Dissolution and Reprecipitation of Sulfur on Carbon Surface

A better understanding of the redox process of lithium polysulfide (LPS) on carbon surfaces is helpful for designing Li/S batteries with better performance. The “shuttle mechanism” can explain the low coulomb efficiency and self-discharge of a Li/S battery, but it cannot explain the fact that battery performance is strongly affected by electrolyte volume and sulfur load. This paper aims to reveal the main redox process of LPS on the surface of carbon by examining the cathodic behavior with different electrolyte volume and sulfur load. Scanning electron microscopy (SEM) images and impedance spectra of the cathode before and after the first discharge were compared, and it was found that the discharge process is the continuous dissolution of sulfur composited with carbon into the electrolyte to form LPS. At the same time, LPS re-precipitates sulfur on the surface of the cathode through a disproportionation reaction to form a solid film. Cyclic voltammetry (CV) curves showed that the solid film passivates the electrode, and the electrode is activated only when the potential is swept negatively and Li2S is generated. When a lean electrolyte is used, there is fluctuation in the CV curves, which proves that the dissolution-reprecipitation of sulfur is the main process of the cathode. The discharge–charge curves of cathodes with different sulfur load were compared, and it was found that there is wavy fluctuation in the discharge curve when the sulfur load increases, which proves again that the sulfur reaction dominates the electrode process.


Introduction
A better understanding of the reaction mechanism of a Li/S battery is helpful for designing a battery with better performance. 1,2 In the past, researchers have used a mechanism similar to that of a lithium-ion battery to describe the process of a Li/S battery-that is, lithium ions shuttle back and forth between the positive and negative electrodes to realize energy conversion. 3 Later, it was recognized that the intermediate product of discharge, lithium polysulfide (LPS), is dissolved in the electrolyte and shuttles back and forth between the positive and negative electrodes (which is the so-called shuttle mechanism), so as to explain low coulomb efficiency or self-discharge or capacity attenuation of a Li/S battery. 4,5 Although the shuttle mechanism has achieved great success in guiding the synthesis of cathode materials, 6,7 the redox process of LPS described by the shuttle mechanism is inconsistent with its cyclic voltammetry (CV). It is believed that LPS is reduced to Li 2 S 4 first and then further reduced to Li 2 S and Li 2 S 2 according to the shuttle mechanism. Nevertheless, CV curves of LPS typically show two reduction peaks and one oxidation peak. 8,9 If the two-step reduction is correct, there should be two reduction peaks and two oxidation peaks on CV curves.
With the deepening of research, we find that more facts cannot be explained by the shuttle mechanism. For example, capacity attenuation is directly related to the amount of electrolyte, 10 and the sulfur load of the cathode has a great impact on cyclic performance. 11 If the redox process of LPS is only the growth and shortening of the polysulfide chain, the amount of electrolyte will not affect the performance of the cathode. Recently, we found that the redox process of LPS is an electrochemical reaction coupled with a chemical process. 12,13 Among them, the disproportionation reaction between LPS and S 8 plays an important role in the process of the electrode. The chemical reaction is described as follows: In the above formula, the forward reaction causes LPS to deposit S 8 on the surface of carbon, which leads to the passivation of cathode; backward reaction causes the S 8 to redissolve into longer-chain LPS, which causes the activation of the electrode. Multistep reactions involving LPS have been studied extensively, 14,15 yet the influence of disproportionation reaction on the electrode process has rarely been involved. It has been extensively reported that dissolved LPS changes its chain length due to its diffusion to the negative electrode and its reaction with metal lithium, 16,17 while less attention has been paid to LPS chain length change caused by disproportionation reaction at the positive electrode. Disproportionation is a liquid-solid-phase reaction, and both the volume of electrolyte and the sulfur load on the electrode affect the concentration of LPS, while the concentration of LPS plays a key role in the equilibrium of the disproportionation reaction. There are few studies related to the LPS reaction in low-dielectric-constant solvents like DME/DOL, while it is the main electrolyte solvent of a Li/S battery. In this paper, we aim to reveal the influence of the disproportionation reaction on the electrode process.

