A team of researchers trained some of the world's most powerful X-rays on a solid-state battery and watched it die. What they found inside — cracks, chemical ghosts, and stress fields strong enough to bend metal — may be the most detailed portrait of battery failure ever made.
Somewhere inside the next generation of electric vehicle batteries, invisible to the naked eye and to most scientific instruments, tiny cracks are growing. They radiate outward from the junction between the battery's electrode and its solid interior, branching and spreading with each charge cycle until one of them tunnels all the way through. At that point, liquid lithium metal seeps in like water finding a fissure, and the battery dies — not with a bang, but with a quiet electrical short that engineers call failure and everyone else calls disappointment. A new study, using X-rays more than a billion times brighter than those in a hospital scanner, has now captured that process in three dimensions for the first time, revealing not just where the damage occurs but why, and what the battery was already carrying long before the first charge was ever applied.
The findings land at an awkward moment for the energy industry. Solid-state batteries have been held up for years as the technology that will finally make electric vehicles genuinely mass-market: safer than today's cells, more energy-dense, and capable of lasting far longer. The research, published in the journal Advanced Energy Materials, does not diminish that promise. But it makes clear that the road to a reliable solid-state battery runs directly through some very fundamental problems of chemistry, mechanics, and manufacturing — problems that, until now, nobody had the tools to see all at once.
THE BATTERY THAT WAS SUPPOSED TO CHANGE EVERYTHING
Every rechargeable battery works the same way at its core: lithium ions shuttle between two electrodes through a middle layer called the electrolyte, driven back and forth by the charging and discharging cycles that power your phone, your car, your laptop. In today's lithium-ion batteries, that middle layer is a liquid — and therein lies the problem. Liquid electrolytes are flammable. They leak. Under stress or heat, they can trigger fires that are difficult to extinguish, a fact that has led to recalled phones, grounded aircraft, and battery packs in electric vehicles that smolder for hours after a collision.
Replace the liquid with a solid, and much of that danger disappears. A solid electrolyte cannot leak and does not ignite. Better still, it opens the door to using pure lithium metal as the battery's negative electrode rather than the graphite found in conventional cells. That matters enormously: lithium metal holds a theoretical energy capacity of around 3,860 milliamp-hours per gram, more than ten times the capacity of graphite, and it has the lowest electrical potential of any known element. In energy density terms, it is a near-perfect electrode material. The combination of a solid electrolyte and a lithium metal electrode is, in theory, a battery that is both safer and dramatically more powerful than anything widely available today.
The catch is durability. Solid-state batteries degrade fast, and they do so through a convergence of mechanical damage, chemical reactions, and electrochemical side effects that all unfold simultaneously, at the microscale, inside a device that is extraordinarily difficult to observe without dismantling. Separating these overlapping failure modes — and understanding precisely which ones matter most — has proven stubbornly difficult. Until this study, no one had managed to do it in three dimensions, without destroying the battery, and in real time.
KEY FIGURES AT A GLANCE
3,860 mAh/g: Theoretical energy capacity of lithium metal; over 10× that of graphite
>20%: Crack fraction near the electrode–electrolyte interface after cycling to failure
30×: How far interface stress exceeded lithium's yield limit (the point of permanent deformation)
15%: Volume of the solid electrolyte that was already impure before cycling even began
AN AUTOPSY WITH THE BATTERY STILL RUNNING
The research team, drawn from University College London, the University of Manchester, the University of Oxford, and the Faraday Institution in the United Kingdom, along with colleagues at the European Synchrotron in Grenoble, France, wanted to do something genuinely novel: map the damage inside a solid-state battery without opening it, while simultaneously capturing the chemistry at every point in its interior and measuring the mechanical forces at work throughout its volume — all in three dimensions, all from the same experiment.
To pull this off, they used one of the most powerful scientific instruments on the continent. The European Synchrotron, known as ESRF, is a ring-shaped particle accelerator roughly the size of a city block that fires electrons around a circular track at close to the speed of light. As those electrons are steered through curves by powerful magnets, they shed X-rays of extraordinary brightness — beams that make hospital X-rays look, by comparison, like a flashlight next to a lighthouse. Crucially, these beams are also extraordinarily precise, narrow enough to probe individual features just a few micrometers across, roughly one-tenth the width of a human hair.
HOW SYNCHROTRON X-RAYS WORK
A synchrotron accelerates electrons around a circular track until they travel close to the speed of light. As they curve through magnetic fields, they release intense bursts of X-rays that can be billions of times brighter than those from conventional sources. These beams penetrate materials without destroying them and can reveal structure and chemistry at the microscale — making them ideal for probing delicate devices, like batteries, that must remain intact and functional during the measurement.
