MENU

Lithium Dendrites Explained | Solid-State Batteries

Back to Technology, Institution & Company Guides

TECHNOLOGY EXPLAINER

What a Lithium Dendrite Is
— why the prediction that a hard electrolyte would stop it was wrong

On every charge, lithium grows out in branches, pierces the electrolyte and shorts the cell. Solid electrolytes were supposed to stop that. The theory said so too. Yet ceramics with a shear modulus above 30 GPa are being pierced by a metal that ought to be soft.

Built from primary sources: a review in the Japan Society of Applied Physics journal (Nagoya University), and published material from Nagoya Institute of Technology, SLAC, Rice University, the University of Houston, Tohoku University and QuantumScape / Last updated September 2026

Conceptual image of a mirror-smooth, unblemished metal disc resting on a dark surface
Conceptual image (AI-generated). An impression of the ideal in which metal plates out flat. It does not represent a real lithium metal anode, its surface condition, or any observation.
What this article covers
  1. What a dendrite is (the short version)
  2. Why anyone still wants to use lithium metal
  3. The prediction that a hard solid electrolyte would stop it
  4. The prediction failed — a 30 GPa material pierced by a soft metal
  5. What is actually happening — the trigger was a void
  6. Our calculation: why a void opens at the interface
  7. A materials engineer's view (1): for lithium, room temperature is hot working
  8. Two things learned in 2026
  9. How far the countermeasures have come
  10. A materials engineer's view (2): there is still no yardstick for comparison
  11. What is still hard / Glossary / Primary sources / Claim-to-source audit
How claims are labelled in this article

Sourced = a value stated in published material from a research institute or manufacturer (link given)
Our calculation = a figure this article derived, with the assumptions spelled out
Not yet confirmed = a plan or a target with no confirmed production record
Structural readings and materials-design interpretations are marked separately as Commentary.

1. What a dendrite is (the short version)

A dendrite is lithium metal plating out in branches during charging. The word comes from the Greek for tree; in Japanese it is called tree-branch growth.

  • What happens: a review by a Nagoya University group states that there would be no problem if Li plated smoothly onto the anode current collector or the Li surface, but in practice it grows as protrusions (dendrite growth)Sourced
  • Why it is dangerous: the same review states that with repeated cycling the Li gradually grows towards the cathode, eventually touches it and causes a short circuit, and that once a short occurs, a large current flows between cathode and anode, generating Joule heatSourced
  • Where that leaves things: Li metal anodes are at present used only in primary batteries — they are still not usable in a rechargeable cellSourced

Flat deposition versus local protrusions

Concept comparison between flat lithium deposition and local protrusions that concentrate stress in a solid electrolyte
  1. Ideal: lithium deposits uniformly and the interface remains intact
  2. Risk: local protrusions concentrate current and stress and may lead to cracking or short circuit
Fig. 1 AI-generated concept illustration. The number, shape and size of protrusions and the crack location are schematic. This is not a dendrite micrograph or a reconstruction of a specific experiment.
The single most important line in this article

The premise that a solid electrolyte is hard, so it can stop dendrites, is coming apart. A ceramic with more than seven times the shear modulus of Li still shorted. And in 2026, two reports arrived in quick succession: that the branches are not a soft metal but brittle needles, and that they start inside the electrolyte.

2. Why anyone still wants to use lithium metal

If it is that dangerous, why not simply avoid it? Because of energy density.

ItemAs the published material puts it
The ceiling for today's lithium-ion batteriesFor a LIB combining lithium cobalt oxide with graphite, about 250 Wh kg⁻¹ is considered the ceiling
With a lithium metal anodeUsing Li metal at the anode allows an energy density of 300 to 500 Wh kg⁻¹, beyond the LIB
Theoretical capacityLi metal has a theoretical capacity an order of magnitude larger than graphite
PotentialThe redox potential of Li is the most negative of any single element in the periodic table
ConclusionLi metal can be called the ultimate anode active material

All Sourced (review in Oyo Buturi vol. 89 no. 4 (2020) [Source 1]).

Our calculation: how much does it change

Put the two published numbers togetherOur calculation.

  • 500 Wh/kg divided by 250 Wh/kg = 2.0 times
  • 300 Wh/kg divided by 250 Wh/kg = 1.2 times
  • Assumption: for an EV with a 500 km range, if the battery mass and everything else stay the same, that is 600 to 1,000 km

Assumptions and limits: the 250 Wh/kg is the ceiling for the LiCoO2 and graphite combination, and the 300 to 500 Wh/kg is what is described as expectedSourced. Neither is an achieved figure. And the conversion into range is a plain proportionality that holds efficiency, vehicle mass and everything outside the battery constant.

