Fundamental studies on behaviors in Li | |SeS2 coin cellsThe discharge/charge reaction mechanisms of S, Se, and SeS2 in the chalcogen family were investigated through cyclic voltammetry (CV) measurements and analyses, as depicted in Fig. 1a. During the discharge process for all three chalcogen positive electrode active materials, two prominent cathodic peaks were observed. The first cathodic peak corresponds to the reduction of S or Se to high-order lithium polysulfides (Li2Sx, 4 ≤ x ≤ 8) or lithium polyselenides (Li2Sex, 4 ≤ x ≤ 8), while the lower plateau is attributed to their further reduction to the final products, namely lithium sulfide (Li2S) or lithium selenide (Li2Se). However, in the case of SeS2, an additional minor cathodic peak appears at 2.1 V, resembling the position of Se’s first cathodic peak. This observation suggests that reduction to lithium polysulfides (LiPS) and lithium polyselenides (LiPSe) occurs in order, and finally precipitate to Li2S and Li2Se at a second plateau region. When examining the anodic scans, anodic peaks emerge, and they are associated to the Li2S and/or Li2Se final products returning to their original states. The position of the first peak in the anodic scan depends on the reactivity of the final products, which causes the first peak to shift to a higher potential region. While the anodic peaks of S and Se overlap due to multiple reactions occurring simultaneously, SeS2 exhibits two distinct anodic peaks. This phenomenon arises from the varying reaction kinetics of the final products. Specifically, Li2Se has an energy bandgap of 2.997 eV, while Li2S is 3.297 eV27,28,29. Generally, it is observed that as the energy bandgap decreases, electrical conductivity tends to increase, therefore, Li2Se displays faster reaction kinetics than Li2S due to its higher electrical conductivity. As a result, the fully discharged SeS2 electrode consists of a mixture of Li2Se and Li2S, with Li2Se exhibiting slower reactivity compared to a pure Se electrode, while Li2S reacts more quickly than a pure S electrode. The presence of these two anodic peaks is attributed to the combination of different reaction kinetics between Li2S and Li2Se.Fig. 1: Electrochemical testing of Li | |S, Li | |Se and Li | |SeS2 cells.a Cyclic voltammograms at a scan rate of 0.1 mV s−1. b Voltage profiles of Li | |S cell at a specific current = 335 mA g−1; Li | |Se cell at a specific current = 135 mA g−1; Li | |SeS2 cell at a specific current = 240 mA g−1 for the first cycle. c Rate performance at different specific currents. d Long-term cycling of Li | |S, Li | |Se and Li | |SeS2 coin cells tested at various specific currents (positive electrode active material mass loading = 2 mg cm−2; ratio of electrolyte to active materials, E/A = 5 μL mg−1; Li | |S current rate = 1675 mA g−1; Li | |Se current rate = 675 mA g−1; Li | |SeS2 current rate = 1200 mA g−1). The electrochemical testing was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode.Figure 1b compared the voltage profiles in the first cycle for various non-aqueous Li | |chalcogen cells with three different chalcogen-based positive electrodes, revealing differences in the 1st cycle specific discharge capacity (Li | |S cell cycled at 335 mA g−1 delivers 1078 mAh g−1, Li | |Se cell cycled at 135 mA g−1 delivers 458 mAh g−1 and, Li | |SeS2 cell cycled at 240 mA g−1 delivers 817 mAh g−1) in line with their respective theoretical capacities. These profiles also align with the CV peaks observed in Fig. 1a. Notably, SeS2 exhibits two additional plateaus at both high and low voltage regions, indicating differing reaction timings between S and Se. This suggests that SeS2 may possess a combination of the advantages of S and Se, offering improved specific discharge capacity due to enhanced electrical conductivity.From this perspective, electrochemical analysis of the Li | |chalcogen cells containing the three different S-based positive electrodes were performed, as shown in Fig. 1c, d. In rate performance, we standardized the specific current based on the current per unit mass of active material at the positive electrodes. The loading of active materials was maintained at 2 mg cm−2, and the ratio of the electrolyte to active material (E/A) was set at 5 μL mg−1. In Fig. 1c, rate performance evaluations were conducted for the Li | |chalcogen cells with the three different S-based positive electrodes at various specific currents, ranging from 0.1 to 1.0 A g−1. Notably, the Li | |S coin cell exhibited reversible capacities of 1132, 1040, and 1013 mAh g−1 at rates of 0.1, 0.2, and 0.3 A g−1, respectively, surpassing the Li | |SeS2 coin cell (852, 750, and 678 mAh g−1) and the Li | |Se coin cell (382, 262, and 182 mAh g−1) in these conditions. However, as the specific currents increased to 0.4 to 1.0 A g−1, reversible capacities of the Li | |S coin cell dropped significantly to 304, 209, and 145 mAh g−1, while the Li | |SeS2 coin cell maintained higher values of 566, 508, and 400 mAh g−1, indicating improved rate performance. In Fig. 1d, we compared the cycle performance at the same C-rate of 1 C for 100 cycles, corresponding to 1675 mA g−1 for Li | |S coin cell, 1200 mA g−1 for Li | |SeS2 coin cell, and 675 mA g−1 for Li | |Se coin cell. We chose this approach over using a fixed specific current because the theoretical capacities of the materials being compared differ, using specific current results in each electrode reacting at a different relative rate. Therefore, the approach using the C-rate rather than specific current offers a more intuitive comparison of the cycle performances among various active materials, ensuring uniform discharge/charge duration. Pre-activation involved 2 cycles at a specific current of 0.1 C-rate to ensure uniform active material distribution and the pre-activation cycle data and its corresponding voltage profiles are shown in Supplementary Fig. 1. Li | |SeS2 coin cell exhibited a significantly higher reversible capacity retention of 70% compared to Li | |Se coin cell’s 7.7% retention and Li | |S coin cell’s reversible capacity of only 210 mAh g−1. It is important to highlight that the rate capability and cyclability tests were performed under relatively low electrolyte to active material (E/A) ratio conditions, using pure S, Se, and SeS2 powders as active materials. Thus, although the battery performances at high specific current are limited, the proof of concept demonstrated in this study could lay the foundation for future work in terms of positive electrode active material engineering improvement.To unravel the dissolution and precipitation mechanism of SeS2, a customized coin cell was designed for operando XRD measurements, with detailed information provided in the method section. Figure 2a shows a voltage profile of a Li | |SeS2 cell subjected to a constant current of 120 mA g−1, with a cell voltage in the 1.7– 2.8 V range during the first cycle, accompanied by the corresponding operando XRD patterns. In the pristine state, the XRD pattern exhibited discernible peaks for SeS2 (JCPDS No. 047-1611) at 20.4°, 23.8°, 24.2°, 25.3°, 26.8°, 27.7°, 28.8°, and 29.5°, as indexed in orange and depicted in Fig. 2b. As the discharge (lithiation) process commenced, the characteristic peaks associated with SeS2 is barely detectable, with an exception being the peak at 23.8°. In the early stage of discharging, this peak shifted slightly to a lower angle before transitioning to a higher angle. Eventually, only the specific peaks attributed to Se remained. This phenomenon suggests that in the ring structured SeS2 composed of both S and Se, S dissolves first in the initial discharge stage. The relatively larger atomic radius of Se than S leads to a larger d-spacing, resulting in the lower-angle shift observed. Once S is completely dissolved, Se takes over, leading to the disappearance of the two characteristic Se peaks. Consequently, Li2S (JCPDS No. 023-0369) is initially generated, followed by the formation of Li2Se (JCPDS No. 023-0072), as marked in green.Fig. 2: Operando XRD measurements of Li | |SeS2 cell.a XRD patterns recorded during the first cycle at a specific current of 120 mA g−1 and the corresponding galvanostatic curve. b An enlarged XRD patterns of the area marked in blue, black and red in representing the latter part of charging, initial stage for Se nucleation, and the early part of discharging, respectively. This measurement was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode (positive electrode active material mass loading = 5 mg cm−2; ratio of electrolyte to active materials, E/A = 20 μL mg−1).Conversely, in the charging (delithiation) process, the Li2Se peak