The Discharge Mechanism for Solid-State Lithium- Sulfur Batteries

The electrochemical discharge mechanism is reported for all-solid lithium sulfur batteries. Upon milling with carbon fibers, the solid electrolyte used within the cathode composite becomes electrochemically active. Analysis with Raman spectroscopy and XPS revealed the importance of bridging S-S bond formation and breaking in lithium polysulfidophosphates during electrochemical lithiation of the active solid electrolyte. Remarkably, when sulfur is introduced as an active material in the cathode composite, lithium polysulfides are formed as an intermediate product before full lithiation into lithium sulfide. The synthesis of materials based on bridging S-S bonds is an important avenue to the design of new cathodes for allsolid batteries.

electrochemical activity of lithium thiophosphate electrolytes [21,22], and here we demonstrate the impact of that activity on the AS-LiS battery discharge mechanism.

EXPERIMENTAL
Chemical Synthesis: Li 3 PS 4 (LPS) was synthesized through mechanical milling of Li 2 S and P 2 S 5 (Aldrich) in a 3:1 molar ratio. The milling was performed on a 2g scale. The milling was performed in ZrO 2 pot (45 ml) and 32 g of 5 mm diameter ZrO 2 balls. Milling is performed at 320 RPM for 30 hs.
Cathode composites: (A) Sulfur/Carbon Fibers (Aldrich)/SE composite: Sulfur : CF : Li 3 PS 4 are combined in a 35.9 : 20.5 : 43.6 mass ratio and mixed with motor and pestle. Then 1g of the mixture was added to a 45 ml ZrO 2 jar with 32 g of 5mm (dia) ZrO 2 balls. The mixture was milled at 500 RPM for 20 h to form the composite. (B) CF:Li 3 PS 4 was combined in a 20.5 : 43.6 mass ratio and mixed with a motor and pestle. Then 1g of the mixture was added to a 45 ml ZrO 2 jar with 32 g of 5mm (dia) ZrO 2 balls. The mixture was milled at 500 RPM for 20 h to form the composite.
Electrochemical Analysis: ~0.2 g of the solid electrolyte, LPS, was pressed at 4 tons/cm2 to form the separator layer. Then the cathode composite, (A) or (B), was added to one side and pressed at 3 ton cm -2 . Mechanically polished lithium foil was added to the opposite side to act as an anode. The cell-stack was held at 2 N•m 2 pressure. X-ray Photoelectron Spectroscopy: XPS analysis was performed on cathode composites after dismantling the cell. Samples were transferred under an inert atmosphere. Peakfitting analysis was performed in the Multipak analysis software. Spectra were collected with a Al Kα 1,2 (1486.6 eV) source with a pass energy of 29.35 eV. The S2p peak was fit after background subtraction (Shirley) and a 2p3/2:2p1/2 ratio of 2:1. All spectra are aligned to adventitious carbon at C1s = 284.7 eV. Quantification is performed using peak intensities and instrument sensitivity factors.
Raman: Raman analysis was performed on cathode composites after dismantling the cell. Samples were transferred under an inert atmosphere. Figure 1 is the 1 st and 2 nd discharge of an all-solid lithium-sulfur battery from at a slow C/30 rate. On the second cycle, we observe an increase in open-circuit potential (OCP), 2.16 V→ 2.35 V, and improved capacity, 2.6 mAh → 3.4 mAh, respectively. As a note, the theoretical capacity of this cathode is 4.2 mAh. First as compared to the galvanic discharge of a similar sulfur cathode, the shape of the potential curve indicates a different reaction mechanism than seen in liquid electrolytes [14]. Based on the shape of the second discharge, we defined three potential-limited regions of reaction; a high-potential reaction region is observed at the beginning of discharge (2.35 V → 2.20 V), a mid-potential reaction region (2.20 V → 2.06 V), and a 2 V-plateau region (2.06 V → 1.5 V). The cathode composite is extracted after galvanic reduction to the labeled State-of-Charge (SOC) or Depth-of-Discharge (DOD), and analyzed with Xray Photoelectron Spectroscopy (XPS) and Raman without exposure to ambient air.