Why this matters for materials engineers: turning the anode from a material into a place

A graphite anode is a material that takes lithium ions in and out between the layers of its crystal. So the mass and volume of a container for holding the lithium are always needed.

With a lithium metal anode that container disappears. And in the design called anode-free, the cell is manufactured with no anode, with the anode created in place on the first charge. QuantumScape describes its own design as one in which the battery is manufactured anode free in a discharged state, and the anode forms in situ on the first chargeSourced.

The anode disappears from the bill of materials as a component. That is not a substitution of material but a change in the parts structure itself. The reason companies keep attacking a phenomenon as awkward as the dendrite is that the prize is that large (our commentary).

3. The prediction that a hard solid electrolyte would stop it

The hope placed in solid electrolytes was originally mechanical.

What the review says

Because an inorganic solid electrolyte is a sufficiently hard material compared with a polymer or Li metal, it is expected to serve as a separator in its own right and to prevent the growth of Li.Sourced

And there was theoretical backing: on the basis of theoretical calculation, Monroe and Newman predicted that if the separator has a shear modulus more than about twice that of Li (about 4 GPa), dendrite growth of Li can be stopped.Sourced (Oyo Buturi vol. 89 no. 4 (2020))

The prediction that hardness would stop it, and reality (shear modulus) Note: shear modulus measures stiffness against shear. All values are quoted from published material. Shear modulus of lithium metal Threshold from theory (about double) Sulfide glass electrolyte (Li2S-P2S5) LLZ (garnet-type oxide) about 4 GPa about 8 GPa 6 to 8 GPa over 30 GPa 0 10 20 30 40 shear modulus (GPa) Note: the prediction was that about twice the shear modulus of Li (about 4 GPa) would stop dendrite growth [Source 1]. Note: sulfide glass is 6 to 8 GPa and LLZ over 30 GPa. The ranking matches theory, yet both shorted [Source 1].
Fig. 2 Drawn from published values (vector drawing). The about 4 GPa for Li, the threshold of about double, the 6 to 8 GPa for Li2S-P2S5 sulfide glass and the over 30 GPa for LLZ all follow the review [Source 1]. The bar lengths illustrate those numbers and are not the materials themselves or measured data. The red dashed line is this article's rendering of the predicted threshold.
MaterialShear modulusIf the prediction held
Lithium metalabout 4 GPa—
Threshold from theoryabout 8 GPa (about twice Li)Above this it should stop
Sulfide glass (Li2S-P2S5)6 to 8 GPa (sulfides are generally below 10 GPa)Just short of it
LLZ (garnet-type oxide)over 30 GPaShould stop it comfortably

All figures are Sourced (review [Source 1]). The right-hand column is this article's framing of what the prediction implies.

4. The prediction failed — a 30 GPa material pierced by a soft metal

Here is what happened.

What the review says

On sulfides: when Nagao and co-workers plated Li on the surface of a Li2S-P2S5 sulfide glass electrolyte, they observed columnar Li penetrating the electrolyte and shorting it.Sourced

On LLZ: research in the last few years, in which Li is repeatedly plated and stripped at the Li/LLZ interface, has been reported by many researchers to short-circuit even LLZ, and it is coming to be recognised that the Monroe and Newman theory does not apply.Sourced

Why this matters for materials engineers: hardness was not the answer

A shear modulus above 30 GPa is ordinary engineering-ceramic territory. Lithium is about 4 GPa. More than a sevenfold difference, and it is still piercedSourced.

What that says is that this is not a hardness contest between material A and material B. If it were, the ratio of hardnesses would settle the outcome.

The review itself lists this as an open problem: understanding is vague as to why a material with as high a shear modulus as LLZ is pierced by a metal as soft as LiSourced.

There are areas where laying out property values and designing from them simply does not work — and that is far from unusual in materials development. Bulk property values do not explain what happens locally (our commentary).

5. What is actually happening — the trigger was a void

So what does happen? The review sets out the process currently thought to be at work. It is presented as a hypothesis: the text says the process leading to a short circuit is thought to be as followsSourced.

