Silver Iodide as a Host Material of Sulfur for Li–S Battery

Preparation of sulfur nanoparticle

Preparation of the S/AgI composite

Materials characterizations

Electrochemical testing

As depicted in Fig. 2a, the sulfur nanoparticles were synthesized firstly by combined a mixture of sodium thiosulfate and hydrochloric acid in the presence of PVP, and then the S/AgI composites were prepared based on the double replacement reaction at room temperature. The morphology of the as-prepared AgI and S/AgI particles were observed by using SEM as shown in Figs. 2b and 2c, respectively. The AgI appears rough surface morphology formed by stacked spheres with inhomogeneous sizes. There is a difference in morphology after integration with sulfur as compared with the AgI, and the S/AgI composite exhibits a closely connected structure. EDS mappings of S/AgI-90 composite is further provided in Fig. 2d, showing the distribution of sulfur, silver and iodine elements. Obviously, silver and iodine elements are randomly distributed, indicating AgI particles are randomly precipitated throughout the sulfur nanoparticles.

Zoom In Zoom Out Reset image size

The XRD patterns were used to investigate the crystal structure and composition of the as-synthesized samples and the results are shown in Figs. 2e and S1 (available online at stacks.iop.org/JES/168/060536/mmedia). For the pure AgI, the diffraction peaks can be easily indexed to the hybrid crystal including hexagonal β-AgI (PDF # 09–0374) and cubic γ-AgI (PDF # 09–0399). ⁴⁶ The XRD result of the S/AgI-90 (Fig. 2e) and S/AgI-85 (Fig. S1) proves that sulfur exists in a crystalline state, and the characteristic peak of (002) at 23.7° for AgI can be observed. Notably, no additional peak is observed, indicating the high purity of the S/AgI-90 sample. The conductivity of pure AgI and S/AgI composite was measured to facilitate the analysis of electrode performance, and the conductivity of S/AgI-90 is significantly higher than that of sulfur (Table SI).

Thermogravimetric (TG) analysis was conducted for the S, AgI and S/AgI composites, as shown in Figs. 2f and S2. It can be seen that AgI is thermally stable within the entire heating range. A dramatic weight loss is observed within the temperature range from 200 °C to 350 °C, which is due to the thermal evaporation of sulfur. To probe whether silver sulfides are formed during thermal treatment, the final product of S/AgI in TG experiment is characterized using XRD (Fig. S3). The final product is the mixture of β-AgI and γ-AgI, and no signals from Ag₂S, Ag₂S₂, or Ag are detected. Thus, the sulfur contents in the S/AgI-90 and S/AgI-85 composites are determined to be 90.4 wt% and 84.7 wt%, respectively. This result proves a high sulfur content in the S/AgI-90 composite.

In order to identify the products of AgI and S/AgI composite at the different charge state, XRD was employed and relevant patterns are shown in Fig. 3. It is clear that the peak of about 65° corresponds to the contribution of the Al foil. ²¹ The XRD pattern of the AgI electrode (Fig. 3a) in the origin state almost consistent with the standard pattern. After discharge to 1.7 V, the diffraction peaks of AgI almost disappear, two new peaks appear at 38.1° and 44.4°, which can be assigned to Ag metal (PDF # 04–0783). When recharged to 2.8 V, the disappearing peaks of AgI reappear, but the peak intensity weakens. Meanwhile, the peak of Ag metal can still be observed, indicating that Ag was not fully oxidized to AgI. However, the reversible oxidation and reduction of AgI via lithiation can be identified from the XRD patterns observed after the charge/discharge processes, and the reaction of AgI electrode in the range of 1.7 V to 2.8 V are as follows:

$$\begin{eqnarray*}&&{\rm{Discharge}}:{\rm{AgI}}+{{\rm{Li}}}^{+}+{{\rm{e}}}^{-}\to {\rm{Ag}}+{\rm{LiI}}\end{eqnarray*}$$

$$\begin{eqnarray*}&&{\rm{Charge}}:{\rm{Ag}}+{\rm{LiI}}-{{\rm{e}}}^{-}\to {\rm{AgI}}+{{\rm{Li}}}^{+}\end{eqnarray*}$$

