Modified Lithium Borate Buffer Layer for Cathode/Sulfide Electrolyte Interface Stabilization

Article information

Abstract

1. Introduction

The advent of lithium-ion batteries (LIBs) has led to a transformative shift in consumers’ mobile-centric lifestyle. From conducting conversations on high-end smartphones to executing tasks on sophisticated laptops, and more recently, traversing distances in electric vehicles, the pervasive influence of LIBs is unmistakable. Although the technology underlying LIBs has exhibited remarkable strides, particularly in augmenting their energy storage capacity, persistent challenges remain [1–9]. Among these challenges is the inherent safety risk posed by the highly flammable organic liquid electrolytes, which is a critical concern reverberating across the entire LIB industry. Against this backdrop, a notable surge in interest in all-solid-state rechargeable batteries as pioneering solutions has been observed recently [10–14]. Distinguished by the use of a nonflammable solid electrolyte, these batteries circumvent the safety issues associated with traditional LIBs. Moreover, they are promising for achieving a higher energy density and broader operating temperature range. All-solid-state batteries are further enhanced when sulfide-based electrolytes are employed, exhibiting electrochemical properties similar to those of LIB systems [15–21]. This is attributable to the exceptional ionic conductivity and pliability of sulfide-based electrolytes, which foster optimal electrode formation.

However, the constrained electrochemical stability window of sulfide electrolytes poses a challenge to their interface stability, particularly with oxide-based cathodes. The elevated reactivity inherent in sulfide electrolytes aggravates this problem, leading to interdiffusion phenomena (chemical mixing) between the transition metals and oxide ions within the cathode, as well as between sulfur (S) and phosphorus (P) ions in the electrolyte [22–25]. This intricate interplay induces the formation of undesirable side reactant layers at the interface, thereby amplifying the interfacial resistance and compromising both the rate capability and cyclic performance. In this context, stable buffer layers have been introduced to enhance the stability of the interface between the sulfide electrolyte and the cathode. Materials such as LiNbO₃ [26,27], LiTaO₃ [28,29], and Li₂ZrO₃ [30,31], which are known for their considerable Li⁺ ion conductivity and low reactivity with sulfide electrolytes, have been employed to mitigate such interdiffusion phenomena. In addition, their low electronic conductivity contributes to minimizing the decomposition of the sulfide electrolyte attributed to their constrained electrochemical stability window. However, the high cost of alkoxide-based source materials that are required for the formation of such buffer layers can hinder their widespread applicability. In contrast, lithium borate has been recently reported as a cost-effective alternative buffer layer between the cathode and sulfide electrolytes [32,33]. Lithium borate can be manufactured at a relatively low cost using affordable boric acid as a source material. It is particularly advantageous for mass production because of its applicability in dry processes.

In this study, we introduce an innovative approach to further enhance the effectiveness of the interfacial stabilization achieved by lithium borate. Despite the efficacy of lithium borate in managing interfacial side reactions, it shares the characteristic of being an oxide, resulting in an inherent difference in Li⁺ chemical potential compared to sulfide electrolytes. This difference leads to an uneven distribution of Li⁺ ions at the interface, impeding the migration of Li ions during charge and discharge cycles [17,18]. To address this issue, the surface of the lithium borate buffer layer was subjected to sulfur modification through a gas reaction involving sulfur. The sulfur-modified lithium borate is expected to reduce the chemical potential difference of Li⁺ owing to its chemical similarity to the sulfide electrolyte, thereby enhancing the electrochemical properties. To validate this hypothesis, a lithium borate-buffer layer was formed on the surface of a high-Ni cathode via a dry process, followed by sulfurization, and its properties were meticulously analyzed. The effects of the modified lithium borate buffer layer were systematically examined and compared via various techniques such as time-of-flight secondary ion mass spectrometry (TOF-SIMS), X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM).

