Chemo-mechanical failure mechanisms of the silicon anode in solid-state batteries

Abstract

Silicon is a promising anode material due to its high theoretical specific capacity, low lithiation potential and low lithium dendrite risk. Yet, the electrochemical performance of silicon anodes in solid-state batteries is still poor (for example, low actual specific capacity and fast capacity decay), hindering practical applications. Here the chemo-mechanical failure mechanisms of composite Si/Li₆PS₅Cl and solid-electrolyte-free silicon anodes are revealed by combining structural and chemical characterizations with theoretical simulations. The growth of the solid electrolyte interphase at the Si|Li₆PS₅Cl interface causes severe resistance increase in composite anodes, explaining their fast capacity decay. Solid-electrolyte-free silicon anodes show sufficient ionic and electronic conductivities, enabling a high specific capacity. However, microscale void formation during delithiation causes larger mechanical stress at the two-dimensional interfaces of these anodes than in composite anodes. Understanding these chemo-mechanical failure mechanisms of different anode architectures and the role of interphase formation helps to provide guidelines for the design of improved electrode materials.

Fig. 1. (Electro)chemical stability of composite Si/LPSCl anodes.

a, HAADF-STEM image of Si particles and the corresponding EDS map. b, TEM image of a Si particle. c, Average-background-subtraction-filtered HAADF-STEM image at the Si|LPSCl interface. d, Electronic and ionic conductivities of the just-mixed Si/LPSCl as a function of time. e, Procedure for resting and impedance measurements based on a three-electrode cell. The inset shows the setup of the three-electrode cell. f, Nyquist plot and the corresponding equivalent circuit used to evaluate the impedance data (working electrode versus RE). g, Nyquist plots of a typical cell with long-term resting. h, R_int as a function of the square root of time (t^0.5).

Fig. 2. Characterization of SEI components at Si|LPSCl interfaces.

a,b, S2p (a) and Si2p (b) XPS spectra of Si|LPSCl before cycling and after different cycles. c,d, HAADF cryo-STEM image (c) and corresponding EDS mapping (d) of Si/LPSCl after 100 cycles. e,f, Li K-edge (e) and O K-edge (f) EELS spectra of Si/LPSCl after 100 cycles. The orange box in the inset in c encloses the region from which EELS spectra were obtained. g, Box plots of SEI-related signal intensities (that is, LiP^–, LiS^–, LiCl^– and SiO^–) from ToF-SIMS surface analyses of Si/LPSCl before and after 100 cycles. The SEI-related signal intensities were normalized by the S^– signal intensity. The lines in boxes depict the median and the lower- and upper-box limits indicate the 25th and 75th percentiles, respectively. The whiskers extend to ±1.5× interquartile range, and the points are outliers. h,i, ToF-SIMS images of the Cl^– fragment, and the product of the LiS^– and S^– fragments in Si|LPSCl composites before cycling (h) and after 100 cycles (i). All the signal intensities were normalized by the total signal intensity. HAADF cryo-STEM, HAADF-STEM operated under cryogenic conditions.

Fig. 3. Ion/electron transport in SE-free Si anodes.

a, Lithiation curves of the SE-free Si and Si/LPSCl anodes at 0.1C. b, Lithium chemical diffusion coefficient of the Li_xSi alloys for different SoCs. c, Simulated crystal and amorphous LiSi₃ structures before and after the melt-and-quench process. d, Simulated ionic conductivity and the corresponding activation energy of amorphous Li_xSi alloys for different SoCs at 300 K. e, Simulated electronic conductivity and the corresponding electron concentration of amorphous Li_xSi alloys for different SoCs at 300 K. f, Resistance changes in the Si anode during the lithiation process.

Fig. 4. Cycling stability at the 2D and 3D Si|LPSCl interfaces.

a,b, Cycling performance of the In/InLi|LPSCl|Si/LPSCl cell (a) and the In/InLi|LPSCl|Si cell (b) at 0.1C under 50 MPa. c–f, Cross-sectional SEM images of the Si/LPSCl anode before cycling (c), after the first lithiation (d), after the first delithiation (e) and after the 100th delithiation (f). g–j, Cross-sectional SEM images of the SE-free Si anode before cycling (g), after the first lithiation (h), after the first delithiation (i) and after the 100th delithiation (j).

Fig. 5. SE-free Si anodes in Si|LPSCl|NCM@LBO full cells.

a,b, Cycling performance of the Si|LPSCl|NCM@LBO full cell (a) and the In/InLi|LPSCl|Si half-cell (b) with a cut specific capacity of 2,700 mAh g^–1 at 0.1C. c,d, Galvanostatic cycling of the In/InLi|LPSCl|Si half-cell (c) and the Si|LPSCl|NCM@LBO full cell (d) along with the measured stack-pressure changes. Each dataset shows six cycles that were performed at 0.05C for the first cycle and 0.10C for the following cycles. The pressure at t = 0 is 50 MPa.