Zoom In Zoom Out Reset image size

The XRD patterns for the S/AgI-90 at different states are shown in Fig. 3b. The uncycled electrode exhibits peaks from crystalline sulfur. After discharge to 1.7 V, the diffraction peak of sulfur can still be observed, meaning that sulfur was not fully reduced, possibly due to the high content. However, a new peak appears at 38.2°, which can be assigned to Ag metal combined with the XRD patterns of AgI. The reappearance of the disappearing peaks of sulfur can clearly be observed after charging to 2.8 V, indicating the formation of sulfur again. In addition, it also can be inferred that the final oxidation product of Ag metal in the S/AgI composites could be a mixture of AgI, Ag₂S and Ag₂S₂ by comparing with the standard card and considering the thermal stability. ⁴⁷ From the above results, we can consider that Ag and LiI are produced during the discharge process, which can improve the conductivity of the electrode. In addition, it is believed that the proper amount of LiI can inhibit the shuttle effect by inducing the formation of protective coating, combined with literature reports and our previous work. ^42,44 This is the reason why the electrochemical performance of the cells with S/AgI as the electrode can been improved.

The Li–S coin cells were fabricated to evaluate the electrochemical performance of various cathodes. The cyclic voltammetry (CV) curves of the S/AgI-90 and S electrodes carried at a scan rate of 0.1 mV s⁻¹ are presented in Fig. 4a. During the cathodic scan, the S/AgI cathode shows two discernible reduction peaks located at 2.23 and 1.91 V, which are associated with the conversion of sulfur to long-chain lithium polysulfides and finally to Li₂S. Actually, the reduction of AgI to Ag and LiI also occurs during this process. However, the S cathode without AgI only exhibits one peak, this phenomenon is caused by the low conductivity and incomplete reaction of the electrode due to high sulfur content. Additionally, deformed and widened peaks are also associated with high content of sulfur. In the subsequent anodic scan, there is merely one oxidation peak of the sulfur electrode at about 2.48 V. In contrast, the S/AgI-90 electrode exhibits two distinct oxidation peaks at about 2.36 V and 2.46 V, indicating the effective reversible conversion of sulfur species. In fact, there is also the oxidation of Ag into AgI, Ag₂S or Ag₂S₂ in the anodic process. However, these processes cannot be distinguished because the redox potentials are close to each other and overlapped. For example, CV curve of the pure AgI electrode was investigated, as shown in Fig. S4. The cathodic peak and anodic peak are broadened, at about 2.45 V and 2.47 V, corresponding to the reduction of AgI to Ag, and the oxidation of Ag to AgI, respectively. According to the previous report, ²⁴ Ag₂S or Ag₂S₂ should be formed and involved in the redox process from the thermodynamical aspect, although the identification is difficult. In general, such a comparative result between S/AgI and S electrodes indicates that the presence of AgI improves the redox environment within the electrode to some extent, the S/AgI-90 electrode has better ionic and electronic conductivity than the S electrode, which benefit from the formation of Ag and LiI.