2. Experimental

In this study, the precursor Ni_0.92Co_0.04Mn_0.04(OH)₂ provided by Ecopro BM was used as the starting material. The pristine Li(Ni_0.92Co_0.04Mn_0.04)O₂ powder was obtained by mixing the precursor with LiOH at a molar ratio of 1:1.03, followed by thorough homogenization. The cathode structure was then formed via heat treatment at 500°C for 5 h and 750°C for 10 h under an oxygen flow rate of 2°C/min. A lithium borate buffer layer was prepared on the cathode surface using a dry method with boric acid (≥ 99.5%, Aldrich) as the source material. The boric acid was ground using a mixer mill (MM400, Retsch) for 5 min at 30 Hz and filtered through a 400 mesh sieve to extract particles smaller than 38 μm. The amount of boric acid was controlled at 2000, 2200, 2400, and 2600 ppm based on the weight of the pristine powder; the boric acid was thoroughly mixed with the pristine powder. Heat treatment was conducted at 300°C for 5 h under an oxygen flow. For further sulfur modification, sulfur powder (99.998%, Sigma–Aldrich) was used as the source material. The lithium borate-coated samples were mixed with sulfur at concentrations of 500, 1000, and 1500 ppm, followed by heat treatment at 300°C for 2 h in a closed tube under an Ar atmosphere. To control the particle size of the coating sources, boric acid and sulfur particles were uniformly ground using a mixer grinder at 30 Hz for 3 min and then sieved through a 400-mesh sieve. TEM images were obtained (JEOL JEM-2100F) to examine the surface morphologies of the pristine and surface-modified samples. TOFSIMS (TOF-SIMS-5, Bi¹⁺) analysis was also performed to observe changes in the surface chemical species upon surface modification.

The electrochemical performance of the pristine and surface-modified cathodes was evaluated using a pressurized cell. Composite electrodes comprising the cathode powder, sulfide solid electrolyte (Li6PS5Cl, POSCO JK), and conductive material (Super P) were mixed in a mass ratio of 62:37:1. The cell assembly process involved pressing the solid electrolyte (0.15 g) at 12 MPa, followed by the addition of the cathode composite (0.015 g), positioning of the Al foil (current collector), and pressing at 44 MPa. The Li-metal foil was placed on the other side, and the mold was sealed under a pressure of 50 kgf·cm. Electrochemical tests were conducted at a constant current range of 2.52 to 4.32 V (vs. Li) and 30°C. Electrochemical impedance spectroscopy (EIS) analysis was performed using an electrochemical workstation (AMETEK VersaSTAT 3) over a frequency range of 0.01 Hz to 500 kHz, with an AC voltage amplitude of 10 mV. Nyquist plot fitting was performed using ZsimpWin 3.60 software. The galvanostatic intermittent titration technique (GITT) was employed to assess the Li⁺-ion-diffusion characteristics post charge–discharge cycling, employing a pulse current of 0.1 C intermittently for 10 min. The Li⁺ diffusion coefficient (D^Li+) was calculated using Eq. (1).

where τ represents the pulse duration (600 s), n_m denotes the molecular weight of the active material, V_m signifies the molar volume of the active material, S denotes the contact area between the solid electrolyte and the active material, ΔE_t represents the transient voltage change, and ΔE_s represents the steadystate voltage change. XPS was used to investigate the cathode/electrolyte interface of the composite cathodes after 300 charge–discharge cycles. The samples were transported via vacuum transfer to minimize unwanted reactions.