A hypothesized path to short circuit

Five-stage concept from lithium dissolution through atomic-vacancy buildup, void formation, current concentration and local plating to penetration and short circuit
  1. Discharge: Li dissolves and leaves atomic vacancies
  2. Vacancies that cannot diffuse away accumulate near the interface
  3. An interfacial void opens and reduces effective contact area
  4. Current concentrates at the remaining contact and overvoltage rises
  5. On the next charge, plating focuses at one point and local stress can cause penetration and short circuit
Fig. 3 AI-generated concept illustration. The five stages follow the review [Source 1]. The source itself presents this as a process thought to occur, so it is not an established mechanism. Vacancy, void and deposit sizes, counts and shapes are exaggerated for explanation and are not observations or analysis results.
Easy to miss: the culprit is discharge, not charge

Dendrites grow during charging. Yet the review places the trigger on the discharge side.

As Li dissolves, voids form and the effective interfacial area decreases, which locally increases the current density of the dissolution reaction — and the short circuit occurs, rather, in the plating immediately afterwardsSourced.

In other words: discharge punches holes in the interface, and the next charge concentrates current on the contact points that remain.

6. Our calculation: why a void opens at the interface

What does "vacancies pile up faster than they can escape" amount to? Use the figures given in the review to check the order of magnitude.

Our calculation: how far they can escape, and how much is stripped

AssumptionsOur calculation

  • The self-diffusion coefficient of lithium is 8×10⁻¹¹ cm²/s at room temperature (as given in the review)Sourced
  • Take the guide to how far diffusion carries them as L = √(D t)
  • Take the areal capacity of the electrode as 3 mAh/cm²
  • Take the density of lithium as 0.534 g/cm³, its atomic mass as 6.94 and the Faraday constant as 96,485 C/mol

Working

  • How far a vacancy can move in one hour (3,600 s): √(8×10⁻¹¹ × 3,600) = about 5.4×10⁻⁴ cm = about 5.4 µm
  • In ten minutes (600 s): about 2.2×10⁻⁴ cm = about 2.2 µm
  • The thickness lost when 3 mAh/cm² of lithium is stripped: 3 mAh = 10.8 C, so 10.8 divided by 96,485 = 1.12×10⁻⁴ mol, × 6.94 g/mol divided by 0.534 g/cm³ = 1.46×10⁻³ cm³, giving about 14.5 µm

An hour strips 14.5 µm while vacancies can move only 5.4 µm. The difference stays near the interface. That is thought to be the entrance to void formation.

Assumptions and limits: √(D t) is a guide to the order of magnitude of a diffusion distance; real vacancy behaviour depends strongly on the stress field, grain boundaries and surface condition. The areal capacity of 3 mAh/cm² is our own assumption. This calculation checks the direction of the mechanism; it does not predict the life or the critical current density of any particular cell.

Why a void opens at the interface (our calculation) A comparison for discharging over one hour 14.5 µm of lithium stripped in an hour at an areal capacity of 3 mAh/cm² Our calculation 5.4 µm vacancies can move in that hour from self-diffusion, as √(Dt) Our calculation More is stripped than can escape, so voids remain at the interface This is the loss of effective interface area that focuses current on the next plating (our commentary) Note: the assumptions and formula are in the calculation box in the text. The self-diffusion coefficient is from [Source 1]. Note: both are order-of-magnitude guides and do not predict the behaviour of a real cell.
Fig. 4 Drawn from our calculation (vector drawing). The self-diffusion coefficient of 8×10⁻¹¹ cm²/s is a value from the review [Source 1], but the 14.5 µm and 5.4 µm are both computed by this article and are not published values. The areal capacity of 3 mAh/cm² is our own assumption.

7. A materials engineer's view (1): for lithium, room temperature is hot working

One sentence in the review is written almost in passing, and it is the key to how this material behaves.

What the review says

The melting point of Li is 181 °C, so room temperature is already, in absolute terms, above half the melting point. At such temperatures, grain-boundary diffusion of atoms and motion of dislocations become active, and strain increases with time under a constant stress even below the yield stress (creep deformation).Sourced

Our calculation: what temperature is that for another metal

In materials engineering, the ratio of temperature to melting point in absolute terms is called the homologous temperature. Work out lithium's, then find the temperature that gives the same ratio in another metalOur calculation.

  • Assumption: room temperature 25 °C (298 K) and the melting point of lithium 181 °C (454 K), the melting point being the review's value
  • 298 K divided by 454 K = about 0.66
  • Assumption: the melting point of iron as 1,538 °C (1,811 K), a widely known property value taken as an assumption here with no source link
  • 1,811 K × 0.66 = about 1,190 K = about 920 °C

Room temperature for lithium corresponds to iron glowing red at close to 1,000 °C.