Zoom In Zoom Out Reset image size

The charge/discharge curves of the S and S/AgI-90 electrodes are presented in Figs. 4b–4d. The 1st charge/discharge curves at 0.1 C (Fig. 4b) and 0.2 C (Fig. 4c) illustrate that the S/AgI-90 electrode exhibits higher capacity and improved Coulombic efficiency as compared with the S electrode. The increased discharge capacity mainly derives from the second plateau, and the improved Coulombic efficiency proves the reduction of the shuttle effect. Besides, the discharge voltage of the S/AgI-90 electrode is more stable and the polarization is smaller than that of S electrode. It was also observed that the oxidation of Li₂S/Li₂S_x in the S/AgI-90 electrode is improved, as demonstrated by the voltage curves at the start of charge at 0.1 C and 0.2 C. After 100 cycles at 0.2 C, S/AgI-90 electrode delivers the discharge capacity of 758.6 mAh g⁻¹ and the Coulombic efficiency are maintained above 98.9%, while the S electrode shows a capacity of only 490.5 mAh g⁻¹ and Coulombic efficiency of only 95.4% (Fig. 4d). The S/AgI-90 electrode exhibits a smaller voltage gap and stable discharge/charge voltage. The improvement in conversation kinetics and electrochemical performance implies that the conductivity and shuttle effect are optimized by introducing AgI and the formation of Ag and LiI. It's worth noting that when the amount of AgI in S/AgI composites is too much, the electrochemical performance of Li–S battery will be adversely affected. The cathodes with AgI content of 15.3% (S/AgI-85) were investigated, and the charge/discharge curves are shown in Fig. S5a. The obvious phenomenon is the charge voltage is fluctuant and the overpotential is increasing after the following several cycles. It can be seen that the charge voltage suffers continuous fluctuation and it cannot reach the upper cut off limit at the 5th cycle, which is mainly due to the excessive LiI produced by AgI during the discharge process. The dissolution of a large amount of LiI and the shuttle of I⁻/I₃ ⁻ between the electrodes, leading to the internal micro-short circuit and severe self-discharge. ^48,49 Additionally, the I₃ ⁻ is possible to oxidize polysulfides while itself reduce to I⁻, which can also cause the above abnormal phenomena. ⁵⁰ Furthermore, the high content of AgI (50%) electrode was also investigated, and the charge/discharge curves are presented in Fig. S5b. The charge/discharge process is failed in the case of high AgI content. The possible cause of the cell failure is the formation and growth of Ag dendrites during the discharge process, which results in the short circuit of the cells.

The cycling stability was performed as illustrated in Fig. 5. The S/AgI-90 composite delivers a specific capacity of 1051.2 mAh g⁻¹ in the first cycle, a larger value of 1070.4 mAh g⁻¹ in the second cycle, and retains 773.3 mAh g⁻¹ after 100 cycles at 0.1 C (Fig. 5a). The Coulombic efficiency is maintained above 97.1% during cycles. By contrast, the sulfur electrode exhibits an initial capacity of ∼1000 mAh g⁻¹, which is slightly lower than the S/AgI-90 composite. Since AgI as a semiconductor has little effect on the improvement of conductivity, the S/AgI-90 composite shows no obvious improvement of the initial capacity. However, the formation of Ag and LiI help maintain the capacity during the long cycling. The pure sulfur electrode exhibits more severe capacity fading (only 540.6 mAh g⁻¹ after 100 cycles), and more poor Coulombic efficiency (about 94.8%). When the rate was increased to 0.5 C, as depicted in Fig. 5b, the Coulombic efficiency is above 96.2% for S/AgI-90 and near 91.9% for sulfur electrode during 500 cycles, indicating the shuttle effect owing to polysulfides could be effectively inhibited by applying AgI to the cathode materials. The S/AgI-90 composite still achieves a high discharge capacity of 502.1 mAh g⁻¹ after 500 cycles, representing the capacity decay rate of 0.092% per cycle. As a comparison, the electrode prepared with sulfur exhibited a capacity of only 166.2 mAh g⁻¹, corresponding to the capacity fade rate of 0.153%. Note that, the reversible specific capacity is calculated by the weight of sulfur, while the contribution from AgI is not considered. Actually, AgI is electrochemically active in the voltage range of 1.7–2.8 V, which has been verified in the previous CV curves. The discharge/charge profiles and cycle performances of AgI were presented in Fig. S6. The discharge capacity contribution of AgI is negligible during long cycles as compared with the large capacity of sulfur.