3. Results and Discussion

The pristine and lithium borate-coated samples were characterized and their initial charge and discharge profiles are presented in Fig. 1a at a current density of 10 mA·g⁻¹. The coating amount was controlled by adjusting the amount of the borate source (boric acid). The boric acid was designed to react with the lithium residues on the surface (e.g., Li₂CO₃ and LiOH) to form a lithium-borate-based coating layer on heat treatment. While the pristine sample exhibited a discharge capacity of ~195 mAh·g⁻¹, the discharge capacities of the samples coated with 2200 and 2400 ppm boric acid (marked as B 2200 ppm and B 2400 ppm, respectively) were slightly higher (~202 and ~196 mAh·g⁻¹, respectively). Conversely, coating conditions using 2000 and 2600 ppm boric acid (marked as B 2000 ppm and B 2600 ppm, respectively) resulted in a decrease in discharge capacity, indicating insufficient or excessive coating layer formation. Fig. 1b and Table 1 present the discharge capacities at current densities of 10, 20, 60, 100, and 200 mA·g⁻¹. The pristine sample exhibited a significant decline in discharge capacity, reaching 135 and 117 mAh·g⁻¹ at current densities of 100 and 200 mA·g⁻¹, respectively. The capacity retention, defined as the percentage of capacity at 200 mA·g⁻¹ relative to that at 10 mA·g⁻¹, was approximately 60%. In contrast, the lithium borate-coated sample exhibited significantly improved discharge capacity at high current densities. All lithium borate-coated samples exhibited slightly elevated discharge capacities at 200 mA·g⁻¹, with their capacity retention ranging from 63% to 65% (Table 1). This improvement in their electrochemical properties can be ascribed to the ability of the coating layer to mitigate side reactions between the cathode and sulfide electrolyte, as corroborated by previous studies [32,33]. Among the experimental conditions, the application of 2200 ppm boric acid was deemed the most optimal, affording the highest discharge capacity and favorable Coulombic efficiency. Subsequently, the lithium borate-coated sample with 2200 ppm boric acid was designated as B-sample and selected for further sulfur modification.

Fig. 1

(a) Initial charge and discharge profiles of the pristine and lithium borate-coated samples at a current density of 10 mA·g⁻¹. (b) Discharge capacities of the pristine and lithium borate-coated samples at current densities of 10, 20, 60, 100, and 200 mA·g⁻¹. TEM images of (c) the pristine and (d) B-sample.

Table 1

Discharge capacities at various current densities, coulombic efficiency, and capacity retention of the pristine and lithium borate-coated samples

Fig. 1c,d present the TEM images of the pristine and B-samples. The pristine sample was examined after washing to eliminate any lithium residues from the surface, resulting in a clean surface. In contrast, despite the same washing treatment, a coating layer of 2–3 nm was distinctly observed on the surface of the B-sample, which was presumed to be a lithium borate layer formed through the coating process. The coating layer was relatively thin and uniformly distributed, suggesting its potential to protect the cathode surface.

Fig. 2a presents a comparison of the discharge capacities of the pristine and B-sample with those of cathodes coated with sulfur-modified lithium borate, measured at current densities of 10, 20, 60, 100, and 200 mA·g⁻¹. To form a coating layer with sulfur-modified lithium borate, B-sample was additionally treated with 500, 1000, and 1500 ppm sulfur. The initial discharge capacity of B-sample measured at 10 mA·g⁻¹ was almost unchanged or even slightly reduced after sulfur treatment. However, as the current density increased, modification with 1000 and 1500 ppm of sulfur resulted in a relatively small capacity decrease compared to that of B-sample. Whereas the capacity retention of B-sample at 200 mA·g⁻¹ was ~64%, those of the samples coated with sulfur-modified lithium borate (1000 and 1500 ppm sulfur) increased to ~67%. Such a finding indicates that the rate capability of B-sample can be further improved by additional sulfur modification. Considering both the discharge capacity and capacity retention, the conditions for sulfur modification using 1000 ppm sulfur appeared to be optimal. Hereafter, we refer to the sulfur-modified lithium borate-coated cathode (with 1000 ppm sulfur) as SB-sample. In the TEM image of SB-sample (Fig. 2b), a coating layer with a thickness similar to that of B-sample is observed; no significant changes are detected in the coating layer owing to sulfur modification. Fig. 2c,d show a comparison of the initial discharge and rate capabilities of the pristine, B-, and SB-samples. While B-sample exhibited a superior discharge capacity and rate capability compared to the pristine sample, more enhancements were observed with sulfur modification in SB-sample. Table 2 summarizes the discharge capacities, Coulombic efficiencies, and capacity retentions of the samples.

