Solid-state batteries have been "a few years away" for years now. The biggest obstacle is dendrites—branches of metallic lithium that grow through the electrolyte with every charge cycle. Most research has focused on preventing these branches from forming in the first place. A team from SLAC National Accelerator Laboratory and Stanford University tried the opposite approach: they squeezed the battery with a shape-memory alloy ring to bend the branches sideways instead. The study, published in Nature on July 1, 2026, and announced by SLAC on August 28, found that even when dendrites grew at densities never before observed in experiments, no short circuits occurred through thousands of charge-discharge cycles.
How Lithium Branches Grow Through Hard Ceramic
Inside a charging solid-state battery, lithium ions that have migrated from the cathode receive electrons at the anode side and revert to metallic lithium. According to SLAC and Stanford's explanation, this metallic lithium infiltrates tiny defects already present in the electrolyte—pores and grain boundaries between crystals. Since the deposited metal has volume, it pushes the surrounding ceramic outward from within. New lithium fills the widened gap, and the branch grows step by step. If it reaches the cathode, the two electrodes become connected by metal, and the battery short-circuits.
Ceramic electrolytes are hard. Hardness was expected to mechanically push back against lithium intrusion—one of the properties that made solid-state batteries promising. But hardness works both ways: once a crack forms in a hard, brittle material, stress concentrates at the tip and propagates rapidly. Dendrites, which are a problem during fast charging in liquid-electrolyte batteries, can pierce through solid electrolytes even during slow charging.
Solid-state batteries aim to improve safety by replacing flammable organic solvents with solid electrolytes, and to achieve high energy density by using metallic lithium anodes. When dendrites connect the two electrodes, both advantages are undermined. A short-circuited cell generates large internal currents and heat, diminishing the benefit of choosing non-flammable materials. The capacity gained from the metallic lithium anode also becomes unusable. Whether mass production is viable has come down to a single question: how many charge-discharge cycles can occur before branches reach the electrodes.
The paper reports that dendrites form in two stages. In early cycles, they invade from the electrolyte's surface; after many charge-discharge cycles, new nucleation sites also emerge inside the electrolyte itself. The theory that dendrites originate internally existed before, but direct evidence was scarce—this gap has now reportedly been filled through direct observation. No matter how smooth the surface is finished, branches will eventually sprout from within during long-term cycling.
For branches entering from the surface, measures that refine the electrolyte-anode interface can delay their formation. But such measures cannot reach branches that originate internally. If nucleation sites also arise from grain boundaries and pores after extended cycling, then no amount of interface engineering can push the lifespan ceiling beyond a certain point. This is where the idea of compression becomes meaningful.
A Shape-Memory Alloy Ring Bends Vertical Branches Sideways
The team compressed the electrolyte using a ring made of shape-memory alloy. Shape-memory alloys have a property of returning to a pre-programmed shape. When this return motion is constrained from outside, a correspondingly larger force is transmitted to the surroundings. The research team surrounded the electrolyte with this ring and heated it to 170°C. The force of the ring trying to contract pressed on the electrolyte from two directions parallel to the electrode surface.
The paper's title refers to this state as "biaxially stressed." A crack trying to pierce vertically between the electrodes has a fracture surface standing perpendicular to the electrode plane. In-plane compression works to press this surface together from both sides. Because the closing force counteracts the opening force, vertical propagation becomes difficult.
If compression were applied in only one direction, crack surfaces standing parallel to that direction would remain unpressed. With an escape route left open, branches could still travel vertically along that surface. But when compressed simultaneously from two perpendicular directions within the electrode plane, a vertically oriented crack surface receives compression regardless of its azimuthal direction. This is precisely why the paper's title emphasizes "biaxial."
Meanwhile, horizontal cracks running parallel to the electrode surface have a fracture plane oriented perpendicular to the direction of compression. Since these surfaces aren't pressed together, they remain open. Lithium chooses the path of least resistance, so if it cannot advance vertically, it grows horizontally instead. The "deflection" in the paper's title refers to this shift in direction. As long as branches never reach the electrode, no matter how many branches exist, no short circuit occurs.
Cells under compression developed dendrites at a density never seen in previous experiments. Yet no short circuits occurred through thousands of charge-discharge cycles, and the batteries kept working. Lead author Teng Cui noted that despite generating an unprecedented number of dendrites, the battery didn't short-circuit, and explained that the interplay between mechanical properties and electrochemistry could lead to new battery design strategies. Reducing the number of branches and protecting the battery turned out not to be the same thing.