gradually diminishes initially, as shown in Fig. 2b. This is followed by the nucleation of Se (JCPDS No. 006-0362), as indicated in black. As the charging progresses post-Se nucleation, it becomes evident that the reaction mechanism toward a crystal structure distinct from the initial state of SeS2. With the progression of Se nucleation, the emergence of beta phase S (JCPDS No. 071-0137), highlighted in yellow, becomes apparent. By the end of the charging state, a mixed crystal structure comprising Se and beta phase S is observed. Thus, this investigation demonstrates that SeS2 does not revert to its original state after the first cycle but instead crystallizes, undergoing separation into Se and beta phase S.Understanding the nucleation and growth mechanisms of SeS2-based positive electrodes during Li | |SeS2 cell chargingTo visualize the formation of Se and S crystals within the SeS2 positive electrode, we employed operando transmission X-ray microscopy (TXM)25,26,30, as shown in Fig. 3, Supplementary Fig. 2 and Supplementary Movie 1. In Fig. 3a, we present the cell voltage profile with specific state of charge (SOC) points, each corresponding to the SOC where Se (highlighted in gray) and S (marked in orange) nucleate, as determined from operando XRD pattern analysis. We denoted these SOC points as ‘S digits’ with, for example, S80 indicating an 80% state of charge. Figure 3b, extracted from Supplementary Movie 1 at S100, provides a visual representation of Se and S particles with high intensities (yellow), while the carbon current collector appears as relatively low-intensity background features in blue. To generalize this phenomenon, we divided the field of view into three regions, as depicted in Fig. 3b. These three regions were magnified and tracked in Fig. 3c, corresponding to specific SOC points, to monitor the growth of Se and S crystals. Using XRD results as indicators of crystal formation points, TXM images were categorized into two groups: from S67 to S80 and from S80 to S100. Further detailed analysis and discussion on distinguishing those crystals can be found in Supplementary Fig. 3 and Supplementary Note 1. The nucleation of Se seed is indicated by sky-blue arrows, commencing around S67 to S71, and these Se particles develop into dendritic structures until S80. As S particle formation begins, we traced the growth of S particles with red arrows, noting that S tended to grow on the surfaces of the already-formed Se crystals. A comparative analysis of TXM images at SOC points S77 and S100 was conducted, as depicted in Supplementary Fig. 3. Marked by red overlays on the grayscale TXM image at S100, these comparisons reveal distinct features, indicating the distribution of S particles around Se crystals. This implies LiPS oxidation on these crystals during the charging process. To further support the notion of these Se nuclei as seeds for S, additional scanning electron microscopy (SEM) analysis was also conducted, as shown in Supplementary Figs. 4, 5 and Supplementary Note 2. The formation of S is more evident in Supplementary Movie 1 than in the figures. The average area of these Se/S crystals was approximately 6700 μm2, calculated using ImageJ software. Moreover, we conducted ex situ SEM imaging in backscattered electron (BSE) mode to reveal the detailed morphology of dendritic Se/S crystals as shown in Fig. 3d. Energy-dispersive X-ray spectroscopy (EDS) analysis of the red box in the SEM image confirmed the presence of S particles (illustrated in yellow) on the surface of Se crystals (displayed in cyan). To ascertain a more accurate Se and S ratio, high-magnification SEM and EDS experiments were conducted, and the results can be observed in Supplementary Fig. 6. As described in Fig. 3e, this result provides experimental proof on the reaction mechanism of non-aqueous Li | |SeS2 cell configuration.Fig. 3: Operando TXM measurements of Li | |SeS2 cell.a Voltage profile of the SeS2 electrode in charging at a specific current of 600 mA g−1 for the first cycle b TXM image of SeS2 electrode at a state of S100. (Color bar is an index of detected x-ray radiation determined by material’s density). Real-time behavior of the SeS2 electrode in operando TXM observation can be identified in Supplementary Movie 1. c Operando TXM images of the SeS2 electrode at three different