Cyclic-anion salt for high-voltage stable potassium-metal batteries

Abstract Electrolyte anions are critical for achieving high-voltage stable potassium-metal batteries (PMBs). However, the common anions cannot simultaneously prevent the formation of ‘dead K’ and the corrosion of Al current collector, resulting in poor cycling stability. Here, we demonstrate cyclic anion of hexafluoropropane-1,3-disulfonimide-based electrolytes that can mitigate the ‘dead K’ and remarkably enhance the high-voltage stability of PMBs. Particularly, even using low salt concentration (0.8 M) and additive-free carbonate-based electrolytes, the PMBs with a high-voltage polyanion cathode (4.4 V) also exhibit excellent cycling stability of 200 cycles with a good capacity retention of 83%. This noticeable electrochemical performance is due to the highly efficient passivation ability of the cyclic anions on both anode and cathode surfaces. This cyclic-anion-based electrolyte design strategy is also suitable for lithium and sodium-metal battery technologies.

. All of the model systems were minimized at the beginning of the simulation. Then, a molecular dynamic of 500 ps was performed for each system at a constant temperature (298.15 K) and pressure (1 atm) (NPT), which brought them into a reasonable preequilibrated configuration for subsequent simulations. Another 500 ps NVT ensemble molecular dynamics simulation was conducted at 298.15 K to track the changes in each system in order to investigate the structure of different electrolytes. In this work, Packmol [5] was used to build all the model systems. LAMMPS [6] and CVFF force field [7] were used to perform the molecular simulations. The time step was fixed to 1.0 fs, and the temperature and pressure were controlled by the Nosé-Hoover thermostat-barostat [8]. A van der Waals interaction cutoff of 1.2 nm was employed, and the PPPM method was used to account for the long-range electrostatic interactions [9]. The geometrical structures for the selected molecular cluster structures with G16 program package based on the Becke's three-parameter hybrid method using the Lee-Yang-Parr correlation functional (B3LYP) method at 6-31+G(d) basis set [10,11]. Supplementary Figure 1| The design logic of cyclic anion and the corresponding LUMO energy levels (cyclic solvents or anions: Bold with italic). Note: In previously studies, the concept of cyclic electrolytes is typically referred to cyclic solvents. And it is reported that the cyclic solvents generally could improve the SEI formation and the oxidative stability of electrolytes. However, this concept is rarely proposed in electrolyte salts, especially the design engineering of anions. It is known that the PF6 − , FSI − and TFSI − are commonly anions in alkali metal batteries systems with molecular structures of point or linear. Herein, we propose a cyclic anion of HFDF − and applied it on potassium metal batteries for the first time, and the cyclic HFDF − based electrolyte also demonstrated excellent SEI formation and superior oxidative stability. And this design concept of cyclic anion is also suitable for other alkali metal battery systems. * The LUMO energy levels of these solvents and anions are calculated using the B3LYP method at 6-311+G(2df,2p) basis set.
Supplementary Figure 3| The HOMO-LUMO energy levels and the electrostatic potential (ESP) maps of different anions and solvents studied in this work.
* The LUMO energy levels of these solvents and anions are calculated using the B3LYP method at 6-311+G(2df,2p) basis set. SMD [24] implicit solvation model with a dielectric constant of 46.0 was used to represent the mixed solvents of EC: DMC (1:1), other parameters are based on the acetone.
Supplementary Figure 5| The room temperature discharge-charge profiles of K||Cu cells at different cycles in different electrolytes.
Note: The discharge-charge profiles (from 10 th to 100 th ) of KHFDF-based electrolyte are well-overlapped compared to other electrolytes, demonstrating its excellent reversibility.
Supplementary Figure 6| The ( Note: Lithium-ion electrolytes with four anions including of PF6 − , FSI − , TFSI − , and HFDF − are also investigated, and denoted as LiPF6-, LiFSI-, LiTFSI-, and LiHFDFbased electrolytes, respectively. The ionic conductivities of these four lithium-ion electrolytes are ranged from 7.2 to 10.6 mS cm −1 (Supplementary Fig. 7a), which are also sufficient for the normal operation of lithium batteries. For the Li||Cu cells, the LiPF6-and LiTFSI-based electrolytes exhibit slightly low CE and failed after 200 cycles ( Supplementary Fig. 7b). By contrast, both the LiFSI-and the LiHFDF-based electrolytes enable higher reversibility in Li||Cu cells. Specially, the LiHFDF-based electrolyte exhibits an average CE over 95% for 260 cycles, slightly higher than the average CE of LiFSI-based electrolyte (~ 93%). The disparity of stability and CE for different electrolytes are further verified by the plating/stripping profiles ( Supplementary Fig. 7c). Moreover, the symmetric Li||Li cell with the cyclic LiHFDFbased electrolyte could deliver a cycling stability over 750 hours ( Supplementary Fig.  7d), further demonstrating the good electrochemical performance of cyclic anion-based electrolytes.
Supplementary Figure 8|  the corresponding electrolytes are labeled as NaDFOB-, NaPF6-, and NaFSI-based electrolytes, respectively. The ionic conductivities of these three sodium ion electrolytes could reach up to over 6.8 mS cm −1 (Supplementary Fig. 8a), which are also sufficient for the normal operation of sodium batteries. Supplementary Fig. 8b shows the statistical CE of Na||Cu cells during the initial three cycles with different Supplementary Figure 18| The SEM elemental mappings and the corresponding elemental distribution percentages collected on Cu foils at the K stripping or plating states after 5 cycles.
Note: The elemental mappings of Cu foil after stripping/plating could be used as to reveal the distribution of SEI elements, 'dead K', and bare Cu. Obviously, the K elemental percentage in KHFDF-based electrolyte after stripping is much lower than the other electrolytes, indicating that less 'dead K' is remained and high stripping efficiency is obtained in this electrolyte, in good accordance with the electrochemical performance in Figure 1b. Furthermore, a high proportion of K and a negligible proportion of Cu is obtained after K metal plating in KHFDF-based electrolyte, indicating that the K metal has the tendency to cover the surface of Cu foil. These results have further verified the good K plating/stripping performance in KHFDF-based electrolyte.
Supplementary Figure 19| Photographs of the Cu foil taken after 5 cycles of stripping (a) or plating (b) in the four electrolytes.
Note: Obviously, a clean Cu foil with negligible metal K in KHFDF-based electrolyte could be obtained after K stripping. Apparently, much area of bare Cu could be observed the Cu foil after plating in KPF6-, KFSI-and KTFSI-based electrolytes. By contrast, a flat Cu foil with relatively uniform metal K could be obtained in KHFDFbased electrolyte. These results are well consistence with the SEM images and the elemental mappings, further demonstrated the good K plating/stripping performance in KHFDF-based electrolyte.
. Figure 20| The SEM images of K metal plating on Cu foil with high areal capacity of 1, 3 and 5 mAh cm −2 in KHFDF-based electrolyte.

Supplementary
Supplementary Figure 21|  Supplementary Figure 22| The O 1s and S 2p XPS spectra collected after K plating the Cu foil in the KPF6-based electrolyte.
Note: The SEI composition in KPF6-based electrolyte is composed of organic (Ccontaining species) and inorganic compounds (O/P-containing species). The signals of O 1s spectra for the three SEI exhibited dominative peaks corresponding to C=O, C-O and M-O groups, which are derived from the decomposition of solvents. For P 2p spectra, it shows apparent enrichment in P-F bond and P-O bond. And big differences could be observed for O 1s and P 2p XPS between the surface and the inner layers, further verifying its two-layered SEI.