Simultaneous Suppression of the Dendrite Formation and Shuttle Effect in a Lithium–Sulfur Battery by Bilateral Solid Electrolyte Interface

Abstract Although the reversible and inexpensive energy storage characteristics of the lithium–sulfur (Li‐S) battery have made it a promising candidate for electrical energy storage, the dendrite growth (anode) and shuttle effect (cathode) hinder its practical application. Here, it is shown that new electrolytes for Li‐S batteries promote the simultaneous formation of bilateral solid electrolyte interfaces on the sulfur‐host cathode and lithium anode, thus effectively suppressing the shuttle effect and dendrite growth. These high‐capacity Li‐S batteries with new electrolytes exhibit a long‐term cycling stability, ultrafast‐charge/slow‐discharge rates, super‐low self‐discharge performance, and a capacity retention of 94.9% even after a 130 d long storage. Importantly, the long cycle stability of these industrial grade high‐capacity Li‐S pouch cells with new electrolytes will provide the basis for creating robust energy dense Li‐S batteries with an extensive life cycle.


Synthesis of the cathode materials
The SPAN was synthesized by pyrolyzing PAN and sulfur powder. A homogeneous mixture of 2.5 g PAN (M w =150,000) and 7.5 g of sublimed sulfur was created and heated to 450 °C (with a heating rate of 5 °C min -1 under nitrogen atmosphere) and held at this temperature for 6 hours. The black colored product, referred as SP-42%, was used as the active material. A high sulfur content SPAN (SP-62%) was also synthesized using the method above but with a higher ratio S (PAN:S = 1:9) and a hold temperature of 300 °C. Ketjen black @sulfur (KB@S) was prepared with ratio of KB:S as 3:7 by first holding the temperature at 155 °C for 12 hours, and then raising it to 200 °C and holding for another 2 hours. Porous carbon @sulfur (PC@S) was prepared by hydrothermal-annealing glucose, then activation with KOH at 700°C, finally the sulfur was infiltrated with 40% content.

Material Characterization
Scanning electron microscope (SEM) and transmission electron microscopy (TEM) was used to characterize the microstructure, and XRD with a 2 range of 5-80 using Cu K radiation (0.154056 nm) was used to characterize the crystal structure. An ESCALAB 250Xi was used for the XPS analysis, a Renishaw 2000 system was used for collecting the Raman spectrum, the WQF-510A was used to measure the FTIR transmission. Finally, a STA 449C thermoanalyzer with heating rate of 5 °C min -1 was used for the TGA analysis.

Electrochemical Characterization
Generally, the electrodes were prepared by coating the SPAN: carbon black: CMC (70:15:15, w/w/w) or SPAN: CMC (90:10, w/w) onto the current collector. For the assembly of the 2032 coin cells, Li foils were used as the counter electrode and the reference electrodes and the galvanostatic charge/discharge was measured using a voltage range of 1.0-3.0 V for all half cells. For the assembly of the full cells, a graphite powder: carbon black: CMC (8:1:1, w/w/w) mixture was coated onto the current collect and used as the anode (graphite anode), the SPAN electrode was used as the cathode, and the new electrolytes were used as the electrolyte. Prior to the assembly of the full cells, the graphite anodes were placed in direct contact with the Li foils for at least 1 hour to ensure lithiation of the graphite anodes. The galvanostatic charge/discharge measurements of the full cells were investigated over the voltage range of (CV) measurements were performed on an electrochemical workstation at various scan rates with a voltage range of 3.0-1.0 V for half cells and 3.0-0.8 V for full cells. The electrochemical impedance spectrum (EIS) measurements were conducted on an electrochemical workstation in the frequency range of 100 kHz~0.01 Hz.
As for the baseline data, the areal mass loading was approximately 1 mg cm -2 with the electrolyte amount encompassing the excess.     Coulombic efficiency at high rates are suppressed. While the battery capacity with TE-II was particularly low at high current densities, it was relatively high with the new NE-I, NE-II, and NE-III electrolytes and exhibited excellent Coulombic efficiency due to the formation of bilateral SEI. The corresponding rate capacities are summarized in Table S1.                indicating that the main current in the peak is capacitive. 4 It is also possible to separate this current from the capacitive and diffusion-controlled processes. The total current and the capacitive current at the scan rate of 0.1 mV s -1 reveals the occurrence of diffusion-controlled process mainly at the peak, which is due to the reaction of S with Li + . The results of the contribution ratios between the capacitive and the diffusion-controlled processes at various scan rates recorded an increase in the capacitive contributions with an increase in the scan rates. 4 The pseudocapacitance of the full cell is possibly from the Li + insertion into the graphite. Furthermore, the calculated average of this Li + diffusion coefficient of the full cell is 4.4×10 -9 cm 2 s -1 , and the high Li + diffusion coefficient is beneficial for the electrochemical performance. The Nyquist plots of the full cell before and after the completion of 50 cycles yielded a very small diameter of the semicircle with the intercept indicative of the diminutive resistance of the charge transfer and the internal resistance, which accounts for the excellent electrochemical performance.

Note S3. The kinetics analysis.
Generally, an analysis of CV curves, via the equation i = av b , is used to identify the contribution of these capacitive or diffusion-controlled processes. 4 The current i obeys a power law relationship with the sweep rate v, with a and b serving as the adjustable parameters. The b-values are calculated from the slope by plotting log i vs. log v. In particular, the b-value of 0.5 represents a fully diffusion controlled process, while a b-value of 1.0 indicates a fully capacitive process. The ratio these contributions can be estimated accordingly from i(V)= k 1 v+k 2 v 0.5 , where i(V) represents the current density at certain sweep rates, and k 1 , k 2 are constants for a given potential. The k 1 is determined as the slope by plotting i(V)/v 0.5 versus v 0.5 , so the ratio of the current due to the capacitive and the diffusion-controlled process can further be obtained by varying the sweep rates.
The Randles-Sevick equation i p = 0.4463nFAC(nFvD/RT) 0.5 is used to estimate the Li + diffusion coefficient for evaluating the kinetics of the electrodes. 5, 6 By plotting i p vs. v 0.5 , which has a Y=BX form where B = (269,000)n 1.5 AD 0.5 C. Here, n represents the number of electrons transferred in the redox process and is equal to 2 in this case, A represents the electrode area (here A=1.13 cm 2 ), F is the Faraday constant (in C mol -1 ), D represents the diffusion coefficient (in cm 2 s -1 ) and C is the concentration of Li ions (here C=0.001 mol cm -3 ).