enlarged regions (I, II, and III) at various SOCs. d Ex situ SEM image in a BSE mode of the SeS2 electrode at a fully charged state and the EDS images of the red box in the SEM image corresponding to the Se and S atom. e Schematic illustration of the growth mechanism in the lithium | |selenium sulfides cell. This measurement was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode (positive electrode active material mass loading = 2 mg cm−2).The DFT calculation results could provide the theoretical support on the reason why S forms preferentially on the surface of pre-formed Se crystals rather than on the carbon current collector. Figure 4 presents first-principles calculations assessing the adsorption energy of LiPS on both graphene layer and Se (100) surfaces. As shown in Fig. 4a, the calculated adsorption energies for LiPS on the carbon surface were −0.58 eV (Li2S), −0.63 eV (Li2S2), −0.53 eV (Li2S4), −0.52 eV (Li2S6) and −0.55 eV (Li2S8). Conversely, on the Se slab, the computed adsorption energies for LiPS were as follows: −2.63 eV for Li2S, −1.50 eV for Li2S2, −1.06 eV for Li2S4, − 1.08 eV for Li2S6, and −1.10 eV for Li2S8, as shown in Fig. 4b. These values highlight the stronger adsorption of LiPS on the Se (100) surface compared to the graphene layer, in line with experimental data, suggesting a preference for LiPS adsorption on Se surface. The transition from Li2S2 to Li2S is identified as a rate-determining step due to its substantial activation energy requirement31,32,33. Our computational analysis reveals a distinction in the binding energies associated with Li2S2 on different substrates: −0.63 eV for graphene compared to −1.50 eV for a Se surface. This difference suggests a stronger affinity of Li2S2 for Se, implying a favorable interaction that can influence the reaction pathway. Additionally, the energy barrier for dissociation (ΔE) of Li2S2 on graphene is computed at 4.23 eV, notably higher than the 0.36 eV observed for the Se surface (Supplementary Fig. 7). The lower energy barrier on Se indicates a facilitated solid-phase conversion process, essential for efficient reaction kinetics. In conclusion, our DFT-based analysis supports the premise that Se acts as catalyst within the Li | |S-based battery system, promoting the critical Li2S2 to Li2S phase transformation by providing lower energy pathways and stronger adsorptive interactions.Fig. 4: Results of first principle calculations.a Atomic configurations and adsorption energies of Li-polysulfides on graphene layer. Li, S, and C atoms are represented by the colors blue, yellow, and brown, respectively. b Atomic configurations and adsorption energies of Li-polysulfides on Se (100) surface. Li, S, and Se atoms are represented by the colors blue, yellow, and brown, respectively.Feasibility of expanding toward blended SexSy
Through XRD analysis, it was demonstrated that the SeS2 active materials eventually undergo reduction into Li2Se and Li2S, and each species subsequently re-oxidizes to form Se and S crystals during the initial charge/discharge cycle of the Li | |SeS2 cell. Notably, due to the preferential absorption of S over Se, S growth occurred in conjunction with the Se nuclei. At this juncture, we focused our attention toward determining if comparable results could be achieved by formulating positive electrodes through the combination of Se and S in an equimolar ratio, similar to that of SeS2.Figure 5a illustrates that both the voltage profiles of the Li | |S coin cells with the SeS2-based and the blend-based positive electrodes, comprising S and Se, exhibited similarities in terms of voltage plateaus, specific capacity, and voltage profiles. As shown in Fig. 5b, c, operando XRD analysis confirmed that the XRD patterns before the discharging differed, but after initial cycle, both patterns became identical, displaying peaks corresponding to Se and beta phase S crystals.Fig. 5: Operando XRD and TXM measurements of Li | |Se:S (1:2) blended cell.a Voltage profiles of the Li | |S cells with the SeS2 and Se:S (1:2) blended electrodes at a specific current of 0.2 A g−1 for the first charge/discharge cycle of the cell. Operando XRD patterns of two electrodes at first cycle in early discharge and final charge states are marked in red and blue each for b, c representing SeS2 and Se:S (1:2) blended