163
Lithium Deposition and Dissolution in Highly Concentrated Electrolyte Solutions

Wednesday, 4 October 2017: 17:00
National Harbor 1 (Gaylord National Resort and Convention Center)
M. Inaba, N. Sekiguchi, M. Iwata, M. Haruta, and T. Doi (Doshisha University)
Lithium metal has a high energy density (3860 mAh g-1, 2062 mAh cm-3), and is the most promising anode in advanced LIBs. It is also important as an anode in Li-S and Li-air batteries in the innovative batteries in the future. Dendrite formation is a serious problem in cycleability and safety for lithium metal anode. Lithium dendrites are formed by (1) inhomogeneous thickness of the SEI formed on lithium metal, (3) local current concentration to the edges of deposited lithium, (3) excess nucleation during deposition, etc. Concentrated electrolyte solutions have attracted attention as electrolytes for LIBs because they have many unique characteristics [1-4]. We have so far investigated concentrated electrolyte systems that are stable against 5 V-class spinal LiNi0.5Mn1.5O4 cathode [5-7]. It has been reported that the use of concentrated LiFSI/tetraglyme (G4) effectively suppress lithium dendrite formation [8]. Unfortunately ether compounds cannot be used for even 4 V-class positive electrodes such as LCO, NCA, and NMC. In this study, we investigated several concentrated systems using carbonate-based solutions to suppress the dendritic growth of lithium during electrochemical deposition.

A thin layer of Au (ca. 30 nm) was deposited on Cu foil and used as a working electrode. Nearly saturated 7.8, 10, and 3.9 mol kg-1 solutions were prepared using PC, DMC, and EC+DEC (1:1 by volume) as solvents and LiFSI was used as an electrolyte. 1 M solutions (0.83, 0.93, and 0.86 mol kg-1, respectively) were also used for comparison. A coin-type three electrode cell was constructed using Li foil as a counter and a reference electrode. Lithium was deposited at a current density of 0.1 or 1.0 mA cm-2 for 10 h, and then dissolved at the same current density to a cut-off voltage of 1.0 V.

Figure 1 shows SEM images of the electrode surfaces after the first lithium deposition. Dendrites were observed in each 1 M solution, but their formation was remarkably suppressed in each concentrated solution. These facts clearly show that the use of concentrated solution is effective for suppressing lithium dendrite formation. Figure 2 compares SEM images of the cross-sections of the electrodes after the 1st and 11th deposition in the concentrated solutions. The particle size of the deposited lithium was larger in 7.8 mol kg-1 PC and 10 mol kg-1 DMC solutions than in 3.9 mol kg-1 EC+DEC after the first deposition, and increased further after the 11th deposition. The high concentration seems to be effective for suppressing excess nucleation and enhancing the particle growth of lithium. In 3.9 mol kg-1 EC+DEC solution, the thickness of the lithium layer was thin after the first deposition, but increased significantly after the 11th deposition. Probably a large amount of decomposition products of the solvent was deposited in the gaps of lithium metal.

Figure 3 shows coulombic efficiencies for lithium deposition/dissolution cycles in 1 M solutions and 7.8, 10 and 3.9 mol kg-1 solutions of LiFSI dissolved in PC, DMC, and EC+DEC (1:1). The coulombic efficiency was low in 1 M PC and DMC solutions, but increased significantly in 7.8 mol kg-1 PC and 10 mol kg-1 DMC solutions. It was high even in 1 M EC+DEC solution and slightly increased in the 3.9 mol kg-1 solution. The average coulombic efficiencies for the 10 cycles were 84.0%, 92.6% and 96.8% in 7.8 mol kg-1 PC, 10 mol kg-1 DMC and 3.9 mol kg-1 EC+DEC, respectively. The EC+DEC solution gave the highest coulombic efficiency, which is due to the formation of an effective SEI layer formed by the decomposition of EC. Unfortunately the solubility of LiFSI in EC+DEC is low (3.9 mol kg-1), which did not effectively suppress the excess nucleation as shown in Fig. 2. More concentrated EC-based solution will suppress the dendrite formation and give a high coulombic efficiency.

This work was supported by NEDO, Japan.

References

[1] S.-K. Jeon et al., J. Power Sources, 175, 540 (2008).

[2] K. Dokko et al., J. Electrochem. Soc., 160, A1304 (2013).

[3] K. Yoshida et al., J. Am. Chem. Soc., 133, 13121 (2011).

[4] Y. Yamada et al., ACS Appl. Mater. Interfaces, 6, 10892 (2014).

[5] T. Doi et al., Electrochim. Acta, 209, 219 (2016).

[6] T. Doi et al., J. Electrochem. Soc., 163 (10) A2211 (2016).

[7] T. Doi et al., J. Electrochem. Soc., 164 (1), A6412 (2017).

[8] H. Wang et al., ChemElectroChem., 2 (8)1144-1151 (2015).