Study on Li Pre-Doping Technique with Li-Naphthalene Solutions Toward Si Negative Electrode for Next-Generation Batteries

Abstract

Si is one of the most attractive negative electrode materials for balanced design of high energy density Li-ion, Li-O2 and Li-S batteries because of the high theoretical capacity of 3580 mAh g-1 delivered by Li3.75Si. Li pre-doping should be an essential technology to reduce the irreversible capacity of Si negative electrodes in the first cycle or to load Li source to positive electrode materials free of Li such as for Li-O2 or Li-S batteries.Various methods of Li pre-doping have been reported: direct contact with Li metal foil or a solution of organo-lithium radical salt like n-butyl lithium, electrochemical Li pre-doping, and mixing Li metal, etc. [1]. So far, we have reported that Si electrode pre-doped by direct contact with Li metal in an electrolyte solution containing fluoroethylene carbonate showed an excellent cycle life and high Coulombic efficiency [2]. Although this method is suitable to shallow impregnation for thin Si electrodes, it is not applicable to thick and dense electrodes of practical use [3].Therefore, we focused on method of solution of organo-lithium, Li-naphthalene, to pre-dope Li into Si. As for the method, the choice of solvent is critical to control the reactivity such as reaction velocity and penetration depth of Li. Five kinds of ether solvents were adopted in this study: three of cyclic ethers i.e. tetrahydrofuran (THF), 2-methyltetrahydrofuran (MeTHF) and tetrahydropyran (THP), and two of linear chain ethers i.e. dimethoxy ethane (G1) and trietylene glycol dimethyl ether (G3). Figure 1 shows discharge capacities of Li pre-doped Si electrodes as a function of pre-dope time at 30 °C. It is explicit that the discharge capacities of cyclic ethers such as MeTHF, THP and THF are higher than those of chain ethers such as G1 and G3. Moreover, MeTHF, THP, and G3 exhibited monotonous increasing capacities as pre-doping time increases, while THF and G1 showed decreasing tendencies. The XRD diffraction results exhibited the formation of Li3.75Si in the Si electrodes doped in the cyclic ether solutions for 24 hours. Figure 2 shows that the equilibrium potential of each Li-NTL solution decreased with increasing temperature in the order of G3 > G1 > THF > THP, MeTHF. The equilibrium potential is a good index of formation of Li-NTL radical salt. Therefore, the order of the potential is consistent with the order of discharge capacities shown in Figure 1, although the order of G1 and G3 is inconsistent between them. Figure 2 also suggests clear dependency of pre-doping rate on temperature: The higher the temperature, the higher the pre-doping rate. The dissociation degree of Li-NTL solutions were estimated from the Walden plots of the solutions by using the data of the fluidity and ionic conductivity. The result demonstrated that the dissociation degree of Li and naphthalene radical was in the order of MeTHF << THP ≤ THF. Considering the above results, this should indicate that Li-NTL(solvent)n contact ion pairs are relatively stable in MeTHF or THP solution without the reductive decomposition compared with those in THF solutions.The present results reveal that Si electrode can be effectively pre-doped with Li by using Li-NTL solution, and that the choice of solvent for Li-NTL solution is one of the critical factors to obtain high Li pre-doping level. Li3.75Si phase was formed by pre-doping for 24 hours in the Li-NTL solution using MeTHF or THP, achieving high Li pre-doping levels over 3000 mAh g-1. The solution pre-doping adopting Li-NTL is one of the promising methods for deep Li-doping, revealing that the lower equilibrium potentials resulting from deep Li doping is due to stable Li-NTL(solvent)n contact ion pairs.This work was partially supported by NEDO RISING2 project (JPNP16001).