electrodes, respectively. This measurement was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode (positive electrode active material mass loading = 5 mg cm−2; ratio of electrolyte to active materials, E/A = 20 μL mg−1). TXM images of Se:S (1:2) blended electrode at a specific current of 600 mA g−1 d from the first nucleation of Se to S79 and e from the first growth of S to S100 of the enlarged area of the yellow box in S79 (scale bar = 50 μm). The corresponding operando TXM movie can be seen in Supplementary Movie 2. f Schematic illustration for SeS2 and Se:S (1:2) blended electrodes for the first cycle. This measurement was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode (positive electrode active material mass loading = 2 mg cm−2).Moreover, we conducted TXM analysis as shown in Fig. 5, Supplementary Fig. 8 and Supplementary Movie 2. TXM images revealed the initial growth of Se seeds, followed by the concurrent growth of S particles alongside the Se nuclei, as depicted in Fig. 5d, e. As shown in Supplementary Fig. 9, a comparative analysis of TXM images was also conducted at SOC points S79 and S100, showing the distribution of S particles around Se crystals. Additionally, we conducted high-magnification ex situ SEM and EDS experiments to confirm the presence of S particles (illustrated in yellow) on the surface of Se crystals (displayed in cyan) as illustrated in Supplementary Fig. 10. Therefore, both the SeS2-based and the blend-based positive electrodes exhibited the same crystal formation mechanism after the initial cell discharging, as illustrated in Fig. 5f. However, the initial state of Se peaks in the Se:S (1:2) blended electrode, which appears weaker and broader compared to those in the fully charged state, is not observed in the SeS2 electrode. These differences in crystallinity between the fully charged state and the initial pristine state of Se, particularly the weaker and broader Se peak around 30 ° in the pristine state, result from the ball-milling process. We analyzed the full width at half maximum (FWHM) values in the XRD data of Se powder before and after ball-milling to evaluate the crystallinity changes during sample preparation. The comparison, presented in Supplementary Fig. 11, demonstrates enlarged XRD peaks at 23.5° and 29.7°, corresponding to the (100) and (101) planes, respectively. The FWMH values for these planes in pristine Se powder were 0.450 and 0.370, respectively, which increased to 0.560 and 0.462 after ball-milling, indicating a slight decrement in crystallinity.Based on these observations, we proposed two cases in Fig. 6a. In Case I, a positive electrode with a high Se ratio in the mixture results in the generation of a significant amount of LiPSe during discharging. During recharging process, once the Se nucleus is formed, there is a preference for growth on pre-existing Se deposits due to favorable energetics when a sufficient amount of LiPSe is supplied. As the Se nucleus grows to a certain extent, the local concentration of LiPSe around the site decreases, leading to a reduced growth rate. Consequently, new nuclei start forming at a distance, and these newly generated nuclei tend to be smaller than the ones already formed, owing to the limited availability of LiPSe. These findings align with results previously reported in the literature, highlighting the impact of the concentration of LiPSe near the formed nuclei on the nucleation and subsequent growth behavior of Se26.Fig. 6: Operando TXM measurements of Li | |Se:S (1:10) blended cell.a Schematic illustration for comparison of Se crystal growth mechanism based on the Se ratio in the positive electrode and the corresponding crystal growth of SexSy in the case II. (Se = gray, S = yellow, and Li = green). b Voltage profile of the Li | |Se:S (1:10) blended coin cell at a specific current of 600 mA g−1 for the first charging process. c TXM images from S70 to S100 (scale bar = 50 μm). The operando TXM can be seen in Supplementary Movie 4. d TXM images of enlarged two areas of the red and green box of S100 in c from S79 to S100. This measurement was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode (positive electrode active material mass loading = 2 mg cm−2).On the other hand, in Case II, a positive electrode with a low Se ratio generates a lower concentration of LiPSe during