Electrochemical Discharge of the Sulfur, Carbon Fiber, and LPS Electrode
Fortunately, there is a wealth of reference information to fingerprint known materials and assign the bonding of the possible sulfide species in the composite cathode [23][24][25][26].
Although the active material sulfur, S 8 , is stable in air, sulfide-based electrolytes are known to decompose upon exposure to air [27,28]. By understanding the state of the cathode under electrochemical discharge in all-solid Li-S batteries, we can design paths to increasing the active material utilization, rate and capacity.

Formation of Active Li 3 PS 4 (μ-S x )S 4 PLi 3 (x ≥ 1)
To ensure good mixing and solid-solid contact, the components of the cathode undergo rigorous milling to form the composite. Even without the presence of elemental sulfur (S 8 ), the composite made from a 6.2:1 (mol:mol) carbon fibers (CF):LPS is electrochemically active with an OCP = 1.68V and an initial discharge capacity of 1.2 mAh (Figure 2a). To investigate the role of carbon, carbon fibers (CF):LPS composites are synthesized in a (mol:mol) 1:0, 1:9, 3:7, 7:3, 9:1 and 1:0 ratios, and only composites formed from the 7:3 → 9:1 ratio are electrochemically active, with the 7:3 ratio showing the highest discharge capacity of 2.0 mAh. Under these conditions, the capacity increases with increasing LPS content, indicating that LPS is the electrochemically active component in the CF:LPS composites. Figure 2b) shows the Raman spectra of a CF:LPS composite (6.2:1), and a broad signal ranging from 380-405 cm -1 is evidence for the formation of (P 2 S 6 ) 4and (P 2 S 7 ) 4anions from the (PS 4 ) 3tetrahedra [29]. The formation of these anionic groups must be counter-balanced with a release of sulfur. In the C:LPS composites, we are not able to observe all the vibrational stretches to indicate S 8 formation. Interestingly, a broad peak centered at 480 cm -1 in the Raman spectrum clearly shows evidence for E 3 bonding signature of S-S bonds, which we assign to the formation of (PS 4+n ) 3anions, similar to the lithium polysulfidophosphates observed by Liang et al. [24]. However, as polysulfidophosphates have only previously been observed through a THF-based solution synthesis, the formation of the sulfur-bridged (PS 4 ) 3units is confirmed with S2p and P2p XPS, as shown in Figure 2c) and d), respectively. Although there is a shift to lower binding energies for the P2p signal in the bridged-anions, distinguishing the presence of bridged-S (green) from terminal-S (orange) is more clearly shown by a S2p peak at 163.5 eV, here use as an indicator of bridging sulfur [25]. Indeed, only the composites made within the active composition range 7:3 → 9:1 show the characteristic new peaks in the XPS HRES spectra. We hypothesize that Li 3 PS 4 (μ-S x )S 4 PLi 3 (x ≥ 1) is the electrochemically active material when cycled with in a full-cell configuration.  Figure 3 shows the S2p spectra of the As-made composite, after the 1 st charge, and limited to selected voltages labeled 1-3: 2.5 V, 2.2 V and 1.5 V, respectively. During oxidation of the cathode, the peak XPS S2p indicative of bridging S-S bonds at 163.5 eV increase in intensity relative to the terminal sulfides at 162.0 eV. Therefore, S-S bonds are formed during the electrochemical de-lithiation of the cathode composite layer. Conversely, the same bonds are broken as the cathode is discharged and lithiated, as the intensity of the bridging sulfur S2p peak decreases and the (PS 4 ) 3tetrahedra are restored. The electrochemical mechanism of the CF:LPS composites echoes the results of Tatsumisago et al. [22], where LPS and acetylene black composites were shown to be electrochemically active. The formation and breaking of linear S-S bonds is essential to understanding the mechanism of sulfur in solid-state composite cathodes.