Fig. 2

(a) Discharge capacities of the pristine, B-sample, and sulfur-modified lithium borate-coated samples at current densities of 10, 20, 60, 100, and 200 mA·g⁻¹. (b) TEM images of SB-sample. (c) Comparison of the initial charge and discharge profiles of the pristine, B-, and SB-samples. (d) Comparison of the rate capability of the pristine, B-, and SB-samples.

Table 2

Discharge capacities at various current densities. Coulombic efficiency and capacity retention of the pristine, B-sample, and sulfur-modified lithium borate-coated samples

The surfaces of the three samples (pristine, B-, and SB-sample) were examined using TOF-SIMS to analyze the chemical species resulting from surface modification. Fig. 3 shows the TOF-SIMS depth profiles of the samples measured by sputtering using a Cs⁺ ion gun. In the profile of the pristine sample (Fig. 3a), only NiO⁻ and SO⁻ were detected in significant amounts, with the quantities of the other species being negligible. The presence of SO⁻ is attributed to the sulfate sources used in the cathode precursor preparation. In the profile of B-sample (Fig. 3b), the quantity of BO⁻ notably increased, particularly at the surface, which is indicative of lithium borate layer formation. The profile of SB-sample exhibits a considerable increase in LiS⁻, SO⁻, and BSO⁻ (Fig. 3c). In Fig. 3d–f, the intensities of LiS⁻, SO⁻, and BSO⁻ in the three samples are compared, suggesting that the sulfur introduced during the modification process reacted with the lithium residues on the surface and with oxygen to form LiS_x and SO_x, respectively. In addition, it partially reacted with lithium borate to produce BSO_x. These reactants may also be formed by the reaction of the cathode with a sulfide electrolyte. Pre-generating these reactants during sulfur modification prevents further reactions between the sulfide electrolyte and cathode. Furthermore, the Li⁺ chemical potential of these sulfur-related reactants is expected to be closer to that of the sulfide electrolyte than to those of the oxide cathode and lithium borate. This closer proximity may facilitate a more uniform distribution of Li⁺ ions, thereby aiding Li⁺ ion migration during charge and discharge cycles.

Fig. 3

TOF-SIMS depth profiles of the (a) pristine, (b) B-sample, and (c)SB-sample. Comparison of the TOF-SIMS depth profiles of the (d) LiS⁻, (e) SO⁻, and (f) BSO⁻ species.

Fig. 4a shows the cyclic performances of the pristine, B-, and SB-samples. Under a current density of 600 mA·g⁻¹, the pristine sample experienced a substantial capacity reduction (31%) over 300 cycles. In contrast, B-sample exhibited a relatively modest capacity reduction of approximately 17%, demonstrating the efficacy of the lithium borate coating. Remarkably, SB-sample featuring sulfur modification, demonstrated a high capacity at a capacity loss of a mere 12% over 300 cycles. Such a finding underscores the synergistic effect of sulfur modification of the lithium borate coating layer, enhancing the performance and bolstering the cyclic stability. The observed improvement is attributed to sulfur modification, which facilitates Li⁺ migration and suppresses side reactions arising from chemical mixing. To further substantiate this hypothesis, impedance, GITT, and XPS analyses were performed.

Fig. 4

(a) Cyclic performance comparison of the pristine, B-, and SB-samples. Nyquist plot of cells containing the (b) pristine, (c) B-, and (d) SB-samples obtained after 1 and 300 cycles. (e) Equivalent circuit and fitting model of the Nyquist plot (example).

