Chemical degradation has dominated the conversation about all-solid-state battery failure for good reason. Sulfide solid electrolytes possess narrow thermodynamic stability windows, making them prone to reaction with cathode active materials. Electrochemical impedance spectroscopy and X-ray photoelectron spectroscopy studies have revealed severe degradation at the LGPS/LiCoO2 interface, and recent work demonstrated that charging beyond 4.2 V versus Li/Li+ facilitates redox reactions between delithiated LiNi0.5Co0.2Mn0.3O2 (NCM) and LPSCl, forming Li2Sn and P2Sx decomposition species that raise the overall cell impedance. Space-charge layer effects have also been widely discussed as contributors to interfacial resistance. Mechanical degradation within cathode composites, by contrast, has received far less scrutiny, even though prior research showed that voids and cracks form near cathode active materials after cycling, causing contact loss at the particle-electrolyte interface and cumulative capacity fading.

Now, researchers have developed an in situ scanning spreading resistance microscopy (SSRM) technique that allows them to watch the local electronic resistance of a working all-solid-state battery electrode evolve as it charges. SSRM is a conductive atomic force microscopy variant that maps microscopic electrical information at a sample surface, and previous ex situ studies had shown that some NCM particles lose contact with neighbors after charging, increasing interparticle electronic resistance. But ex situ measurements suffer from an equivocal interpretation: securing an identical viewing field between measurements is difficult, and voltage relaxation occurs during sample transfer. The new in situ approach overcomes these obstacles by enclosing the air-sensitive sulfide electrolyte in a specially designed restraining attachment operated within a dried argon atmosphere, with an appropriate confining stress applied during electrochemical operation. The preparation proved remarkably gentle: the cell retained an accessible capacity of 77 percent, or 213 mAh g−1, of theoretical capacity at 4.03 V versus Li-In, corresponding to 4.65 V versus Li/Li+, indicating negligible damage from ion milling and sample transfer.

The resistance maps that emerged are striking. In the discharged state, different NCM particles displayed different local resistance values, a signature that prior SSRM simulation and experimental work attributes primarily to variation in the contact resistance between particles rather than differences in their intrinsic electrical resistivity. Upon charging to 3.63 V versus Li-In, the overall resistance of the NCM regions decreased, consistent with a semiconductor-to-metal transition that raises the intrinsic electronic conductivity of NCM as it delithiates. But in the upper right portion of the observation area, some NCM particles jumped to extremely high resistance, revealing electronic isolation caused by the volume shrinkage of the particles as lithium exits the crystal lattice. Previous X-ray diffraction studies quantified this shrinkage: NCM contracts by 3.5 volume percent from the discharged state to 3.63 V, and by another 3.2 volume percent from 3.63 to 3.93 V. For a spherical particle 5 micrometers in diameter undergoing isotropic contraction, that corresponds to a diameter reduction of roughly 50 nanometers, enough to break the fragile localized contacts through which particles connect.

As the charging potential climbed above 3.83 V, the overall local resistance of the NCM regions rose steadily. The mechanism traces back to the crystallography of NCM itself. An in situ X-ray diffraction study showed that the lattice parameter along the crystallographic c-axis gradually expands up to about 60 percent delithiation, beyond which it abruptly contracts. That 60 percent delithiation threshold falls between 3.73 and 3.83 V in this cell. At roughly 60 percent delithiation, NCM undergoes an H2-to-H3 phase transition accompanied by abrupt volume shrinkage and changes in particle geometry, which reduce the effective contact area between neighboring particles and thereby increase interparticle contact resistance. Remarkably, the in situ technique also captured a reconnection event: electrically isolated NCM particles in the bottom right region of the observation area reconnected to the current collector after charging to 3.93 V, possibly because the direction of lattice distortion reverses at high charging potentials. The isolation and reconnection behavior was reproduced in an independently prepared cell, confirming the phenomenon’s reproducibility.

Resistance histograms extracted from the mapping images told a consistent quantitative story. For cathode composites charged up to 3.73 V, the resistance distribution shifted toward lower resistance as the potential increased, reflecting the improving intrinsic conductivity of delithiated NCM. Once the charging potential reached 3.83 V or higher, the primary resistance distribution gradually dispersed toward the high-resistance side, indicating gradual electrical isolation of NCM particles. Critically, cross-sectional scanning electron microscopy of the same electrodes revealed no obvious cracks at the NCM/electrolyte interface or within secondary particles, even after charging to 4.03 V. The absence of visible cracks in morphological characterization does not preclude mechanical degradation driven by subtle structural changes such as lattice distortion and volume shrinkage. Because NCM particles connect through localized contact regions, minor structural changes can alter interparticle contact states, leading to isolation and reconnection that morphology alone cannot detect. Local resistance analysis thus provides a semi-quantitative evaluation of mechanical degradation invisible to conventional imaging.

