Energy Storage

Lithium Dendrite Failure

Tiny "lithium trees" pierce the separator and short the cell into thermal runaway

Lithium dendrite failure occurs when metallic lithium plates onto the anode surface during charging instead of intercalating into the graphite. The plated lithium grows tree-like protrusions ("dendrites") tens of nm to several µm long that pierce the polymer separator, short the anode to the cathode, and trigger thermal runaway — internal temperature can climb past 800 °C in seconds. Triggered by fast-charging, low-temperature charging (<0 °C), aged solid-electrolyte interphase (SEI), or manufacturing defects. Solid-state batteries with ceramic electrolytes (LLZO, Li-PS) aim to mechanically suppress dendrite growth.

  • Dendrite onsetCharging at C/2 below 0 °C, or aged cells at any C-rate
  • Dendrite size50 nm initial → multi-µm
  • Separator16–25 µm polyolefin (PE/PP), pierced once
  • Thermal runaway temp~150 °C trigger, peaks 800 °C+
  • SuppressionSolid-state, lithium-metal foil, surface coatings
  • First documented1970s lithium-metal cells

Interactive visualization

Press play, or step through manually. The visualization is yours to drive — try it before reading on.

Open visualization fullscreen ↗

Watch the 60-second explainer

A condensed visual walkthrough — narrated, captioned, under a minute.

Why dendrite failure matters

  • Phone fires. Roughly 25,000 lithium-ion device incidents per year reach U.S. fire departments; dendrite shorts are the dominant internal-fault root cause.
  • EV battery recalls. The 2020–2021 Chevy Bolt recall pulled 142,000 vehicles after dendrite-driven shorts in LG Chem cells; the cost exceeded $1.9 billion.
  • Boeing 787 grounding. In January 2013 the FAA grounded all Dreamliners after two thermal runaway events in the auxiliary lithium-ion battery; root cause traced to internal short-circuits consistent with dendrite growth.
  • eVTOL safety. Electric vertical-takeoff aircraft demand 6 to 8 C discharge rates and millions of cycles; dendrite-suppressing chemistries (silicon-anode, solid-state) are gating certification.
  • Grid storage. 100 MWh battery installations need cell-to-cell propagation barriers; one dendrite short can ignite an entire container.
  • Cycle-life economics. Even non-catastrophic dendrites trap "dead lithium" — lithium isolated by SEI rupture — which permanently lowers capacity by 1 to 3 percent per fast-charge cycle.
  • Cargo aviation. The IATA bans bulk lithium-metal shipments and restricts lithium-ion to 30 percent state-of-charge precisely to limit dendrite-driven cargo fires.

Common misconceptions

  • "All lithium-ion fires are dendrites." Many are not. External shorts, manufacturing defects (metal swarf, foil burrs, misaligned tabs), and crush/puncture damage cause most field fires. Dendrite-driven shorts are a specific internal failure mode that requires growth over many cycles.
  • "The BMS prevents dendrites." A battery management system measures pack voltage, current, and external temperature. It cannot see the microstructure inside a cell. By the time a BMS detects an internal short, the cell is already in runaway. BMSes prevent overcharge and overdischarge, not microstructural failure.
  • "Every fast-charge causes dendrites." Properly thermally managed fast-charging at 25 to 40 °C with current tapered to anode capacity does not plate lithium. Modern EVs charge from 10 to 80 percent in under 20 minutes without dendrite formation, because thermal control and current schedules respect the diffusion limits.
  • "Solid-state is dendrite-free." Ceramic electrolytes (LLZO, sulfide glasses) actually develop a different failure: dendrites propagate along grain boundaries and pre-existing microcracks in the ceramic. Lab cells routinely fail at < 1 mA/cm² because of this. Defeating intergranular dendrites is the open problem in 2026.
  • "Dendrites only form in lithium-metal anodes." Graphite anodes plate metallic lithium when local concentration saturates. Once plated, that metal grows dendritically just like lithium-metal cells. Most field failures are from graphite-anode cells, not lithium-metal.
  • "Cooling the cell stops runaway." Once initiated past ~150 °C, the exothermic chain (SEI breakdown → separator melt → cathode oxygen release → electrolyte combustion) self-sustains internally. External cooling can prevent propagation to neighbor cells but cannot reverse runaway in the source cell.
  • "Dendrites are visible." They are not. Optical microscopy through opaque electrodes is impossible; standard SEM uses room-temperature samples that volatilize the lithium. Cryo-electron microscopy at -170 °C, pioneered by Yi Cui's group at Stanford in 2017, is the only direct imaging method.

