Solid-state batteries (SSBs) are emerging as next-generation energy storage devices offering high energy density and improved safety compared to conventional liquid electrolyte batteries. The rigid solid-solid contacts in SSBs lead to a more prominent role of chemo-mechanics, with complex interactions arising from the differing chemical and mechanical properties of the solid electrolyte (SE) and active materials, particularly at interfaces. Silicon (Si) is a promising anode material for SSBs due to its high theoretical specific capacity (3,590 mAh g⁻¹), low lithiation potential (avoiding lithium dendrite growth), and low cost. However, significant volume changes during lithiation/delithiation (~300%) pose significant chemo-mechanical challenges, including SEI formation at the Si|SE interface, insufficient ion/electron transport in thick SE-free Si anodes, and potential contact loss at interfaces during delithiation. This work aims to understand these challenges and failure mechanisms in both composite and SE-free Si anodes.

Previous research highlights the instability of Si with sulfide SEs at low lithiation potentials, resulting in SEI formation. Studies on surface modification of Si particles to mitigate this are limited. The use of compact, SE-free Si anodes offers the advantage of a planar interface, reducing SEI degradation and irreversible lithium loss. However, the ionic and electronic conductivity of these anodes, especially at different states of charge (SoC), requires further investigation. The stability of interfaces during delithiation, particularly for the 2D interface in SE-free anodes, remains an open question. Existing literature lacks a comprehensive understanding of the interplay between lithium transport, microstructure evolution, and mechanical misfit effects across the heterointerfaces in Si anodes for SSBs.

This study employed a multi-scale approach combining experimental techniques and theoretical simulations to investigate the chemo-mechanical failure mechanisms of both composite Si/LiₓPSₓClₓ (LPSCI) and SE-free Si anodes. A three-electrode battery setup was used to quantify SEI growth rate. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), time-of-flight secondary ion mass spectrometry (ToF-SIMS), and electron energy loss spectroscopy (EELS) were used to characterize the microstructure and chemical composition of the anodes and SEI. Density functional theory (DFT) calculations were performed to simulate the ionic and electronic conductivity of LiₓSi alloys at different SoCs. A chemo-mechanical phase-field fracture model was developed to simulate stress and void formation at the interfaces. The electrochemical performance of both half-cells (In/InLi|LPSCI|Si and In/InLi|LPSCI|Si/LPSCI) and full cells (Si|LPSCI|NCM@LBO) were evaluated using galvanostatic intermittent titration technique (GITT) and electrochemical impedance spectroscopy (EIS).

The study revealed several key findings: (1) SEI growth kinetics and compositions: Impedance analysis showed that the SEI growth rate is significantly faster in composite Si/LPSCI anodes (k' = 10.10 h⁻¹/²) compared to SE-free Si anodes (k' = 0.30 h⁻¹/²). The SEI comprises LPSCI decomposition products (Li₃P, Li₂S, LiCl) and SiO₂-derived components (SiO₂, Li₂O, Li₄SiO₄) from the SiO₂ surface layer on Si particles. (2) Lithiation/delithiation kinetics: SE-free Si anodes exhibit high lithium diffusion coefficients (DLi = 1.0 × 10⁻⁸ cm² s⁻¹) and high ionic (σion = 1.5 × 10⁻³ S cm⁻¹) and electronic (σele = 4.4 × 10⁻⁴ S cm⁻¹) conductivities, confirmed by both GITT measurements and DFT simulations. This enabled high specific capacities (~3400 mAh g⁻¹), exceeding those of composite anodes (~2600 mAh g⁻¹). (3) Chemo-mechanics of Si anodes: Despite high ionic/electronic conductivity, SE-free Si anodes showed poor cycling stability due to void formation at the 2D Si|LPSCI interface. Phase-field modeling indicated that high stress (0.3 GPa) and plastic strain (~10%) in the LPSCI separator during lithiation and delithiation contribute to void formation. A polypropylene carbonate (PPC) layer improved cycling stability by suppressing interface degradation and alleviating stress. Full cells showed better cycling performance than half-cells, likely due to the pressure changes during cycling.

The findings demonstrate that while SE-free Si anodes offer intrinsically superior ionic and electronic conductivity leading to high specific capacity, the chemo-mechanical challenges associated with volume changes during cycling must be addressed. The formation of voids at the 2D interface due to stress accumulation highlights the importance of considering both electrochemical and mechanical properties when designing Si anodes. The successful mitigation of these issues using a PPC modification layer demonstrates the potential of interface engineering to enhance cycling stability. The difference in performance between half-cells and full cells highlights the importance of testing under realistic cell configurations. The relatively low cycling stability of even modified SE-free anodes underscores the need for further optimization of interface materials and cell designs.

This study provides a comprehensive understanding of the chemo-mechanical failure mechanisms of Si anodes in SSBs. The superior intrinsic properties of SE-free Si anodes need to be coupled with strategies to manage stress and void formation during cycling. Interface modification using materials such as PPC shows promise for enhancing cycling stability. Future research should focus on developing more effective interface modification layers with high ionic conductivity, exploring alternative anode architectures, and optimizing cell design to minimize stress and improve overall performance. Despite the challenges, the potential of Si anodes for high energy density SSBs remains high.

The study focused on specific solid electrolytes and Si anode architectures. The generalizability of the findings to other SEs and Si morphologies warrants further investigation. The DFT simulations considered idealized bulk structures, while real materials contain defects and grain boundaries that may affect the conductivity. The phase-field modeling involved simplifications of the complex material behavior. The comparison of half-cells and full cells revealed differences, highlighting the need to investigate the impact of other cell components on the chemo-mechanical behavior.