The SSRM data also exposed design weaknesses in the composite electrode itself. The black regions in the electron microscopy image, predominantly occupied by acetylene black conductive additive, exhibited local resistance comparable to that of the NCM regions, indicating discontinuous conductive paths through the additive, a result consistent with previous SSRM studies showing that acetylene black regions display resistance similar to NCM even when the additive is present. Some acetylene black particles showed high resistance in the mapping, suggesting they were not functioning as conductive additives at all. Meanwhile, small NCM particles visible in the electron microscopy image were electrically isolated in the SSRM results, implying that cathode powders pulverized beyond necessity easily disconnect from their electronic conduction pathways. This finding carries a caution for electrode engineers: while trace amounts of conductive additive are likely to reduce electronic contact resistance between NCM particles and suppress their isolation during charging, the microstructure must be designed deliberately, since electronic transport in all-solid-state electrodes is highly sensitive to microstructural characteristics in ways that conventional lithium-ion batteries are not.

To connect the microscopic resistance evolution to macroscopic battery performance, the team cycled cells at five different upper cut-off potentials: 3.63, 3.73, 3.83, 3.93, and 4.03 V versus Li-In. First charge capacities rose from 117 to 209 mAh g−1 across that range, and discharge capacities after discharging to 2.38 V reached 97, 136, 159, 167, and 163 mAh g−1, with coulombic efficiencies of 83, 79, 85, 89, and 78 percent respectively. But the capacity loss over 50 cycles increased with the applied upper cut-off potential: the 50th-cycle discharge capacities were 88, 118, 145, 135, and 99 mAh g−1. Impedance spectroscopy added supporting evidence, showing that a semicircle with a peak frequency of 63 kHz enlarged with charging potential, attributed to increased charge-transfer resistance at the NCM/electrolyte interface, while the real part of the impedance in the high-frequency region also grew, associated with electronic contact resistance between NCM particles. Differential capacity curves revealed a peak near 3.1 V corresponding to the H1-to-M phase transition, whose shape evolved differently depending on the operating voltage range.

The decisive step came from comparing the experimental differential capacity curves after 50 cycles with three-dimensional voltage profile simulations built around three characteristic degradation modes. The electronic isolation model inserted insulating domains between neighboring NCM particles at a given ratio, representing loss of electronic percolation; the grain-boundary resistance model incorporated a one-pixel-thick resistance layer with conductivity spanning three orders of magnitude, from 6.4 × 10−4 to 2.6 × 10−7 S cm−1, matching the variation observed in the local resistance distribution; and the interface resistance model placed a one-pixel-thick layer with conductivity of 1.4 × 10−10 S cm−1 at the NCM/electrolyte interface to represent chemical side reactions. The simulations showed distinct signatures: electronic isolation reduced discharge capacity by rendering NCM particles electrochemically inaccessible, grain-boundary resistance produced gradually increasing overpotential with state of charge from severe lithium-concentration gradients within the electrode, and interface resistance yielded a constant overpotential across the entire state-of-charge range with a discharge peak distinctly shifted to lower potentials but no capacity loss.

The match between experiment and simulation was compelling. For cells cycled below 3.83 V, the evolution of the experimental differential capacity curves closely matched the electronic isolation model, indicating that charging below 3.83 V causes capacity loss through electronic isolation of some NCM particles, an interpretation directly supported by the electrically isolated particles detected at 3.63 V in the SSRM observations. For cells cycled above 3.93 V, the experimental curves showed expanded peak width without drastic peak shift, in good agreement with the grain-boundary resistance model, demonstrating that increased interparticle contact resistance at high potentials is a key mechanical degradation mechanism contributing to cycle fading. The overpotential observed in cells cycled above 3.93 V additionally suggested chemical degradation at the NCM/electrolyte interface under high-voltage operation. Unlike previous three-dimensional simulation studies that assumed idealized electrodes with perfect electronic contact between active material particles, these models successfully reproduced the voltage profiles of cells operated across the full range of upper cut-off potentials. The overall conclusion is sobering and actionable: mechanical failure between cathode particles is a major contributor to capacity degradation over a wide operating voltage range in all-solid-state batteries, and finding strategies to reduce and prevent the increase in electronic contact resistance between active materials will be critical for developing these batteries with long lifetimes.

Subject of Research: Mechanical degradation and interparticle contact loss in sulfide all-solid-state lithium-ion battery cathodes studied by in situ local resistance microscopy

Article Title: In Situ Local Resistance Analysis of Mechanical Degradation in All‐Solid‐State Batteries

Article References: Gamo, H., Maeda, Y., Yamagishi, Y., Kiyobayashi, T., Ishida, N., Siroma, Z., Kuratani, K., Takeichi, N., & Sano, H. (2026). In Situ Local Resistance Analysis of Mechanical Degradation in All‐Solid‐State Batteries. Advanced Science, 13(59), Article e76825. https://doi.org/10.1002/advs.76825

Image Credits: AI Generated

DOI: 10.1002/advs.76825

Keywords: all-solid-state batteries, solid electrolytes, scanning spreading resistance microscopy, mechanical degradation, NCM cathode, interparticle contact resistance, differential capacity analysis, 3D battery simulation, capacity fading, sulfide electrolytes, Situ, Local