Impedance analysis can be used to identify the interfacial resistance components due to surface modification. In Fig. 4b–d, Nyquist plots depict the measurements after the 1st and 300th cycles. The semicircle in the Nyquist plot signifies the cell resistance components. Notably, B-and SB-samples exhibited lower resistance compared to the pristine sample. For further examination, the resistance components observed in the Nyquist plots were fitted using the equivalent circuit illustrated in Fig. 4e. The Nyquist plot of an all-solid-state cell can be fitted to the bulk resistance (R_b), grain boundary resistance of the sulfide electrolyte (R_g), and resistance at the cathode/solid electrolyte (R_C/E) and anode/solid electrolyte (R_A/E) interfaces [34,35]. The obtained impedance values are presented in Table 3. As anticipated, the R_C/E value was notably influenced by the surface modification of the cathodes, exhibiting a distinct disparity compared to the other impedance values. The pristine sample exhibited an R_C/E value of approximately ~42 Ω after the 1st cycle; this value increased significantly to ~89 Ω after 300 cycles. Such a finding underscores the rapid increase in the resistance component linked to the cathode/sulfide electrolyte interface due to side reactions such as electrolyte decomposition and chemical mixing between oxygen and transition metals in the cathode and S and P in the electrolyte during charge and discharge. In contrast, B-sample exhibited a substantial reduction in the R_C/E value to ~8 Ω after the 1st cycle and ~17 Ω after the 300th cycle, indicating the efficient reduction of interfacial resistance through lithium borate coating. Furthermore, the sulfur-modified SB-sample displayed even lower R_C/E values after the 1^st cycle (~2.4 Ω), maintaining low values (< 10 Ω) even after 300 cycles. This confirms that the sulfur modification enhances Li⁺ migration, which is attributed to the improved electrochemical performance of SB-sample (Fig. 2d and 4a).

Table 3

Impedance values of the pristine, B-, and SB-samples calculated from the Nyquist plots

Fig. 5 presents the GITT curves obtained after 300 cycles, enabling the estimation of the Li diffusion coefficient ( ) values for the pristine and surface-modified cathodes within the voltage range of 3.65–4.2 V. This analysis aimed to explore the impact of lithium borate coating and sulfur modification on Li⁺ diffusion. Both B-and SB-samples demonstrated higher D^Li+ values compared to the pristine sample, suggesting that surface modification facilitates Li⁺ ion diffusion (Fig. 5). Notably, in the high-voltage range (4.0–4.2 V), the D^Li+ values of SB-sample significantly exceed those of B-sample. This observation confirms that sulfur-modified lithium borate layers are more efficient in enhancing the Li⁺ ion mobility compared to lithium borate layers alone.

Fig. 5

Li-diffusion coefficients (D^Li+) of the pristine, B-, and SB-samples estimated from GITT curves measured after 300 cycles.

To directly validate the efficacy of the sulfur-modified lithium borate coating on side reaction suppression, XPS analysis of the composite electrode (cathode + sulfide electrolyte) obtained from the cell before and after 300 cycles was conducted. The XPS spectra of the composite electrode after cycling revealed the side reactants generated during the charging and discharging processes. In Fig. 6a, the S 2p spectrum of the pristine sample before cycling exhibits large peaks at 161.6 and 162.7 eV (marked in yellow), assigned to the PS₄³⁻ unit of the argyrodite structure of the electrolyte [36,37]. Peaks at 160.3 and 161.4 eV (marked in red) are indicative of transition metal sulfides or lithium sulfide resulting from the chemical mixing between the cathode and sulfide electrolyte [38–40]. The green peaks at 163.2 and 164.3 eV and the blue peaks at 163.7 and 164.8 eV correspond to the P–[S]n–P and S–S bonds, respectively, generated from electrolyte decomposition [41,42]. Although the red peaks in Fig. 6a have a low intensity, they significantly intensified after 300 cycles (Fig. 6b). This indicates that chemical mixing progressed during cycling, leading to an increase in the number of side reactants. The accumulation of these side reactants causes capacity reduction and rate capability degradation during cycling. However, the intensities of these peaks were notably reduced in the S 2p spectra of B-and SB-samples after 300 cycles (Fig. 6c,d). Such a result confirms the successful suppression of the surface modification of the side reactions.