the discharging. Due to the reduced and insufficient resource supply, the growth of Se occurs slowly, allowing simultaneous nucleation rather than individual growth. This process leads to smaller and more evenly distributed crystals. A more detailed description of each step can be found in Supplementary Fig. 12 and Supplementary Note 3. Concerning SexSy-based positive electrodes, our previous calculation data confirmed that S particles tend to grow on Se seeds preferentially over carbon fiber. A positive electrode with a lower ratio of Se is expected to result in smaller seeds for the growth of S particles and promote homogeneous SexSy crystal growth.To validate our hypothesis, operando TXM analysis was conducted on a blend-based positive electrode, comprising the Se:S molar ratio as 1:5 and 1:10, as shown in Supplementary Figs. 13 and 14, respectively. TXM images were extracted from Supplementary Movie 3 and Supplementary Movie 4. As shown in Fig. 6b, the voltage profile of the Li | |Se:S (1:10) blended coin cell during charging enables the visualization of the gray- and orange-colored circles where each selenium and sulfur crystals nucleated. As shown in Fig. 6c, Se crystals, indicated by white circles, appeared smaller than those in positive electrodes with a higher Se ratio such as SeS2. The growth of sulfur particles was indicated by red circles from S79 to S100. In Fig. 6d, the red and light-green square boxes from Fig. 6c were enlarged to provide more detailed views, demonstrating the growth of sulfur particles along with the selenium seeds, as tracked by the respective red and light-green arrows. The average area of the Se:S (1:10) crystal was approximately 545 μm2. Supplementary Fig. 15 illustrates distinct Se nuclei observed in TXM images of each electrode captured prior to S formation. It is apparent that as the Se ratio decreases within the electrode, there is a concurrent increase in the number of Se nuclei along with a decrement in their size. This observation aligns with our hypothesis, providing confirmation of the impact of the Se ratio on nucleation behavior.Our prior investigation has substantiated that SeS2 exhibits a closely analogous reaction behavior when compared to Se:S (1:2) blended electrodes. This empirical verification is particularly conspicuous within the SEM images. Figure 7 presents SEM images in BSE mode, portraying positive electrodes characterized by varying the Se-to-S ratio. These images offer a comprehensive depiction of crystal morphology and size comparison. We illustrate this tendency in Fig. 7a. Manipulating chalcogen ratios to achieve more uniform distribution of active materials results in enhanced homogeneity in electrochemical processes in lithium chalcogen batteries34,35,36.Fig. 7: Ex situ SEM measurements of positive electrodes containing various Se:S molar ratio.a Schematic illustration for the comparison of different Se-to-S ratio positive electrode active materials at fully charged state (Se nuclei = black, Se = grey, and S = dark yellow). b Ex situ SEM images at fully charged state at a specific current of 600 mA g−1 after the first charging process in a BSE mode. The red boxes represent the gradually enlarged SEM images in the vertical direction. This measurement was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode (positive electrode active material mass loading = 2 mg cm−2).As shown in Fig. 7b, the appearance of Se in SeS2 and Se:S (1:2) blended electrodes is very similar in terms of shape and dimension. The substantial size of these Se crystals, often ranging in the order of hundreds of micrometers, making difficult to distinguish the presence of S on Se crystal surface. Nonetheless, Fig. 3d provides evidence through EDS analysis, corroborating the simultaneous presence of S on the Se surface. Notably, the influence of Se in facilitating S growth becomes pronounced as Se concentration decreases. Reduced Se concentration leads to a diminishment in Se crystal dimensions to the scale of tens of micrometers, thus establishing catalytic sites conductive to S deposition. Furthermore, an examination of the seed surface in Se:S (1:10) reveals the envelopment of numerous S clusters around Se, thereby reinforcing the claims presented earlier. These findings support Se’s pivotal role as a catalytic site in promoting