By mounting a live battery cell on a rotation stage and bombarding it with these beams while switching between two types of detector, the team could gather two entirely different kinds of information from the same experiment within minutes of each other. The first technique, X-ray microcomputed tomography, works much like a medical CT scan: thousands of images taken at different angles are reconstructed into a three-dimensional picture that shows every void, crack, and structural feature inside the electrolyte. The second technique, X-ray diffraction computed tomography, reads the crystallographic fingerprint of every material present at every location — and, crucially, detects tiny distortions in those fingerprints that betray the presence of mechanical stress.
Think of it this way. The first technique tells you that a crack exists and shows you its shape. The second tells you what chemical compounds surround that crack, whether they got there through air exposure or electrochemical reaction or incomplete manufacturing, and whether the crystal lattice of the surrounding material is under compression or tension, and by how much. Together, they constitute something close to a full forensic reconstruction of battery failure — a pathology report with the patient still breathing.
"Near the failed interface, stresses reached more than thirty times the threshold at which lithium metal permanently deforms — enough to push the metal into every crack it could find."
— From the Research Findings
TWO CELLS, TWO VERY DIFFERENT STORIES
The experiment involved two small cylindrical cells, each built from the same design: a pellet of solid electrolyte sandwiched between two discs of lithium metal foil and housed in a compact Swagelok-style casing. The electrolyte chosen was lithium thiophosphate chloride — Li₆PS₅Cl in chemical shorthand — a sulfide-based material widely studied in the field for its relatively high ionic conductivity and commercial availability. One cell was scanned directly after assembly, never charged or cycled. The other was run through alternating charge and discharge cycles at room temperature until it short-circuited and failed, roughly twelve hours after cycling began. Then both cells were imaged at the synchrotron.
The difference between them was arresting. The fresh cell looked almost featureless under X-ray scanning: a largely uniform disc of electrolyte with a handful of isolated voids scattered through its interior. The failed cell was another world. Cracks fanned outward from both electrode interfaces, many of them spanning the full diameter of the pellet. Some extended all the way through the electrolyte's thickness from one face to the other — a continuous path from one lithium electrode to the other, open for business as a short-circuit highway. The researchers called these "terminal cracks," and named them well. More than 20 percent of the material near each interface consisted of voids and cracks in the failed cell. Further from the electrodes, that figure dropped to roughly 1 to 2 percent.
The location of the damage was itself significant. Solid-state battery researchers have long suspected that the junction between the lithium metal electrode and the solid electrolyte is the most vulnerable zone in the device. These results put that suspicion on much firmer quantitative ground: the interface is where the punishment concentrates, and it is where the battery ultimately breaks.
READING THE CHEMICAL RECORD
Physical cracks, it turns out, are only part of the story. The diffraction data painted a richer, and in some ways more troubling, chemical portrait of the failed cell — one that implicated not just cycling but also the way the battery was made and assembled.
Some of what the scans revealed was expected. When lithium metal contacts the solid electrolyte during operation, it reacts chemically, leaving behind small deposits of lithium sulfide and lithium phosphide. These byproducts are well-documented in the literature; they form a kind of reactive boundary layer at the interface and are thought to contribute to the rising internal resistance that slows a battery down over its lifetime. Their presence here confirmed that the electrochemical degradation was proceeding as predicted.
What was not predicted — and considerably harder to explain — was the presence of lithium carbonate scattered throughout the failed cell. Lithium carbonate forms when the sulfide electrolyte reacts with carbon dioxide, meaning it requires exposure to ordinary air. The cells, however, had been assembled inside a sealed argon-filled glovebox, an enclosure explicitly designed to keep both moisture and air out. The finding pointed to trace air infiltration through imperfect sealing in the cell casing — an uncomfortable reminder that even carefully controlled laboratory conditions leave room for contamination. The carbonate deposits were concentrated near crack-rich regions at the outer edge of the electrolyte, suggesting that small voids in the pellet had acted as conduits, channeling ambient air inward to pockets of material already made vulnerable by structural defects.
Then there was the impurity problem hiding in plain sight. The diffraction maps also turned up pockets of unreacted precursor materials: chemicals that are used as starting ingredients in the synthesis of the electrolyte but had never fully converted during the heating process. These were distributed through both the fresh and the failed cell alike, accounting for roughly 15 percent of the total electrolyte volume in the uncycled cell — before any charging had occurred, before any electrochemical stress had been applied. A brand-new, never-used battery, assembled under controlled conditions, was already carrying a measurable chemical impurity burden. The implications for manufacturing quality are direct.