The electrolyte used was a garnet-type ceramic with the composition Li6.6La3Zr1.6Ta0.4O12. X-ray measurements at the Stanford Synchrotron Radiation Lightsource reportedly showed that the electrolyte's crystal structure itself remained unchanged even after dendrites formed internally—the bulk crystal structure was preserved. Researchers from Kyungpook National University and Arizona State University also participated in the study, which was funded by the U.S. Department of Energy.
From Root-Cause Analysis to Design: Four Reports in Quick Succession
MIT research reported by XenoSpectrum in July 2026 identified space-charge layers forming at solid electrolyte grain boundaries as a source of dendrite generation. In April of the same year, the Max Planck Institute demonstrated a mechanism whereby hydrostatic pressure generated by lithium infiltrating cracks applies tensile stress to the solid electrolyte, promoting brittle fracture—here it's the electrolyte, not the dendrite, that breaks. Both studies focused primarily on the root causes: why branches form and under what conditions electrolytes fail.
In December 2025, Zikang Yu, Brian Sheldon, and colleagues at Brown University published a different approach in Joule. Their method creates a temperature gradient across both sides of the electrolyte, generating internal compressive stress. This reportedly suppressed dendrite penetration and tripled the critical current density. This announcement came roughly six months before SLAC's paper appeared in Nature.
Brown's method and SLAC's experiment share the same underlying idea: controlling electrochemical behavior through mechanical stress. However, the role assigned to stress differs. Brown uses compression to prevent branches from entering the electrolyte at all, while SLAC uses it to redirect branches after they've already entered. The former suppresses formation itself; the latter neutralizes the danger after allowing formation to occur.
Over roughly six months, from December 2025 to July 2026, four reports on dendrite mechanisms and control emerged in succession—three of them addressing mechanical stress. The field's center of gravity appears to be shifting from identifying causes toward incorporating mechanical stress as a design variable in batteries. Co-author and Stanford University associate professor Wendy Guo noted that these batteries could ultimately lead to higher energy density and improved reliability.
What all four studies share is that they no longer treat dendrites as purely an electrochemical problem. Alongside questions of how high current density can go and how electrolyte composition should change, a new variable has entered the discussion: what stress is acting on the material. Stress can be generated through temperature differences, external clamping, or the cell's structure itself.
What Remains to Be Verified Before Limited Rollout in 2027–2028
Toyota has announced plans for a limited rollout of EVs equipped with solid-state batteries in 2027–2028. Under a supply security plan certified by Japan's Ministry of Economy, Trade and Industry, combined production capacity for two next-generation battery projects and one solid-state battery project is set to reach 9 GWh annually by 2030, with production beginning in stages from 2026 onward—though the production scale for solid-state batteries alone has not been disclosed. QuantumScape, meanwhile, has non-exclusively licensed its technology to Volkswagen Group's PowerCo and is collaborating toward mass production at a scale of 40 GWh annually. Neither company has officially announced a specific year for the start of mass production.
Both plans assume that designs capable of suppressing dendrite-induced short circuits to factory-shipment standards will be ready in time. With only two to three years between the 2027–2028 limited rollout and 2030 mass production, behavior confirmed in laboratory-scale cells must be reproduced in the large-area cells used in vehicles within that window. As the area expands, so does the number of defects in the electrolyte—and with them, the number of sites where branches can originate. For domestic manufacturers' mass-production designs, whether uniform compression can be sustained across the entire surface will become a design consideration alongside electrolyte material selection.
In this experiment, a shape-memory alloy ring heated to 170°C maintained the compression on the electrolyte. Whether the same compression can be sustained for years within the structure of a vehicle battery's casing or pack is a separate challenge. To move toward mass production, it will be necessary to confirm that the effects of biaxial compression can be reproduced in electrolytes other than garnet, and that the compression doesn't degrade through repeated charge-discharge cycles and temperature changes.
Cui does not describe this achievement as a finished solution, saying the goal is to build a battery that is highly reliable, has high energy density, and can charge quickly—and that this research shows several of the necessary steps toward that goal. William Chueh, who leads the SLAC-Stanford Battery Center, similarly explained that these insights point toward a viable path for solving the larger challenge of energy storage. This research doesn't show how to eliminate dendrites. It shows a structure that prevents them from reaching the electrodes even as they form. The conditions for mass production may well fall into place starting from this design.