the uniform deposition of S.Prior researches on chalcogen composites primarily focused on their application in a singular composite form or materials engineering aimed at achieving Li-based batteries with improved electrochemical energy storage performances37,38,39. Te has been recently introduced as an additive into S and Li2S-based positive electrodes, resulting in the formation of a solid electrolyte interphase (SEI) on the surface of the Li metal electrode to enhance battery performance40. While that study primarily addressed the effects on the lithium metal negative electrode, the present study demonstrates the pivotal role of positive electrode active material morphological evolution in affecting battery performance.Optimization of the Se-S-based positive electrode formulationTaking into account the results presented in this article, we could summarize that: when a small amount of Se is present, resulting in a low concentration of LiPSe, Se serves as a nucleation seed that uniformly permeates the electrode during the initial charging. Afterward, S growth occurs along these Se seeds. Therefore, by carefully adjusting the Se-to-S ratio, it is possible to grow S on a minor quantity of Se seeds. This is expected to yield a synergetic effect, capitalizing on the rapid kinetics of Se and the high capacity of S simultaneously.To experimentally validate these claims, we fabricated three electrodes with varying molar ratio of Se:S, ranging from 1:2 to 1:10. We compare the voltage profiles of Li | |S coin cells containing three different Se-S-based positive electrodes varying the specific currents as shown in Fig. 8a–c. It is evident that the galvanostatic discharge/charge voltage profiles of the three Li | |Se-S cells gradually follow the voltage profile of the Li | |S cell as the relative amount of Se decreases. The discharge capacities obtained from the long plateau at 2.1 V and a specific current of 0.1 A g−1 are 759, 691, and 605 mAh g−1 for the Li | |Se-S cells with the 1:10, 1:5, and 1:2 electrodes, respectively. Additionally, the stair-like profile observed in the nearly final discharge state, for the cells with the 1:5 and 1:2 electrode and resembles Li | |SeS2 cell (Fig. 1b), is not observed in the Li | |Se:S (1:10) blended cell. This latter cell, which contains an optimized amount of Se, exhibits the longest voltage plateaus and the least polarization under all specific current conditions. The rate capabilities of these positive electrodes tested in Li | |S coin cell configuration are presented in Fig. 8d. Among the three samples, the 1:10 electrode enables the highest cell reversible capacities, achieving 1077, 908, 844, 815, 801, and 769 mAh g−1 at specific currents of 0.1, 0.2, 0.3, 0.4, 0.5 and 1.0 A g−1, respectively. These capacities represent values for intermediate cycles at each given specific current. Particularly, after enduring high specific current, the discharge capacity rebounds to 853 mAh g−1 when the specific current is reduced to 0.2 A g−1. In comparison with the other electrodes, the Li | |Se:S (1:10) blended coin cell exhibits the highest reversible capacity and rate capability, thanks to faster redox reaction kinetics facilitated by the homogeneous nucleation seeds of Se.Fig. 8: Battery performances of Li | |S coin cells with positive electrodes containing various Se:S molar ratio.a–c Voltage profiles of Li | |S coin cells with various SeS2-based positive electrodes corresponding to the cycle numbers 3, 8, 13, 18, 23, 28, and 33 in (d). d Rate capability tests of Li | |S coin cells with SeS2-based positive electrodes having various Se:S molar ratio (Se:S = 1:2, 1:5 and 1:10) and 2 mg cm−2 of active material loading at various specific currents. e Long-term cycling stability of Li | |S coin cells (at a specific current of 1.0 A g−1) with various SeS2-based positive electrodes with active material mass loading of 2 mg cm−2 and E/A = 5 μL mg−1. This measurement was conducted at 25 °C, and the mass of the specific current refers to the mass of the active material in the positive electrode.Furthermore, as shown in Fig. 8e, we conducted long-term cycling tests for all the various Li | |S cells with an active material loading of 2 mg cm−2 at a specific current of 1.0 A g−1 for 100 cycles. After pre-activation for 2 cycles at a specific current of 0.1 A