WHY IMPURITY MATTERS SO MUCH
Foreign material embedded in a solid electrolyte is not merely a quality-control inconvenience. Pockets of unreacted precursor or air-exposure byproducts create mechanical weak points — sites where stress concentrates during cycling and where cracks are more likely to nucleate. They can also impede lithium-ion transport, raise the cell's internal resistance, and react with the electrodes in ways that generate further degradation products. In sulfide-based electrolytes, which are among the most reactive battery materials known, even trace contamination can have outsized consequences.
WHERE THE METAL BENDS
The stress measurements were the most visceral part of the story. By tracking the precise spacing of atoms in the electrolyte's crystal lattice at every point throughout the battery's volume, the team calculated the local mechanical stress — whether the material was being compressed or pulled apart, and by how much — at a spatial resolution of just a few micrometers. It was, effectively, a three-dimensional stress map of a battery failure rendered at near-cellular scale.
In the fresh cell, the stress field was essentially flat. Nothing unusual; the material was at rest. In the failed cell, approaching the electrode interface was like walking toward a fault line. The mean stress climbed steadily and, upon reaching the last 20 micrometers or so before the lithium metal, jumped sharply upward. The spread of values around that mean widened dramatically too, signaling a highly uneven stress landscape — regions of extreme tension sitting next to regions of heavy compression, sometimes within a hair's-breadth of each other.
The numbers, when converted into real physical terms, were striking. The mean stress at the failed interface reached approximately 30 megapascals — more than thirty times the stress at which bulk lithium metal permanently deforms. Peak stresses were higher still. For context, lithium metal has a yield stress of roughly 1 megapascal, the threshold at which the metal stops behaving elastically and starts to flow and creep like a soft putty under pressure. The conditions measured at the interface in the failed cell would push lithium metal well past that threshold. The mechanical logic of what follows is grimly simple: under that kind of pressure, liquid-like lithium is forced into every available crack, advancing through the electrolyte until it bridges the two electrodes and brings the battery to a halt.
Importantly, not every crack in the failed cell carried an elevated stress signature. Many of the cracks scattered through the bulk of the electrolyte showed no particular stress concentration around them at all. This distinction, between cracks that matter and cracks that do not, turns out to be one of the most practically useful insights in the study. A terminal crack — one that ran all the way through the electrolyte thickness and was the probable site of the short circuit — showed a clear spike in tensile stress near 20 megapascals, far exceeding the yield point of lithium. A deflected crack that dead-ended in the bulk showed nothing of the kind. Knowing the difference is the beginning of knowing where to look, and where to intervene.
THREE CULPRITS IN ONE CELL
One of the study's more underappreciated achievements is what it reveals about the timeline of damage. Battery researchers working with solid-state cells have generally assumed, reasonably enough, that if a cell is assembled in an inert-gas glovebox and sealed carefully, then any degradation found afterward can be attributed to cycling. These results complicate that picture substantially.
The damage inside the failed cell could be sorted into three distinct buckets, each with a different origin. The first is manufacturing: the synthesis of the electrolyte involves heating the powder at high temperatures inside carbon-coated containers, and this leaves behind residues — carbon particles and a compound called lithium acetylide — that appeared throughout both the fresh and the failed cells. These were present from the very beginning, baked in during production.
The second source is cell assembly. Despite the glovebox environment, the assembled cell showed unmistakable signs of air exposure in the form of lithium carbonate deposits concentrated near the outer edges of the electrolyte pellet. The researchers found evidence that small void networks between the pellet edge and the cell casing had provided pathways for trace amounts of ambient air to reach pockets of unmixed precursor material, reacting there to form carbonate and creating localized stress concentrations — all before a single charge cycle had run. The implication is pointed: in sulfide-based electrolytes, which react aggressively with trace moisture and carbon dioxide, the gap between "made in a glovebox" and "made in perfectly inert conditions" is not as narrow as it might appear.
Electrochemical cycling, of course, added its own far heavier damage on top — the dense network of interface cracks, the electrochemical reaction products at the electrode boundaries, and the high stress concentrations that ultimately pushed lithium metal into a terminal crack and ended the battery's life. But the study's point is precisely that cycling damage does not exist in isolation. It builds on a foundation that was already compromised. Fifteen percent of the electrolyte's volume was secondary-phase material in the fresh, uncycled cell. That is not a clean starting point.