Zoom In Zoom Out Reset image size

The rate performance of the S/AgI-90 and S electrodes were investigated at various current densities, measuring for ten cycles at each current density from 0.1 to 2 C and returning to 0.1 C, as shown in Fig. 5c. The S/AgI-90 composite always exhibits higher discharge capacity under the corresponding current density in general. For the S cathode, a sharp decline of the discharge capacity occurred when the current density was switched from 0.2 to 0.5 C. The discharge capacity of the 10th cycle at 0.2 C was 788.3 mAh g⁻¹, whereas the initial capacity was only 240.2 mAh g⁻¹ at 0.5 C. A similar dramatic decrease in capacity was also observed on the S/AgI-90 cathode, but at a higher current density (i.e., from 0.5 to 1 C). The capacity drops rapidly with the increase of current density might be due to slow kinetics of cathode reactions caused by the high sulfur content, compact electrode structure and continuous passivation layer. ⁵¹ But it should be noted, when the current density is decreased to 0.1 C again, the specific capacity of S/AgI-90 cathode can be recovered to ∼89.9% of its final capacity at 0.1 C (872.2 mAh g⁻¹ vs 970.1 mAh g⁻¹). However, the S cathode only restores to 613.1 mAh g⁻¹, which just corresponds to 68.5% of the capacity recovery. The reason for the rapid drop in capacity and poor reversibility may be related to a very high overpotential combined with the observation in Fig. 4d. Generally, the discharge capacity of the S/AgI-90 composite is not satisfactory at a high rate according to the current literature reports, but the results also indicate that the consecutive cycling reliability is improved compared with that of the pure S electrode.

The electrochemical impedance spectroscopy (EIS) was measured for the S and S/AgI-90 electrodes before and after cycling, and the equivalent circuit models are shown in Fig. 6. Before cycling, one semicircle in the high frequency and a sloped line in the low-frequency region can be observed. The former corresponds to the charge transfer resistance (R_ct ) and the latter stands for Warburg resistance (W_o ). ^52,53 Comparing the Nyquist plots of both electrodes, the S/AgI-90 holds lower charge transfer resistance than the S electrode, indicating the improved conductivity and the sulfur redox reactions in the cathode by the introduction of AgI. This inference has also been confirmed by the conductivity measurement data and previous electrochemical results. After cycling, both Nyquist plots are consisted of two semicircles in the high-medium frequency region and a sloped line in the low-frequency region, corresponding to the formation of interfacial films, charge transfer, and ion diffusion process, respectively. The interfacial films include solid electrolyte interface, insulating Li₂S₂/Li₂S layers, and protective coating by the reaction of DME and I⁻. ⁵⁰ The S/Ag-90 electrode has lower interfacial resistance (R₁ ) and charge transfer resistance (R_ct ) than the S electrode, suggesting that the S/AgI cathode possesses more favorable interfacial properties and higher activity of charge transfer reaction, which is inseparable from the active role of the discharge products Ag and LiI. The Ag metal increases the conductivity of the electrode based on its inherent properties, and LiI can improve the interface properties of electrodes by dissolving into the organic electrolyte and increasing the contact between the electrolyte and the electrodes, as well as inducing the formation of protective coating. For the S cathode, however, there is no such advantage. After cycling, as the side products passivate the electrode interfaces, the effective cathode surface is decreased and Li⁺ transport pathway is hindered, thus the S cathode processes larger R₁ and R_ct . Based on this point, the introduction of AgI and the presence of Ag and AgI help improve the electrochemical performance.

Zoom In Zoom Out Reset image size

In summary, we have fabricated S/AgI composites as cathode materials by a facile chemical process for Li–S battery. With high sulfur content of up to 90.4%, the electrochemical performance of the S/AgI electrode has been effectively improved due to the presence of Ag and LiI. In particular, the Coulombic efficiency and cycling stability are improved by optimizing the conductivity and inhibiting the shuttle effect of polysulfides. The S/AgI-90 cathode exhibited a discharge capacity of 502.1 mAh g⁻¹ after 500 cycles, accounting for a capacity decay rate of only 0.092% per cycle, and the Coulombic efficiency remained above 96.2% during the whole cycles at 0.5 C. These electrochemical performances are significantly better than the pure sulfur cathode without AgI, suggesting the S/AgI composite would be a promising candidate for the Li–S battery. This work proposes a reactive metal halide as a host material for the cathode of Li–S battery. It is necessary to further optimize the electrode structure and investigate the effect of halides for the better electrochemical performance of the Li–S system.

Financial support from the National Natural Science Foundation (22075151, 21373118) and Fundamental Research Funds for the Central Universities of China are gratefully acknowledged.