Preparation of Electrolyte and Electrode
The required LiCF 3 SO 3 mass for preparing 1 mol/L electrolyte was weighed and baked at 120°C for 2 h. Then the LiCF 3 SO 3 was dissolved in 30 ml ethylene glycol dimethyl ether and 1,3-dioxolane mixed solvent (DME/DOL, volume ratio 1:1) to form ether electrolyte.
S/MC (MC: mesoporous carbon) composite cathode: The required mass of S/MC (S content 50 wt.%; MC surface area 519 m 2 /g) and polyvinylidene fluoride was weighed (weight ratio 9:1), and 1-methyl-2-pyrrolidone was added after full mixing and continuously ground to form a slurry that was coated evenly on an aluminum sheet with a diameter of 1.3 cm and dried at 60°C in vacuum for 2 h. Two kinds of sulfur loading on the cathode were applied: 1.37 mg/cm 2 and 4.27 mg/cm 2 .

Cell Assembly
The coin cell (CR2025) was assembled in a glove box filled with argon. The S/MC composite was placed into the bottom shell, onto which the electrolyte was dropped, followed by a separator, stacked lithium sheet and foam nickel, and covered with a top shell. Then the cell was removed from the glove box and sealed with a sealing machine. The volume of electrolyte was about 0.3 ml, and UBE porous film (Japan) was used as separator.

Electrochemical Measurements
Galvanostatic charge/discharge performance was assessed in the potential range of 1.5-3.0 V at 25°C using a LAND CT2001A (Wuhan, China) with constant current of 60 mA/g-S. Cyclic voltammetry was carried out at room temperature using a three-electrode method on a CHI660d (Shanghai, China) electrochemical workstation; the working electrode was S/MC, the reference electrode was silver wire, and the counter electrode was a lithium sheet. Scan voltage range was −2.5∽ + 1 V versus Ag 2 S/Ag with a scan rate of 1 mV/s. Scanning was conducted for three cycles. Two electrolyte volumes were applied: 0.3 ml and 5 ml. Alternating current (AC) impedance of the coin cell was tested on the CHI660d workstation. The spectra were recorded via applying a 5-mV perturbation (100 kHz-100 MHz) in the open circuit.

