content in the electrode composite is expected to enhance the active material–electrolyte interfacial area and ion conduction pathways among the electrolyte particles [8]. Additionally, soft sulfide electrolytes have been reported to buffer the stress arising from the volumetric expansion of the active material, thereby improving the mechanical properties [36]. Figure 5 shows the constant-current charge–discharge curves and cycle-to-cycle changes in charging capacity across different voltage regions for all-solid-state cells fabricated using anode composites with weight ratios of 30:70 and 20:80. In both cases, a pronounced decline in the charging capacity, which is typical for a 50:50 composite, was absent. The 30:70 composite maintained a stable discharge capacity of approximately 300 mAh g⁻¹ from the 2nd to the 18th cycle, followed by a gradual decrease from the 19th to the 30th cycle, mainly due to reduced charging capacity below −0.45 V. In the 20:80 composite, the discharge capacity increased up to the 30th cycle, reaching a maximum of 365 mAh g⁻¹, corresponding to a capacity retention of 98.1%, which indicates excellent cycling stability. Furthermore, a clear voltage plateau at −0.55 V corresponding to the two-phase reaction during lithium intercalation from the stage 2 phase (LiC₁₂) to the stage 1 phase (LiC₆) was observed [37, 38]. This suggests that a more uniform intercalation reaction occurred within the composite, thereby improving the utilization and retention of the graphite particles down to a lower cutoff voltage. Furthermore, a 30 mAh g⁻¹ discharge plateau near −0.61 V, attributed to lithium dissolution, indicates that graphite utilization improved to approximately 90%.

Fig. 5

(a), (b) Charge-discharge curves and (c), (d) capacity variations in different charging voltage regions of >−0.45 V (gray), −0.45 to −0.53 V (red), −0.53 to −0.62 V (blue), and <−0.62 V (green) of graphite (8 µm)-Li₁₀P₃S₁₂Br anode composites with different weight ratios of (a), (c) 30:70 and (b), (d) 20:80. The digits in (a), (b) represent the cycle number. (online color)

Figure 6 shows schematic diagrams of the microstructural changes in the graphite–Li₁₀P₃S₁₂Br composites with different compositions before and after cycling. Within the composite, the Li₁₀P₃S₁₂Br particles can be divided into those that are connected to Li₁₀P₃S₁₂Br in the separator layer, thereby forming an interelectrode ion conduction pathway, and those that are not connected to this pathway. When the graphite content was high, the connectivity of the Li₁₀P₃S₁₂Br particles was disrupted by the graphite particles, resulting in a lower proportion of Li₁₀P₃S₁₂Br contributing to the inter-electrode ion conduction pathway. Lithium intercalation proceeds from the graphite particles that are in contact with the Li₁₀P₃S₁₂Br particles, forming the conduction network. In contrast, lithium is indirectly supplied to graphite particles that are not connected to the conduction network, through already intercalated graphite. Intercalation through the stage 1 phase (LiC₆), which exhibits poor lithium mobility, results in high resistance. Furthermore, because the lithium diffusion distance varies among different graphite particles, a nonuniform reaction distribution is easily established [34]. In the 50:50 composite, where the Li₁₀P₃S₁₂Br content was low, a significant fraction of graphite and Li₁₀P₃S₁₂Br particles was not connected to the inter-electrode ion conduction pathway (Fig. 6(a)), resulting in the absence of a clear voltage plateau in the stage 1 region and a lower intercalation capacity. In contrast, in composite with a higher Li₁₀P₃S₁₂Br content (20:80 wt.%), almost all the graphite and Li₁₀P₃S₁₂Br particles are connected to the ion conduction network (Fig. 6(b)), leading to a distinct voltage plateau in the stage 1 region and a higher maximum capacity than that of the 50:50 composite. Additionally, during cycling, the volumetric expansion and contraction of graphite are expected to partially disrupt the conduction pathways within the composite. In composites with high graphite content (50:50 wt.%), the fraction of graphite and Li₁₀P₃S₁₂Br particles not connected to the inter-electrode ion conduction network further increases (Fig. 6(c)), so that lithium cannot reach the isolated graphite regions, resulting in a sudden decrease in capacity. Conversely, increasing the Li₁₀P₃S₁₂Br ratio reduces overall volumetric expansion and contraction per unit volume of the composite, thereby suppressing irreversible microstructural changes. Even if some conduction pathways are blocked, lithium (de)intercalation can proceed through alternative contact points that remain connected to the inter-electrode ion conduction network, preventing a significant increase in the lithium diffusion distance (Fig. 6(d)). Consequently, the 20:80 composite exhibited superior cycle retention compared to of that the 50:50 composite. In summary, these results demonstrate that Li₁₀P₃S₁₂Br effectively functions as a solid electrolyte in graphite-based composites. Moreover, ensuring good interfacial contact between graphite and Li₁₀P₃S₁₂Br and minimizing lithium diffusion distance are key for improving cycle stability.

