Understanding Silicon Electrode Surface Reactivity through Model Silicate Thin Film Layers

Abstract

Development of higher capacity anodes in lithium ion batteries for use in electric vehicles is necessary in order to further enhance their energy density. Silicon anodes are being considered for these next generation lithium ion batteries due to its exceptionally high specific capacity (3579 mAh/g). One main drawback to silicon anodes is the formation of an unstable solid electrolyte interface (SEI), a surface film formed by decomposition of the electrolyte components during cycling of the battery. This unstable SEI continuously consumes Li ions and solvents from the electrolyte and stunts battery life. One major cause of this SEI instability is due to silicon anode volume expansion and contraction of up to 300% during cycling[1]. However, there is still much to learn about the chemical reactions occurring at the silicon surface before, during and after the first cycle of the battery. A useful way to narrow observations to just silicon surface reactions with no convolution from binders or conductive additives is by studying thin film silicon anodes. A neutron reflectivity study done by Veith et al utilized thin film silicon to note that the complexity of the reactions at the surface of silicon is increased by the chemical reactivity of the with the electrolyte at open circuit voltage before the anode was even cycled[2]. This implies that the length of time and composition of the surface of the anode before cycling will have an impact on how the SEI forms during the first cycle. To better understand the nature and evolution of this SEI layer formed prior to any cycling of the silicon anode and how it impacts the performance of the silicon anode, model SEI layers were deposited on clean 50 nm thick silicon thin films using RF magnetron co-sputtering. Thin film chemistries from SiO2 to Li4SiO4 were synthesized in order to model the proposed lithiation of the oxide layer during the first cycle. Model thin films were soaked in 1.2M LiPF6 in EC:EMC 3:7 wt% electrolyte from 30 minutes to 3 days in an inert atmosphere glovebox, removed and rinsed with DMC and studied using ATR IR, XPS, XRR and FIB CS. Once the pre-lithiatied film was studied using these techniques, half cells with these same silicate model films were cycled in order to observe any differences in SEI formation or cell performance. Impedance measurements were taken on model films before and after lithiation in order to better understand lithium transport of these model lithium silicates during cycling. Initial results on these model films after soaking in the electrolyte indicate a dependence on stoichiometry for the time resolved surface reactivity. While electrodes with unlithiated oxide (SiO2) soaked in electrolyte for up to 24 hours before discernable surface reaction peaks presented in their ATR IR spectra, the electrodes with lithium silicates showed significantly faster appearance of reaction peaks. This could indicate that lithium silicates passivate more quickly against the electrolyte. Preliminary FIB CS data also indicates that the SiO2 layer may be consumed with time, perhaps producing gaseous products instead of a passivating surface film. These model systems have started to tease apart the complexity of the surface reactivity and lithiation kinetics that manifest during storage and cycling of silicon anodes and could provide key insight into the low cycling stability of silicon anodes observed in practice. Ko, M., S. Chae, and J. Cho, Challenges in Accommodating Volume Change of Si Anodes for Li-Ion Batteries. Chemelectrochem, 2015. 2(11): p. 1645-1651.Veith, G.M., et al., Direct measurement of the chemical reactivity of silicon electrodes with LiPF6-based battery electrolytes. Chemical Communications, 2014. 50(23): p. 3081-3084. Figure 1