Assumptions and limits: equal homologous temperature does not make creep behaviour identical between metals. This is a conversion for intuition, not a claim that iron and lithium behave alike.

Why this matters for materials engineers: designing for creep at room temperature is not normal work

For a metals engineer, creep belongs to high-temperature design: turbines, boilers, furnaces. Worrying about creep at room temperature almost never happens.

In a lithium metal anode, room temperature is already the creep regimeSourced. That carries two meanings.

  • The bad side: a cell sitting on a shelf under stress keeps changing shape. The assumption that nothing happens in storage does not hold
  • The good side: apply pressure and the lithium flows and fills the voids, with no heating at all

Indeed, the review states that raising cell pressure or temperature accelerates the deformation of Li and promotes the disappearance of voids, and several groups including the authors have reported that short circuits are suppressed as a resultSourced.

All-solid-state cells are designed around being clamped because they are using this property. Clamping is not only to keep things in contact but to keep creep working and the voids filled as they appear — read that way, the design requirements for the clamping mechanism change (our commentary).

8. Two things learned in 2026

Conceptual image of a thin glass needle lying snapped in two on a dark surface
Fig. 5 Conceptual image (AI-generated). A metaphor for the property of snapping like glass. It does not depict a lithium dendrite itself, and it is neither a micrograph nor an experimental result.

(1) The branches start inside the electrolyte

In July 2026, SLAC National Accelerator Laboratory announced results from a paper published in NatureSourced. The team showed that dendrites form at internal defects within the electrolyte rather than at its surface, settling a long-running argument in the fieldSourced.

They also placed a shape-memory alloy ring around the solid electrolyte and heated it to 170 degrees Celsius so that it shrank down to compress the batterySourced. Under compression, dendrites still formed during charging, but they spread horizontally instead of verticallySourced.

(2) The branches were not soft. They were brittle needles

In March 2026, Rice University announced results from a paper published in ScienceSourced. Lithium dendrites exhibit unexpectedly high strength and brittle behavior under mechanical stress — they did not behave like bulk lithiumSourced.

The cause given is the SEI (solid electrolyte interphase) shell. The very thin SEI covering the dendrite enhances their structural rigidity and prevents the dendrites' lithium core from deforming plasticallySourced. The University of Houston release puts it as having proven they are actually brittle and snap like glassSourced.

Two findings reported in 2026 (our summary) 1 It starts at internal defects Nature (July 2026), SLAC and Stanford Dendrites start not at the surface but at defects inside the electrolyte Compressed, they grow sideways 2 The branches were brittle needles Science (March 2026), Rice and Houston A dendrite wears an SEI shell cannot flow plastically, and is brittle It snaps like glass Note: 1 follows the SLAC announcement about the Nature paper [Source 4]; 2 the Rice announcement about the Science paper [Source 5]. Note: the drawings are schematic, not observations. Strength values could not be confirmed and are not given.
Fig. 6 Conceptual diagram (vector drawing). Item 1 follows the SLAC National Accelerator Laboratory announcement [Source 4]; item 2 follows Rice University [Source 5] and the University of Houston [Source 6]. The shapes are schematic and show no distribution of internal defects or real needle shape or size. Strength values could not be confirmed in public information and are not given.
Why this matters for materials engineers: now the failed theory makes sense

The Monroe and Newman theory in Section 3 rested on the premise that lithium is a soft metal. If something soft pushes something hard, the soft one gives way — hence, a hard electrolyte should stop it.

But the 2026 reports state that the mechanical properties of a dendrite differ from those of bulk lithiumSourced. The surface SEI becomes a shell and stops the core deforming plastically. Not a soft metal but a hard, brittle composite.

Anyone who works with composites knows the picture. Put a thin hard coating on a slender core and the apparent strength rises while ductility disappears. Something of exactly that kind had been growing inside the cell, unintended (our commentary).

Put that alongside SLAC's result that the starting point is a defect inside the electrolyte, and the target for countermeasures moves. However well you finish the surface, an internal defect starts it anyway. SLAC's announcement likewise gives as design directions electrolytes with fewer internal defects and designs that build in mechanical compression from the outsetSourced.