Fig. 6

Schematics of lithium-conductive pathways due to microstructural changes (a), (b) before and (c), (d) after charge-discharge cycling in graphite/Li₁₀P₃S₁₂Br anode composites with different weight ratios of (a), (c) 50:50 wt.% and (b), (d) 20:80 wt.%. (online color)

4. Conclusions

LGPS-type Li₁₀P₃S₁₂Br was evaluated as a solid electrolyte for anode composites. Cyclic voltammetry of the Au/Li₁₀P₃S₁₂Br/Li cell showed reversible redox reactions at the electrode interface without an increase in interfacial resistance, confirming a sufficiently wide electrochemical window for low-potential anodes. The graphite/Li₁₀P₃S₁₂Br composite anode exhibited charge–discharge capacities exceeding 300 mAh g⁻¹, indicating effective lithium intercalation at the graphite/Li₁₀P₃S₁₂Br interface. Increasing the volume fraction of Li₁₀P₃S₁₂Br in the anode composite improved cycle retention of the charge–discharge capacity, reaching 365 mAh g⁻¹ by the 30th cycle. These results demonstrate that Li₁₀P₃S₁₂Br functions effectively as a solid electrolyte in anode composites, thereby expanding the range of viable electrolyte options. To further improve energy density and cycling stability, shortening the lithium diffusion distance within graphite and optimizing the interfacial area between graphite and Li₁₀P₃S₁₂Br is essential.

Acknowledgments

This work was partially supported by JSPS KAKENHI Grants JP19H05793 and JP24H00042.