g−1, the Li | |S cells with the 1:10, 1:5, 1:2 and SeS2 positive electrodes exhibited capacity retentions of 72%, 64%, 52%, and 67%, respectively. Particularly, the Coulombic efficiency (CE) over 100% during the early cycles in 1:2 electrode is observed as shown in Fig. 8e and the CE data plotted in the 90–103% range are reported in Supplementary Fig. 16. The high CE (>100%) means that the discharge capacity was larger than charge capacity, indicating 1:2 electrode does not undergo the pre-activation as well as other electrodes during early cycles. As shown in this article, the 1:2 and SeS2 electrodes undergo electrochemical reaction through the same mechanism, but the difference in reactivity is observed. In addition, the capacity fading observed in the early cycle performance of the cell with the 1:2 electrode, as depicted in Fig. 8d, e, is more evident when compared to the one of the cell with the SeS2 electrode, despite undergoing similar mechanism. In spite of the similarity, variations in electrochemical properties can be present between SeS2, featuring S and Se in a specific atomic configuration, and the 1:2 blended electrode, which constitutes a simple mixture at a molar ratio in bulk. These distinction impacts the reaction kinetics during cell discharge and charge cycles. To further substantiate the differences in reactivity between SeS2 and the 1:2 electrodes, cyclic voltammetry (CV) analyses were performed, as illustrated in Supplementary Fig. 17a. During the cathodic scan, three representative peaks were observed. We further analyzed the peak potential (Supplementary Fig. 17b) and onset potential (derived by using the tangent method41, Supplementary Fig. 17c) obtained by the CV (Supplementary Fig. 17a). SeS2 exhibited a higher peak potential (2.33 V, 2.17 V, and 2.04 V) and onset potential (2.41 V, 2.30 V, and 2.10 V) than the 1:2 electrode at peaks 1, 2, and 3, respectively, indicating a lower overpotential. Additionally, CV measurements at various scan rates were conducted to examine reaction kinetics, as shown in Supplementary Fig. 18. Analysis of the cathodic peak currents linearly fitted according to scan rates showed that SeS2 exhibited steeper slopes at all peaks, indicating faster redox reaction kinetics. This finding supports our hypothesis that SeS2, with its atomic-level composites, demonstrating improved performance compared to bulk-mixed blended electrodes with the same ratio. At low Se concentrations, the incorporation of Se seeds enables the highest cell rate capabilities and long-term performance, particularly in catalyzing the uniform deposition of S. This can be attributed to Se’s role in facilitating the even distribution of S, resulting in enhanced S utilization and rapid reactivity. Consequently, this leads to the efficient execution of higher-rate charge and discharge processes, ultimately extending the cell’s cycle lifespan.In summary, this study sheds light on the intrinsic morphological behavior of chalcogen-based positive electrode active materials, particularly focusing on SexSy for non-aqueous lithium sulfur batteries. Our study revealed the behavior of SeS2 as it undergoes separation into Se and S during the initial cycle. During the subsequent charging process, dissolved Se forms seeds with a low nucleation voltage, while S preferentially deposits and grows on these seeds. We found that a mixture of Se and S exhibited analogous behavior to SeS2, emphasizing the catalytic effect of Se seeds and the potential for achieving enhanced performance compared to the separate reactions of Se and S.In contrast to many prior studies that primarily emphasized performance improvement and focused on the influence at the Li metal negative electrode, our research involves real-time observation of the internal behavior occurring at the positive electrode. Based on this understanding, we have introduced an optimized strategy to maximize the catalytic effects of Se. Ultimately, our work demonstrates that by utilizing a trace amount of Se to create catalytic seeds in the S-containing positive electrode, a synergistic effect can be achieved. This effect enables high cell capacity while enhancing reaction kinetics.
Understanding the electrochemical processes of SeS2 positive electrodes for developing high-performance non-aqueous lithium sulfur batteries