A CLEANER PROBLEM STATEMENT
The structural weakness in much of the existing research on solid-state battery degradation is that the models and the measurements rarely speak the same language. Theoretical models propose detailed mechanisms by which stress accumulates and cracks propagate, but they are typically validated against a single type of experimental observation, usually CT imaging that shows where cracks have formed. The chemical environment around those cracks, and the actual measured stress fields, have largely been inferred rather than observed directly. The result is models that fit specific datasets well but struggle to generalize.
This study offers something different: a single, coherent dataset in which the morphology, chemistry, and mechanics of failure are all captured together, at the same locations, at the same time. When a model predicts that a particular crack geometry should produce a certain stress distribution, there is now measured data to test that prediction against. When a model assumes that air-exposure products are negligible, the data suggests that assumption deserves a second look.
There is also a conceptual simplification on offer. If you can identify which specific degradation mechanisms are creating the most dangerous stress concentrations — and which cracks carry elevated stress and which do not — then the problem of designing a longer-lasting solid-state battery can be reframed. Instead of trying to simultaneously suppress every possible failure mode, you can focus engineering attention on the mechanisms that actually matter most. A complex electro-chemo-mechanical problem becomes, in the researchers' framing, a stress minimization problem. That is more tractable. It points toward specific targets: cleaner synthesis, better sealing, microstructure designs that distribute stress more evenly, electrolyte formulations that passivate the electrode interface less aggressively.
BROADER IMPLICATIONS: WHAT THIS MEANS FOR THE WORLD
Electric vehicles: Solid-state batteries with extended lifetimes could make EVs more affordable, safer, and capable of longer ranges, accelerating the shift away from fossil fuels.
Consumer electronics: Safer, more energy-dense batteries mean thinner, lighter phones and laptops that hold a charge longer — without fire risk.
Grid energy storage: More durable cells could underpin renewable energy infrastructure by storing solar and wind power more reliably and at lower cost.
Battery manufacturing: Discovering that 15% of an electrolyte is impure before cycling begins puts direct pressure on synthesis and assembly protocols — improvements that could have immediate commercial payoff at scale.
Safety standards: Three-dimensional stress and degradation maps could inform more precise safety criteria for next-generation battery technologies in vehicles, aircraft, and grid systems.
Scientific modeling: The 3D stress and phase datasets produced here give the computational modeling community experimentally validated benchmarks they have lacked until now.
THE ROAD AHEAD
This study was, by the researchers' own accounting, an early step. Both cells were small laboratory constructs, and the measurements captured a single moment in time: the end-of-life state of a battery that had already failed. What remains to be done — and what the team views as the natural next phase — is to run these same measurements while the battery is actively cycling, watching the crack networks grow and the stress fields evolve in real time rather than reconstructing them after the fact. In the jargon of experimental science, the goal is to move from in situ to operando.
That transition is not straightforward. Synchrotron scan times are long, and batteries fail fast. To make live cycling experiments feasible, the cells will need to be made smaller and the scanning will need to become faster, without sacrificing the spatial or chemical resolution that makes the data useful. Sealing cells well enough to prevent air infiltration during hours of scanning at a synchrotron facility presents its own engineering challenge. And the researchers note that care will need to be taken to distinguish artifacts of the cell design from genuine electrochemical behavior — the very problem that this study has helped to make visible.
Beyond solid-state batteries, the researchers argue that the combined X-ray imaging and diffraction approach should be applicable to virtually any energy storage system that degrades through multiple simultaneous mechanisms. That is most of them. Cathode materials based on high-nickel or cobalt-free compositions, silicon-based anodes where large volume changes during charging are an ongoing engineering headache, tin-based electrode systems: all could in principle be mapped with the same methodology. The framework is general enough to migrate.
None of this makes the solid-state battery problem easy. The gap between a laboratory curiosity and a commercial product that outlasts the vehicle it powers remains real and substantial. But there is a difference between a problem that is hard and a problem that is hard and poorly understood. This study has shifted solid-state battery degradation, at least somewhat, from the second category into the first. The cracks were always there. Now, for the first time, scientists can read them.
Publication Details: Year Published : 2024 (published online: December 4, 2024); Journal: Advanced Energy Materials: Publisher: Wiley-VCH GmbH; DOI / Link: https://doi.org/10.1002/aenm.202404231
Credit & Disclaimer: This article is a popular science summary written to make peer-reviewed research accessible to a broad audience. All scientific facts, findings, and conclusions presented here are drawn directly and accurately from the original research paper. Readers are strongly encouraged to consult the full research article for complete data, methodologies, and scientific detail.