REFERENCES

• 1)   N.  Kamaya,  K.  Homma,  Y.  Yamakawa,  M.  Hirayama,  R.  Kanno,  M.  Yonemura,  T.  Kamiyama,  Y.  Kato,  S.  Hama,  K.  Kawamoto and  A.  Mitsui: A lithium superionic conductor, Nat. Mater. 10 (2011) 682–686. doi:10.1038/nmat3066
• 2)   A.  Sakuda,  A.  Hayashi,  Y.  Takigawa,  K.  Higashi and  M.  Tatsumisago: Evaluation of elastic modulus of Li₂S–P₂S₅ glassy solid electrolyte by ultrasonic sound velocity measurement and compression test, J. Ceram. Soc. Jpn. 121 (2013) 946–949. doi:10.2109/jcersj2.121.946
• 3)   H.  Yang and  N.  Wu: Ionic conductivity and ion transport mechanisms of solid-state lithium-ion battery electrolytes: A review, Energy Sci. Eng. 10 (2022) 1643–1671. doi:10.1002/ese3.1163
• 4)   W.D.  Jung,  J.S.  Kim,  S.  Choi,  S.  Kim,  M.  Jeon,  H.G.  Jung,  K.Y.  Chung,  J.H.  Lee,  B.K.  Kim,  J.H.  Lee and  H.  Kim: Superionic Halogen-Rich Li-Argyrodites Using In Situ Nanocrystal Nucleation and Rapid Crystal Growth, Nano Lett. 20 (2020) 2303–2309. doi:10.1021/acs.nanolett.9b04597
• 5)   L.  Zhou,  A.  Assoud,  Q.  Zhang,  X.  Wu and  L.F.  Nazar: New Family of Argyrodite Thioantimonate Lithium Superionic Conductors, J. Am. Chem. Soc. 141 (2019) 19002–19013. doi:10.1021/jacs.9b08357
• 6)   R.  Xu,  Z.  Wu,  S.  Zhang,  X.  Wang,  Y.  Xia,  X.  Xia,  X.  Huang and  J.  Tu: Construction of All-Solid-State Batteries based on a Sulfur-Graphene Composite and Li_9.54Si_1.74P_1.44S_11.7Cl_0.3 Solid Electrolyte, Chem. Eur. J. 23 (2017) 13950–13956. doi:10.1002/chem.201703116
• 7)   Y.  Li,  S.  Song,  H.  Kim,  K.  Nomoto,  H.  Kim,  X.  Sun,  S.  Hori,  K.  Suzuki,  N.  Matsui,  M.  Hirayama,  T.  Mizouchi,  T.  Saito,  T.  Kamiyama and  R.  Kanno: A lithium superionic conductor for millimeter-thick battery electrode, Science 381 (2023) 50–53. doi:10.1126/science.add7138
• 8)   Y.  Yamada,  K.  Watanabe,  H.S.  Kim,  K.  Suzuki,  S.  Hori,  R.  Kanno and  M.  Hirayama: Microstructure Control of LiCoO₂-Li₁₀GeP₂S₁₂ Composite Cathodes by Adjusting the Particle Size Distribution for the Enhancement of All-Solid-State Batteries, Batter. Supercaps 6 (2023) e202300261. doi:10.1002/batt.202300261
• 9)   J.Y.  Kim,  J.  Park,  M.J.  Lee,  S.H.  Kang,  D.O.  Shin,  J.  Oh,  J.  Kim,  K.M.  Kim,  Y.G.  Lee and  Y.M.  Lee: Diffusion-Dependent Graphite Electrode for All-Solid-State Batteries with Extremely High Energy Density, ACS Energy Lett. 5 (2020) 2995–3004. doi:10.1021/acsenergylett.0c01628
• 10)   Y.  Zhu,  X.  He and  Y.  Mo: Origin of Outstanding Stability in the Lithium Solid Electrolyte Materials: Insights from Thermodynamic Analyses Based on First-Principles Calculations, ACS Appl. Mater. Interfaces 7 (2015) 23685–23693. doi:10.1021/acsami.5b07517
• 11)   Y.  Kato,  S.  Hori,  T.  Saito,  K.  Suzuki,  M.  Hirayama,  A.  Mitsui,  M.  Yoneyama,  H.  Iba and  R.  Kannno: High-power all-solid-state batteries using sulfide superionic conductors, Nat. Energy 1 (2016) 16030. doi:10.1038/nenergy.2016.30
• 12)   D.H.  Kim,  D.Y.  Oh,  K.H.  Park,  Y.E.  Choi,  Y.J.  Nam,  H.A.  Lee,  S.M.  Lee and  Y.S.  Jung: Infiltration of Solution-Processable Solid Electrolytes into Conventional Li-Ion-Battery Electrodes for All-Solid-State Li-Ion Batteries, Nano Lett. 17 (2017) 3013–3020. doi:10.1021/acs.nanolett.7b00330
