Cycle Stability of All-Solid-State Lithium-Ion Batteries with Nanoporous Si Composite Anodes Prepared from SiO2 Fume

Abstract

A lot of anode materials have been reported to enhance the storage capacity and energy density of lithium-ion batteries (LIBs). Among such advanced candidates, Si is the most promising, mainly for the following reasons: 1) high theoretical capacity, 2) low operating potential, and 3) abundance in nature. However, the cycle life is short because the decomposition proceeds by the large volume change (> 300%) with the storage and release of Li+ ion. Very recently, we have found that sulfide-based all-solid-state LIBs (ASSLIBs) with Si composite anodes exhibited high capacity retention [1,2]. Here, nanoporous-Si particles, which were prepared through the air-oxidation of Mg2Si, were adapted as the anode active material. In the present study, the reaction of Mg2Si and SiO2 fume under vacuum was used to obtain the nanoporous-Si particles. In PRiME, we will report the cycle stability of ASSLIB comprising new Si composite anode.The nanoporous-Si particles were prepared as follows under Ar atmosphere, unless otherwise specified. Mg2Si (382.9 mg), SiO2 fume (300.0 mg), and Mg (12.1 mg) were mechanically milled at 1800 rpm for 15 min by using a high-speed mixer. The mixture was heated to 700 oC with a ramp rate of 100 oC min-1, and it was held at 700 oC for 12 hours in vacuum. The products were soaked in 1M HCl to remove MgO and unreacted Mg2Si, and then collected by centrifugation. Finally, the precipitates were rinsed by deionized water and ethanol three and two times, respectively. Anode composite material was comprised of the nanoporous-Si particles (36wt%), Li3PS4 electrolyte (55wt%), and acetylene black (9wt%). In an electric insulation tube with 10 mm diameter, the Li3PS4 powder (80 mg) and the anode composite material (2 mg) were pressed under 330 MPa to make two-layered pellet. As the counter electrode, Li-In foil was attached on the electrolyte side. Finally, the three-layered pellet was compressed at ca. 75 MPa using stainless-steel disks to prepare the all-solid-state half cells. By a charge/discharge measurement device (BTS-2004, Nagano), electrochemical tests were conducted in the constant current mode of 0.127 mA cm-2 (0.033 C) for initial 3 cycles and 0.3 mA cm-2 (0.077 C) for the following cycles at 30 oC. The cut-off voltages were 0.88 and -0.58 V vs Li-In, corresponding to 1.50 and 0.04 V vs Li+/Li.All the peaks in XRD pattern of the nanoporous-Si particles were indexed to a diamond cubic Si. The calculated lattice constant of a = 0.544 nm was close to the reported value of a = 0.543 nm (JPCDS 27-1402). By using Scherrer equation, the crystallite size was estimated to be about 39.8 nm. In the Raman spectrum, there was a broad peak at 496 cm-1. Compared with the strong scattering at 520 cm-1 in single crystal Si, assignable to the k~0 transverse optical (TO) mode, the signal shifted to lower frequency with broadening. This is due to the decrease in the crystallite size. As shown in Figure 1, the disordered nanometer pores were observed in TEM image. Based on Brunauer-Emmett-Teller analysis, the specific surface area and pore size were found to be approximately 213 m2 g-1 and 5.9 nm, respectively. These experimental results indicate that the nanoporous-Si particles were successfully prepared by the reaction of Mg2Si and SiO2 fume under vacuum. Figure 2 shows the cycle characteristics of the half cell with the nanoporous-Si composite anode. The measurement result on the cell with the composite anode containing commercially available non-porous Si particles is also exhibited to compare the features. Though the initial discharge capacities were similar in both cells, there were large difference in the cycle stability. Namely, the nanoporous-Si cell maintained the capacity retention of 78% at 50th cycle, while 17% in the non-porous Si cell. These results strongly suggest that the volumetric expansion is buffered by the shrinkage of pores in the Si particles. Our finding provides a novel and unique process to improve the cycle stability, and valuable information about the design of composite anodes for ASSLIBs.