Frequently asked questions

What is a lithium dendrite, physically?

A dendrite is a needle-like or branched filament of metallic lithium (Li-0) that grows from the anode surface during charging. Instead of lithium ions slipping between graphite layers (intercalation), they accept an electron at the surface and deposit as solid metal. Initial nuclei are 50 to 100 nm wide. Continued plating grows them outward into mossy or branched structures that can reach several micrometers, comparable to the 16 to 25 µm separator thickness. Cryo-electron microscopy reveals two morphologies: "mossy" lithium (rounded grains) and "whisker" lithium (sharp filaments). The whiskers are the dangerous form.

Why does cold charging cause dendrites?

Below 0 °C the diffusion coefficient of lithium in graphite drops by an order of magnitude. Ions arrive at the anode surface faster than they can diffuse into the bulk. The local lithium concentration at the surface saturates and the electrochemical potential drives plating instead of intercalation. The threshold depends on charge rate: at C/2 below 0 °C, plating begins; at 1C, plating begins below 10 °C. This is why phones disable fast-charging in cold weather and electric vehicle BMS systems heat the pack before high-current charging.

How does the separator fail?

Polyolefin separators (16 to 25 µm of polyethylene or polypropylene with 40 to 50 percent porosity) have ~100 nm pores. A growing dendrite enters a pore and applies localized mechanical stress. Once the dendrite tip reaches the cathode side, electrons flow directly from anode to cathode, bypassing the external circuit. The short-circuit current can exceed 100 A in a few cubic millimeters. Joule heating melts the polymer (PE softens at 130 °C, PP at 165 °C) and the pierced region grows. Ceramic-coated separators add Al2O3 layers to slow this propagation but do not stop a determined dendrite.

What is thermal runaway and why is it self-sustaining?

Thermal runaway is a positive feedback loop. Above ~80 °C the SEI layer begins to decompose exothermically. By ~120 °C the separator melts. Above ~150 °C the cathode (NMC, LCO, or NCA) decomposes, releasing oxygen. The released oxygen oxidizes the lithiated graphite anode and the carbonate electrolyte, releasing more heat. Each reaction raises temperature, which accelerates the next. Peak temperatures inside the cell exceed 800 °C and can reach 1000 °C. Once initiated, no external intervention stops it — the cell vents flammable gases (H2, CO, methane, ethylene) and burns until the lithium and carbonate are consumed. Adjacent cells in a pack ignite within seconds; this is propagation thermal runaway.

Why don't all cells fail this way?

Three layers of defense. First, the SEI layer (a thin solid film of Li2CO3, LiF, ROCO2Li that forms during the first few cycles) blocks direct lithium plating in normal conditions. Second, charging current is matched to the anode's intercalation capacity — manufacturers spec maximum C-rates at given temperatures. Third, the BMS monitors voltage and temperature and throttles. Failure occurs when all three break down: an aged SEI cracks, exposing fresh lithium-philic graphite; the cell is charged outside its temperature window; manufacturing defects (metal particles, electrode misalignment, pinhole separator) create localized hot spots. Roughly 1 in 10 million cells fails this way over its lifetime — but with billions of cells in service, that's hundreds per year.

Will solid-state batteries solve this?

Solid-state batteries replace the liquid carbonate electrolyte with a ceramic (LLZO garnet, Li7P3S11 sulfide) or polymer. The ceramic is mechanically rigid (Young's modulus 100+ GPa for LLZO) and is meant to physically block dendrites. In practice dendrites still propagate through ceramic electrolytes by following grain boundaries and microcracks; this "intergranular dendrite" problem is the major hurdle to solid-state commercialization in 2026. Sulfide electrolytes are softer and more workable but react with moist air to release H2S. Toyota, QuantumScape, and Solid Power have prototype cells; mass production for EVs is targeted for 2027 to 2030.