Fig. 6

S 2p XPS spectra of the (a) pristine sample before the test, (b) pristine sample after 300 cycles, (c) B-sample after 300 cycles, and (d) SB-sample after 300 cycles.

Fig. 7 shows the P 2p spectra of the samples before and after cycling. The P 2p spectrum of the pristine sample (Fig. 7a) exhibits two main peaks at 131.9 and 132.7 eV (marked in yellow), corresponding to the PS₄³⁻ bonds of the electrolyte [27,43], and several side peaks. The red peaks at 130.7 and 131.5 eV represent reduced phosphates due to electrolyte decomposition [44]. The blue peaks at 132.8 and 133.6 eV are associated with P₂S_x and Li₃PO₄, respectively, whereas the green peaks at 133.7 and 134.6 eV signify the presence of transition-metal phosphates, indicating chemical mixing at the interface due to side reactions [37,39]. After 300 cycles, the intensities of the red, blue, and green peaks in the P 2p spectra increased owing to the progress of the interfacial side reactions (Fig. 7b). However, as shown in Fig. 7c,d, the B-and SB-samples with a surface coating exhibited reduced intensity of these peaks compared to the pristine sample. Furthermore, considering the lower peak intensity of SB-sample compared to B-sample, sulfur modification seems more efficient in mitigating interfacial side reactions. This is likely due to the synergistic effect between the protective effect of lithium borate and its improved compatibility with the sulfide electrolyte through sulfur modification. The structure of the sulfur-modified lithium borate formed on the surface cannot be precisely determined, therefore, the Li⁺ chemical potential cannot be directly measured; however, we can indirectly observe that side reactions, such as chemical mixing, are effectively controlled. The enhanced rate capability and cyclic performance observed after sulfur modification can be attributed to the suppression of side reactions and enhanced Li⁺ migration (accompanied by a decrease in the interfacial resistance). Fig. 8 presents the effects of sulfur-modified lithium borate coating.

Fig. 7

P 2p XPS spectra of the (a) pristine sample before the test, (b) pristine sample after 300 cycles, (c) B-sample after 300 cycles, and (d) SB-sample after 300 cycles.

Fig. 8

Schematic of the effects of sulfur-modified lithium borate coating.

4. Conclusions

This study focused on augmenting the interfacial stability in all-solid-state batteries by introducing a lithium borate coating layer on the cathode, which was subsequently refined through sulfur modification. Despite its inherent cost-effectiveness, lithium borate possesses oxide characteristics, leading to a discernible difference in the chemical potential of Li⁺ compared with sulfide electrolytes. This discrepancy induces an uneven distribution of Li ions at the cathode/electrolyte interface, impeding the smooth migration of Li⁺ ions during both charge and discharge cycles. In response to this challenge, a pioneering approach was adopted, wherein the lithium borate buffer layer underwent sulfur modification through a gas reaction involving sulfur. Characterization of the samples revealed notable enhancement of the electrochemical properties of the lithium-borate-coated samples. Moreover, sulfur modification of the cathode further augmented the rate capability and cyclic performance of all-solid-state cells. The reduced interfacial resistance identified in the impedance analysis, coupled with the diminished undesirable reactants verified through XPS analysis, underscore the efficacy of the lithium borate coating. This effect was further amplified by sulfur modification, confirming that sulfur modification enhanced the capability of lithium borate for controlling side-reactions. This heightened efficacy is ascribed to the role of sulfur modification in reducing the chemical potential difference of Li⁺ between the cathode and sulfide electrolyte, aligning with the anticipated benefits outlined earlier. The innovative approach presented in this study not only offers valuable insights into the realm of all-solid-state batteries but also holds significant promise for steering future research endeavors in this burgeoning field.

Article information Continued