Unraveling the Catalyst‐Solvent Interactions in Lean‐Electrolyte Sulfur Reduction Electrocatalysis for Li−S Batteries

Abstract Efficient catalyst design is important for lean‐electrolyte sulfur reduction in Li−S batteries. However, most of the reported catalysts were focused on catalyst‐polysulfide interactions, and generally exhibit high activity only with a large excess of electrolyte. Herein, we proposed a general rule to boost lean‐electrolyte sulfur reduction by controlling the catalyst‐solvent interactions. As evidenced by synchrotron‐based analysis, in situ spectroscopy and theoretical computations, strong catalyst‐solvent interaction greatly enhances the lean‐electrolyte catalytic activity and battery stability. Benefitting from the strong interaction between solvent and cobalt catalyst, the Li−S battery achieves stable cycling with only 0.22 % capacity decay per cycle with a low electrolyte/sulfur mass ratio of 4.2. The lean‐electrolyte battery delivers 79 % capacity retention compared with the battery with flooded electrolyte, which is the highest among the reported lean‐electrolyte Li−S batteries.

3 dimethyl sulfoxide (d6-DMSO). Before battery cycling, known amounts of uncycled electrolyte is injected into the coin cells, extracted following the same process and used as a reference (black curves in Figures   1a and 1b). Fluorobenzene is selected as the internal reference for both 1 H and 19 F tests because its chemical shift is far from those of DOL/DME solvent and TfSI − anions in both 1 H and 19 F spectra. 4.7 μL of fluorobenzene is added to the extracted solution for NMR quantifications. The 1 H and 19 F peaks of fluorobenzene are normalized to 100%. Due to the known content of internal reference, the retention of both DOL/DME solvent or TfSI − anions in electrolyte can be quantified by comparing the peak areas.
Electrochemical Tests. For Li−S battery testing, the sulfur cathode was prepared by mixing 80 wt.% of S with 20 wt.% of the catalyst. The catalyst/S mixture was ball-milled with LA-133 binder with a mass ratio 90: 10. Total sulfur content in the cathode was 72 wt.%. Li−S batteries were assembled with CR2032 coin-cell in an Ar-filled glove box by coupling a piece of Li-metal anode with 1 mol L -1 LiTfSI in DOL/DME (1:1 volume ratio) with 0.4 mol L -1 LiNO3 as the electrolyte. The cyclic voltammetry (CV) measurements were carried out from 1.7 to 2.8 V with a scan rate of 0.2 mV s -1 . For electrochemical test of Li−S batteries including rate and cycling performance, the batteries were galvanostatically charged and discharged at the selected current rates (1C = 1675 mAh g -1 ) and cycles. The galvanostatic charge-discharge of all the batteries was carried out using Neware battery test system (CT-4008T-5V50mA-164, Shenzhen, China).
For the Li2S nucleation tests, the coin-cell was galvanostatically discharged to 2.06 V at 0.0785 mA, and then potentiostatically discharged at 2.02 V until the current was < 10 −5 A. For Li2S nucleation testing, the Li2S8 catholyte was prepared by combining sublimed sulfur and Li2S powder in a molar ratio of 7:1 in LiTfSI/DOL/DME electrolyte under vigorous stirring for 24 h. Carbon-fiber paper (CP) disks with a diameter of 10 mm were used as the substrates to load Co, Rh and Pt catalysts with 1 mg cm -2 to assemble the coin-cells. Li-foil was used as the counter electrode. 10 μL of Li2S8 (0.25 mol L -1 ) catholyte was added on the cathode and 5 μL of blank electrolyte on the anode. Therefore, the Li2S nucleation was tested under lean electrolyte (15 μL) conditions. Computational details. Computations for this work were carried out using density functional theory (DFT) as implemented in Vienna ab-initio simulation package (VASP) code. Electronic exchangecorrelation energy was modeled using the Perdew-Burke-Ernzerhof (PBE) function within a generalized gradient approximation (GGA). The projector-augmented wave (PAW) method was used to describe the ionic cores. For the plane-wave expansion a 450 eV kinetic energy cut-off was used following testing a series of different cut-off energies. A Monkhorst-Pack 2×2×1 k-point grid was used to sample the Brillouin zone. Convergence criterion for the electronic structure iteration was set to 10 -5 eV, and that for geometry optimizations was 0.02 eV Å -1 on force. A Gaussian smearing of 0.1 eV was applied during geometry optimization and for total energy computations.
Binding energy (Eb) was computed by subtracting the energy of substrate and absorbed molecule from the energy of whole system. For example, for DOL on Co metal surface, the binding energy is computed as follows: where E (Co+DOL) is the DFT-based total energy of DOL on Co, eV, E (DOL) the energy of individual DOL in the same supercell, eV, and E (Co) the energy of Co, eV. A more negative value means stronger binding interaction.             Intensity / a.u.
Intensity / a.u. Figure S16. Raman spectra of pristine DOL, DME and DOL/DME mixed solvent.          (111). The selected Co atom shows a higher value of electron depletion for both DOL and DME molecules. This confirms its strong binding with the solvent molecules. DME on Pt (111) DME on Rh (111) DME on Co (111)