Discharge Mechanism of S:CF:Li 3 PS 4
The previous results indicates the presence of two active materials in the S:CF:LPS composites; the lithiated form of the CF:LPS composite and the un-lithiated S 8 active material. Therefore, the OCP of the As-made composite is mixed-potential of a discharged and charged cathode active materials. Consequently, the cathode was cycled galvanically for 1 cycle to electrochemically delithiate the entire cathode composite before the second discharge; therefore, the mechanism begins from a full state-of-charge (100 SOC). Figure 4a) and b) show the Raman and XPS S2p results, respectively, as the composite cathode is discharge through the high-and mid-potential regions. In the Raman spectra, we identified these peaks as lithium polysulfide Li 2 S n (n ≥ 2) and the active form of the LPS, Li 3 PS 4 (μ-S x )S 4 PLi 3 . Here, we designate the peak to (PS 4+n ) 3-(n ≥ 1), and observe decrease in intensity as the cathode is discharged from 100 SOC to 25 DOD. From the XPS data, the S2p peak representative of bridging sulfur decreases in intensity which confirms the cleaving of the bridging S-S bonds is responsible for the high potential discharge capacity. Therefore, the electrochemical activity within these potentials is the lithiation of the active LPS. As shown in Figure 4 in the mid-potential region, the sulfur ring, S 8 , peak intensity decreases in both the Raman and XPS spectra. In addition, the S2p peak at 163.5 eV grows in intensity, which is counter-intuitive to the response seen for the lithiation mechanism of Li 3 PS 4 (μ-S x )S 4 PLi 3 . Indeed, the intensity for PS 4+n 3peak in the Raman spectra deceases in the Raman spectroscopy, as the shoulder for lithium polysulfides (Li 2 S n , 451 cm -1 ) increases, thus revealing that the increase in the peak intensity for bridging S-S bonds in the S2p is the formation of lithium polysulfides from S 8 . Lithium polysulfide formation is well known in liquid-based sulfur cathodes, however, the formation of linear S-S bond structures also plays a key role in the solid-state. The analysis shows that lithium polysulfides are necessary in the discharge mechanism of sulfur in the solid-state. Finally shown in the Figure 5a) Raman analysis of the 2 V-plateau (50 DOD → 100 DOD), the LiS n shoulder decreases in intensity concurrent to the growth in the Li 2 S peak in the XPS S2p (Figure 5b). Lithium polysufides are transformed into the fullylithiated Li 2 S. Due to the strong ionic bond, the S2p XPS shows a clear growth of the peak at 160.7 eV representing Li 2 S, a key discharge product whose intensity is often difficult to detect with Raman. Figure 5c) shows the atomic % of sulfur species present at different depths-of-discharge calculated from the peak-fitting analysis. Considering, that the capacity of the cathode is 3.4 mAh (1353 mAh/g sulfur ), the figure shows that from 100 SOC to 25 DOD of the S:CF:LPS composite, the first 0.85 mAh worth of charge is due to the the lithiation of the active LPS. Therefore, LPS accounts for ~0.85 mAh or ~25 % of the contribution to the total capacity. Importantly, Figure 5c) only quantifies the surface of the cathode and unreacted sulfur was detected within the cathode by Raman. Our analysis indicates that improving sulfur utilization is vital to maximizing the energy from all-solid lithium-sulfur cells.

CONCLUSION
In conclusion, the electrochemical discharge mechanism for all-solid Li-S batteries was determined to be combination of the intended sulfur active material, and the unintentional capacity gained from an active form of the solid-electrolyte. The formation and scission of linear S-S linkages are key to the electrochemical reaction with Li + ions, and to achieve full utilization of the sulfur cathode requires improved cathode composite engineering. Also, metallic lithium dendrite penetration into the cathode must be stopped to analyze the cathode under extended cycling. Indeed, a deep investigation of the charging and cycling mechanisms will reveal the stability and capability of sulfur as a cathode for all-solid batteries.