Passive Film on Carbon Surface
The changes in morphology on the surface of the sulfur electrode before and after the first discharge are compared in Fig. 1. Before discharge, the solid sulfur was closely combined with the carbon matrix, and most of it was encapsulated in the pores of porous carbon. Therefore, only the morphology of spherical porous carbon can be seen from the scanning electron microscopy (SEM) images, while the morphology of sulfur cannot be seen. After discharge, the surface of the carbon electrode is covered by a loose solid layer which is mainly composed of spherical solids with large particle size. The solid layer was considered to be composed of Li 2 S in previous reports. [18][19][20] However, this is not supported by the dramatic volume change after discharge. From SEM, we can see that the volume of discharge products increases more than tenfold, covering the carbon surface and forming a solid film. Assuming that Li 2 S is generated from sulfur, the solid volume will increase by only one fifth. This result can be calculated from the density difference between sulfur (2.07 g/cm 3 ) and Li 2 S (1.66 g/cm 3 ). In addition, the volume of sulfur encapsulated in porous carbon is less than half of the pore volume. If the solid product is Li 2 S, it is not enough to generate a continuous solid film after discharge. The energy-dispersive x-ray analysis (EDX) results show that there is only sulfur and no lithium after the first discharge. We confirmed that the main component of the solid film is S 8 . 12 This conclusion can also be understood as follows: The disproportionation reaction is the nature of LPS, and the main product of LPS disproportionation reaction is elemental sulfur. 21 After discharge, most of sulfur encapsulated in carbon pores is dissolved in the electrolyte and then re-precipitated on the surface of the carbon electrode through disproportionation reaction. The result of this process is that sulfur is transported from the pores to the surface of carbon, which leads to the formation of a continuous passivation layer on the carbon surface. Additional evidence of sulfur production in the later stage of discharge is that the discharge capacity of the Li/S battery is 1254 mAh/g (only 3/4 of the theoretical capacity). In other words, one fourth of the sulfur cannot be used after discharge. The existence of residual elemental sulfur in the later stage of discharge has been reported previously. 22 Our previous studies also confirmed that the electrode process is an electrochemical reaction accompanied by the disproportionation of LPS to generate S 8 . 13 The disproportionation reaction of LPS precipitates elemental sulfur, resulting in the formation of solid film on the electrode.
The solid film of the carbon electrode was also analyzed by impedance spectroscopy. The impedance spectra of the sulfur cathode before and after discharge are compared in Fig. 2. It can be seen that the overall impedance of the cathode decreases after discharge. This phenomenon cannot be explained by generation of Li 2 S, which is also insulating. The decrease in impedance can only be explained by the improvement in cathode conductivity caused by dissolution of sulfur in carbon pores. In addition, the impedance after discharge has one more arc (shown in the inset) than that before discharge. The arc in the middle frequency band (8.2-176 Hz) is a symbol of film formation. 23 Equivalent circuits of the impedance spectrum before and after discharge are shown in Fig. 3. The difference of two circuits is that impedance caused by (R g CPE2) occurs after discharge, which is indicated by the pink part in Fig. 3. (R g CPE2) is employed to fit impedance of solid film produced after discharge. The meaning of elements in circuits is the same as that in literature. 23 The fitting results fits well with experimental data (shown in Fig. 3). Detailed results are listed in Table I. Impedance curves before discharge are composed of one depressed semicircle in high frequency and a slash in low frequency. The intersection of the high-frequency end and the x-axis represents the resistance of the electrolyte (R e , the fitting value = 12.69 Ω). The intersection of the low-frequency part and the x-axis represents R e + R ct . R ct = charge transfer resistance; the fitting value = 52.89 Ω. The depressed semicircle in the impedance spectrum is characteristic of a porous electrode. After discharge, the impedance spectrum turns into two depressed semicircles followed by a long sloping line. Another depressed semicircle at the middle frequency range (8.2-176 Hz) represents the formation of solid film on the electrode surface. R ct decreases to 6.68 Ω. This is because elemental sulfur embedded in carbon pores dissolves into electrolyte to form LPS during discharge, resulting in increased conductivity of the electrode and concentration of long-chain LPS. R g represents the resistance of solid film; the fitting value of R g is 6.852 Ω. The value is not high enough that formation of solid film does not block the cathodic reaction.
In brief, the discharge process may be as follows: At the beginning of discharge, sulfur in close contact with carbon obtains electrons and is reduced to Li 2 S. Li 2 S reacts with sulfur through a solid-phase mechanism to form LPS, which is dissolved in electrolyte. During the discharge, most of the sulfur is dissolved in electrolyte by reacting with LPS. At the same time, LPS re-precipitates sulfur on the surface of the cathode through disproportionation reaction to form a solid film. The subsequent cathodic process is mainly a sulfur dissolution-reprecipitation process.

