Solid-state batteries may fade because cathode particles shrink apart

Yet the cells still lose capacity as they cycle. Most of the blame has fallen on chemistry. Sulfide electrolytes tolerate only a narrow voltage range before they react with the cathode material, and studies have found heavy degradation at the boundary between one such electrolyte and LiCoO2. Charging past 4.2 V versus lithium has been linked to reactions that produce decomposition products and push up the cell's resistance.

Mechanical damage, by comparison, has had far less scrutiny, even though earlier work found voids and cracks forming near cathode particles after cycling, which cut contact and add up to lost capacity.

What did the researchers actually watch?

Think of a cathode as a crowd of people passing a bucket along a line. If one person steps back by an arm's length, the bucket stops there, even though that person is perfectly healthy. The particles in this study behave a bit like that: nothing is wrong with them chemically, but they can no longer hand off electrons.

The team used scanning spreading resistance microscopy, a variant of conductive atomic force microscopy that maps tiny differences in electrical resistance across a surface. Earlier versions of this work were done ex situ, meaning the sample was measured, moved and measured again. That makes it hard to find the same spot twice, and the cell's voltage relaxes during transfer, so readings can be ambiguous.

The new in situ approach keeps the cell running during measurement. The air-sensitive sulfide electrolyte sits in a purpose-built restraining holder in a dry argon atmosphere, with a confining pressure applied during operation. The preparation appears gentle: the cell still delivered 77 percent of its theoretical capacity, 213 mAh per gram, at 4.03 V versus a lithium-indium reference (4.65 V versus lithium), which the authors read as little damage from ion milling and transfer.

What did the resistance maps show?

The cathode material here is NCM, a nickel-cobalt-manganese oxide (LiNi0.5Co0.2Mn0.3O2 in the earlier chemistry work cited). In the discharged state, different particles showed different local resistance. Prior simulation and experiments attribute that mostly to differences in contact resistance between particles, not to differences in the particles' own conductivity.

On charging to 3.63 V versus lithium-indium, resistance across the NCM regions fell overall. That fits a known shift in which NCM goes from semiconductor-like to metal-like as lithium leaves it. But in the upper right of the viewing area, some particles jumped to extremely high resistance. The authors read this as electronic isolation caused by the particles shrinking as lithium exits the crystal.

Earlier X-ray diffraction work put numbers on the shrinkage: NCM contracts 3.5 percent in volume from discharged to 3.63 V, and another 3.2 percent from 3.63 to 3.93 V. For a spherical particle 5 micrometers across, that works out to a diameter loss of roughly 50 nanometers, enough to break the delicate point contacts holding particles together.

At higher charging voltages the team also saw contact resistance between particles rise, including around the 3.83 V point where NCM goes through an abrupt phase transition and shrinks. Electrochemical simulations based on a 3D model qualitatively reproduced the measured voltage curves and pointed to lost contact between cathode particles as a mechanical cause of fade.

What does this mean for engineers, and what's missing?

The practical message is that a solid-state cell can lose capacity without any new chemical reaction. If particles are cut off electrically, the lithium inside them is stranded. That suggests designers should think about cathode structure, particle size and the pressure applied to the cell, not only about coatings and electrolyte stability. The source summary does not detail specific fixes, though the editor's framing hints at engineering remedies.

Several things are still unclear. The paper's summary says cells were tested over 50 charge and discharge cycles at various voltage limits, but we could not see how much of real-world fade this mechanism explains compared with interface chemistry. The imaging also covers a small area of one electrode, so it shows that the mechanism exists, not how large a share it plays in a full-size cell. We also have no reaction from other battery labs or automakers, and no timeline for when any design change based on this work could reach a production cell.

In short, solid-state batteries may be fighting two battles at once: a chemical one at the interface and a mechanical one between particles. The second now has a way to be seen directly. If you follow battery news, which do you think is the harder problem to solve: a better electrolyte, or a cathode that holds still?

The finding suggests mechanical failure deserves as much engineering attention as chemical interface degradation.