• 13)   M.  Calpa,  N.C.  Rosero-Navarro,  A.  Miura and  K.  Tadanaga: Argyrodite solid electrolyte-coated graphite as anode material for all-solid-state batteries, J. Sol-Gel Sci. Technol. 101 (2022) 8–15. doi:10.1007/s10971-021-05634-7
• 14)   Z.  Zhang,  J.  Wang,  Y.  Jin,  G.  Liu,  S.  Yang and  X.  Yao: Insights on lithium plating behavior in graphite-based all-solid-state lithium-ion batteries, Energy Storage Mater. 54 (2023) 845–853. doi:10.1016/j.ensm.2022.11.023
• 15)   K.  Takada,  T.  Inada,  A.  Kajiyama,  H.  Sasaki,  S.  Kondo,  M.  Watanabe,  M.  Murayama and  R.  Kanno: Solid-state lithium battery with graphite anode, Solid State Ionics 158 (2003) 269–274. doi:10.1016/S0167-2738(02)00823-8
• 16)   M.  Otoyama,  A.  Sakuda,  A.  Hayashi and  M.  Tatsumisago: Optical microscopic observation of graphite composite negative electrodes in all-solid-state lithium batteries, Solid State Ionics 323 (2018) 123–129. doi:10.1016/j.ssi.2018.04.023
• 17)   L.  Höltschi,  J.F.  Jud,  C.  Borca,  T.  Huthwelker,  C.  Villevieille,  V.  Pelé,  C.  Jordy,  M.E.  Lazzi and  P.  Novák: Study of Graphite Cycling in Sulfide Solid Electrolytes, J. Electrochem. Soc. 167 (2020) 110558. doi:10.1149/1945-7111/aba36f
• 18)   M.  Otoyama,  H.  Kowada,  A.  Sakuda,  M.  Tatsumisago and  A.  Hayashi: Operando Confocal Microscopy for Dynamic Changes of Li⁺ Ion Conduction Path in Graphite Electrode Layers of All-Solid-State Batteries, J. Phys. Chem. Lett. 11 (2020) 900–904. doi:10.1021/acs.jpclett.9b03456
• 19)   H.J.  Lee,  H.R.  Kim,  K.J.  Lee,  Y.  Lee,  H.K.  Kim,  W.Y.  Yoon and  J.P.  Ahn: Crack Healing Mechanism by Application of Stack Pressure to the Carbon-Based Composite Anode of an All-Solid-State Battery, ACS Appl. Energy Mater. 5 (2022) 5227–5235. doi:10.1021/acsaem.2c00603
• 20)   G.  Fukunishi,  A.  Ikezawa,  T.  Okajima,  F.  Kitamura,  K.  Suzuki,  M.  Hirayama,  R.  Kanno and  H.  Arai: Impedance Analysis and Cyclability Evaluation of Graphite Composite Electrodes with All-Solid-State Three-Electrode Cells, ACS Appl. Energy Mater. 6 (2023) 10908–10917. doi:10.1021/acsaem.3c01656
• 21)   S.  Song,  S.  Hori,  Y.  Li,  K.  Suzuki,  N.  Matsui,  M.  Hirayama,  T.  Saito,  T.  Kamiyama and  R.  Kanno: Material Search for a Li₁₀GeP₂S₁₂-Type Solid Electrolyte in the Li–P–S–X (X = Br, I) System via Clarification of the Composition–Structure–Property Relationships, Chem. Mater. 34 (2022) 8237–8247. doi:10.1021/acs.chemmater.2c01608
• 22)   S.  Yang,  M.  Takahashi,  K.  Yamamoto,  K.  Ohara,  T.  Watanabe,  T.  Uchiyama,  T.  Tamaki,  A.  Sakuda,  A.  Hayashi,  M.  Tatsumisago and  Y.  Uchimoto: Studies on the inhibition of lithium dendrite formation in sulfide solid electrolytes doped with LiX (X = Br, I), Solid State Ionics 377 (2022) 115869. doi:10.1016/j.ssi.2022.115869
• 23)   W.J.  Li,  M.  Hirayama,  K.  Suzuki and  R.  Kanno: Fabrication and electrochemical properties of a LiCoO2 and Li10GeP2S12 composite electrode for use in all-solid-state batteries, Solid State Ionics 285 (2016) 136–142. doi:10.1016/j.ssi.2015.05.007
• 24)   A.L.  Santhosha,  L.  Medenbach,  J.R.  Buchheim and  P.  Adelhelm: The Indium–Lithium Electrode in Solid-State Lithium-Ion Batteries: Phase Formation, Redox Potentials, and Interface Stability, Batter. Supercaps 2 (2019) 524–529. doi:10.1002/batt.201800149