Activation of the Cathode
Cyclic voltammetry of the LPS solution on the porous cathode is shown in Fig. 4. Unlike the two reduction peaks and one oxidation peak usually reported in previous literature, there is no reduction peak or oxidation peak on the curves. This is because the CV in previous studies was usually measured with two electrodes, while the CV in Fig. 4 is measured with three electrodes. It is known from our previous study that S 8 is suspended in LPS solution in the form of colloidal particles. 13 Therefore, it will also be deposited on the surface of the cathode. In the voltage range of −1∽ + 1 V (versus Ag 2 S/Ag), the current is very small. The cathode is considered to be passivated because its surface is covered with sulfur. When the potential is lower than −1 V, the current increases linearly. According to the impedance results, solid film covers the cathode surface during sulfur reduction. If the reduction product is Li 2 S, its generation will reduce the current rather than increase it because Li 2 S is an insulating solid. A possible reason can only be the decrease in the sulfur-covered area on the cathode. The process may be as follows: At the beginning, most of cathode surface is covered with sulfur, so the current is very small. With a negative sweep of potential, sulfur in close contact with carbon obtains electrons and is reduced to Li 2 S. Li 2 S reacts with sulfur to form LPS, and is thereby dissolved in electrolyte, resulting in the decrease of the sulfur-covered area. The reaction is as follows: The area covered by sulfur decreases, which leads to activation of the electrode, and therefore the current increases. This process is self-accelerated.
It is well known that the electrochemical reaction rate is much higher than the chemical reaction rate, but when the electrode surface is passivated by chemical reaction products, the whole reaction rate is determined by the chemical reaction rate. The dissolution rate of sulfur due to disproportionation reaction is slow. As can be seen from Fig. 4, the current of the second cycle and third cycle was less than that of the first cycle in the (−1∽ −2.5 V) voltage range. Although LPS undergoes multiple electrochemical-chemical reactions, the dominant reaction is the disproportionation reaction between LPS and S 8 .
The electrode in Fig. 5 is the same as that in Fig. 4, except that the volume of electrolyte used is much smaller than that in Fig. 4, which is only 1/20 of that in Fig. 4. The sulfur load is 2.5 mg/cm 2 . The overall trend of current in Fig 5 is similar to that in Fig. 4, except that current shows wavy twists and turns when potential was lower than −1 V (shown in the  inset by a circle). The reasons for the wave in current are as follows: When the potential is swept negatively, Li 2 S is generated and reacts with S 8 to form LPS. Since the volume of electrolyte is small, the concentration of LPS near the electrode soon reaches a large enough value; that is, high concentration and long-chain LPS is produced. The high concentration of LPS has great solubility in sulfur, so sulfur dissolves into electrolyte, resulting in reduced electrode passivation and increased current. Nevertheless, the disproportionation reaction is the nature of LPS. Higher concentration leads to faster generation of sulfur. Therefore, electrode passivation becomes heavier, and the current decreases. Figure 6 shows the first discharge-charge curve in a Li/S battery with sulfur load of 1.38 mg/cm 2 , which is similar to that reported in usual literature. The discharge-charge process can also be explained by dissolution-reprecipitation mechanism of sulfur. 13 The average chain length of LPS determines the discharge voltage. In a discharge platform, the average chain length of LPS remains basically unchanged. When the solubility of sulfur changes, the chain length of LPS will change, thereby leading to a swing of the platform. The average chain length of LPS decreases during discharge and increases during charging. LPS is highly unstable, and disproportionation reaction can easily lead to change of sulfur solubility. It is obvious when the sulfur load increases. Wavy fluctuation is seen in the discharge platform with a sulfur load of 4.27 mg/cm 2 , which is shown by the inset circle in Fig. 7. The platform is 1.95 V, which is different from the platform (2.05 V) in Fig. 6 and the usual report of 2.1 V. We believe that the small change in the platform originates from the change in sulfur solubility in electrolyte, and the generation of wavy fluctuation also originates from the change in sulfur solubility. When the sulfur load increases, the passivation of the electrode will increase, resulting in a decrease of the platform. With the progress of discharge, more short-chain LPS is generated; thus the solubility of LPS to sulfur decreases. The steep drop in voltage in the later stage of discharge is due to the quick decrease in sulfur solubility. The reasons for the wave in the platform are as follows: When the electrode contains more sulfur, the bulk sulfur will dissolve into the electrolyte, and thus the electrolyte near the electrode contains a high concentration of long-chain LPS. The sulfur dissolution rate on the electrode is accelerated, so the electrode is activated, and the voltage increases. High-concentration LPS can be diluted by bulk electrolyte. When the solubility of LPS to sulfur decreases due to dilution, the sulfur dissolution rate slows down, resulting in electrode passivation and decreased voltage. Results shown above prove that sulfur is dissolved and deposited at all times in the discharge cycle. In the charging platform, we also observed the wavy fluctuation. Based on the dissolution and deposition of sulfur, we use a novel mechanism to explain the charge discharge curve. 13 It is proved that sulfur dissolution-reprecipitation runs through discharge-charge cycles of a Li/S battery.

Conclusion
The kinetic process of a cathode in a Li/S battery is an electrochemical process accompanied by chemical equilibrium. Elemental sulfur dissolves and redeposits in the form of LPS through disproportionation reaction at all times in the process of electrode reaction. The carbon electrode is covered by a solid film composed of deposited elemental sulfur after Fig. 6 First discharge-charge curve in a Li/S cell; sulfur load = 1.38 mg/cm. 2 Fig. 7 The first discharge-charge curve in a Li/S cell; sulfur load = 4.27 mg/cm. 2 the first discharge. The dissolution and deposition of sulfur runs through the whole process of battery reaction. The chemical process is obvious when lean electrolyte and high sulfur load are present.