• 25)   O.  Kwon,  M.  Hirayama,  K.  Suzuki,  Y.  Kato,  T.  Saito,  M.  Yonemura,  T.  Kamiyama and  R.  Kanno: Synthesis, structure, and conduction mechanism of the lithium superionic conductor Li_10+δGe_1+δP_2−δS₁₂, J. Mater. Chem. A 3 (2015) 438–446. doi:10.1039/C4TA05231E
• 26)   J.T.S.  Irvine,  D.C.  Sinclair and  A.R.  West: Electroceramics: Characterization by Impedance Spectroscopy, Adv. Mater. 2 (1990) 132–138. doi:10.1002/adma.19900020304
• 27)   X.  Wang,  T.  Nishima and  I.  Uchida: Evaluation of the Stability of Organic Solutions and Metals by Using Microelectrode Technique, Electrochemistry 67 (1999) 145–150. doi:10.5796/electrochemistry.67.145
• 28)   D.H.S.  Tan,  E.A.  Wu,  H.  Nguyen,  Z.  Chen,  M.A.T.  Marple,  J.M.  Doux,  X.  Wang,  H.  Yang,  A.  Banejee and  Y.S.  Meng: Elucidating Reversible Electrochemical Redox of Li₆PS₅Cl Solid Electrolyte, ACS Energy Lett. 4 (2019) 2418–2427. doi:10.1021/acsenergylett.9b01693
• 29)   T.  Hakari,  M.  Nagao,  A.  Hayashi and  M.  Tatsumisago: All-solid-state lithium batteries with Li3PS4 glass as active material, J. Power Sources 293 (2015) 721–725. doi:10.1016/j.jpowsour.2015.05.073
• 30)   A.  Ikezawa,  G.  Fukunishi,  T.  Okajima,  F.  Kitamura,  K.  Suzuki,  M.  Hirayama,  R.  Kanno and  H.  Arai: Performance of Li4Ti5O12-based reference electrode for the electrochemical analysis of all-solid-state lithium-ion batteries, Electrochem. Commun. 116 (2020) 106743. doi:10.1016/j.elecom.2020.106743
• 31)   M.S.  Zheng,  Q.F.  Dong,  H.Q.  Cai,  M.G.  Jin,  Z.G.  Lin and  S.G.  Sun: Formation and Influence Factors of Solid Electrolyte Interphase Film on the Negative Electrode Surface in Lithium-Ion Batteries, J. Electrochem. Soc. 152 (2005) A2207. doi:10.1149/1.2060669
• 32)   D.  Guerard and  A.  Herold: Intercalation of lithium into graphite and other carbons, Carbon 13 (1975) 337–345. doi:10.1016/0008-6223(75)90040-8
• 33)   H.  Fujimoto,  M.  Murakami,  T.  Yamanaka,  K.  Shimoda,  H.  Kiuchi,  Z.  Ogumi and  T.  Abe: Synchronized Operando Analysis of Graphite Negative Electrode of Li-Ion Battery, J. Electrochem. Soc. 168 (2021) 080508. doi:10.1149/1945-7111/ac18e4
• 34)   M.  Heß and  P.  Novák: Shrinking annuli mechanism and stage-dependent rate capability of thin-layer graphite electrodes for lithium-ion batteries, Electrochim. Acta 106 (2013) 149–158. doi:10.1016/j.electacta.2013.05.056
• 35)   S.  Schweidler,  L.  de Biasi,  A.  Schiele,  P.  Hartmann,  T.  Brezesinski and  J.  Janek: Volume Changes of Graphite Anodes Revisited: A Combined Operando X-ray Diffraction and In Situ Pressure Analysis Study, J. Phys. Chem. C 122 (2018) 8829–8835. doi:10.1021/acs.jpcc.8b01873
• 36)   A.  Sakuda,  A.  Hayashi and  M.  Tatsumisago: Sulfide Solid Electrolyte with Favorable Mechanical Property for All-Solid-State Lithium Battery, Sci. Rep. 3 (2013) 2261. doi:10.1038/srep02261
• 37)   S.  Taminato et al.: Real-time observations of lithium battery reactions—operando neutron diffraction analysis during practical operation, Sci. Rep. 6 (2016) 28843. doi:10.1038/srep28843
• 38)   H.  Zhang,  Y.  Yang,  D.  Ren,  L.  Wang and  X.  He: Graphite as anode materials: Fundamental mechanism, recent progress and advances, Energy Storage Mater. 36 (2021) 147–170. doi:10.1016/j.ensm.2020.12.027

Related Articles

• This Paper was Originally Published in Japanese in J. Jpn. Soc. Powder Powder Metallurgy 72 (2025) 129–136.