9. How far the countermeasures have come

Where the countermeasures stand (our summary) Note: none of them is reported to have achieved complete suppression. 1 An alloy layer 2 Remove damage 3 Heat and pressure 4 Compress it 5 Change electrolyte A thin film of a metal that alloys with Li on the LLZ surface but it vanishes after hundreds of cycles not an essential solution Removing the damaged layer left by polishing with hydrochloric acid improves wetting to molten Li cycling stability improved Raising cell pressure and temperature makes Li creep and fill the voids reported by several groups A shape-memory alloy ring clamps the electrolyte Branches still form but grow sideways and do not short thousands of cycles run A chloride electrolyte (monoclinic LiAlCl4) formed by cold pressing 94% relative density stable over 70 cycles Nagoya Institute of Tech. Note: 1, 2 and 3 follow the review [Source 1]; 4 the SLAC announcement [Source 4]; 5 Nagoya Institute of Technology [Source 3]. Note: no approach reported so far has achieved complete suppression of short circuits [Source 1]. Note: putting five side by side is our own. The electrolytes and conditions differ, so this is not a like-for-like comparison.
Fig. 7 Conceptual diagram (vector drawing). Items 1, 2 and 3 follow the review [Source 1], item 4 SLAC National Accelerator Laboratory [Source 4] and item 5 Nagoya Institute of Technology [Source 3]. Setting five side by side is this article's own and is not an industry-standard classification. The material systems and evaluation conditions differ, so this is not a ranking.
ApproachWhat it doesPublished resultSource
An alloy layer at the interfaceDeposit a thin film of a metal that alloys with Li on the LLZ surface to improve wettingIt works, but after hundreds or thousands of plating and stripping cycles it disappears from the interface into the bulk Li, so it cannot be an essential solutionReview (2020)
Removing the damaged layerRemove the defect-rich damaged layer left by polishing, by immersion in aqueous hydrochloric acidWettability to molten Li improved and the cycling stability of the Li metal anode improved dramaticallyReview (2020)
Heat and pressureRaise cell pressure and temperature so that creep in the Li fills the voidsSeveral groups including the authors have reported that the disappearance of voids is promoted and short circuits suppressedReview (2020)
Compression from outsidePlace a shape-memory alloy ring around the electrolyte and heat it to 170 °C so that it shrinks and compressesDendrites still form, but they spread horizontally instead of vertically. The batteries still worked after thousands of charge cyclesSLAC National Accelerator Laboratory (August 2026)
Changing the electrolyteUse monoclinic LiAlCl4 (a chloride) and form it by cold pressing aloneRelative density 94% with solid-to-solid resistance at 7.5% (against 63% and 99.9% for the garnet-type oxide). Stable cycling over 70 cyclesNagoya Institute of Technology and others (June 2020)

All Sourced (review [Source 1], SLAC [Source 4], Nagoya Institute of Technology [Source 3]).

Why this matters for materials engineers: the other road that 5 points to

What stands out in the Nagoya Institute of Technology announcement is reaching 94% relative density through a simple, low-environmental-impact process of cold pressing aloneSourced.

The garnet-type oxide given for comparison had a relative density of 63% with solid-to-solid resistance at 99.9% of the totalSourced. In other words, almost all of the resistance was between the grains.

Of the chloride ion, the announcement states that it has a low charge density and a weak Coulomb interaction with the lithium ion, and that since its polarisability is also high, the particles can be expected to deform under pressureSourced.

Not stopping it with hardness, but bringing things into contact through ease of deformation. That is the opposite direction from the "hard enough will stop it" thinking of Section 3. Sulfides, too, were chosen for softness. In this field, situations where deformability beats hardness keep recurring (our commentary).

10. A materials engineer's view (2): there is still no yardstick for comparison

Having lined up the countermeasures, there is in fact no way to compare how much each of them achieved. The review says so plainly.

What the review says

There is no unified metric for evaluating the cycling characteristics of a Li metal anode. The magnitude of the critical current density that induces a short circuit, the number of cycles before a short, the charge passed in one cycle, the cumulative charge and so on all differ between papers, and these factors are all mutually dependent. The effect of each approach should therefore be discussed across several evaluation axes.Sourced

Why this matters for materials engineers: comparing cycle counts alone does not work

That sentence changes how published data should be read.

  • "1,000 cycles achieved" — without how much lithium was moved each time, there is nothing to compare
  • "Critical current density of X mA/cm²" — if the charge per cycle is small, the same current density shorts less readily
  • These are explicitly stated to be all mutually dependentSourced

The table in Section 9 puts "70 cycles" and "thousands of cycles" side by side, but they cannot be set against each other as better or worse. The materials differ, the charge per cycle differs, the temperature and pressure differ.

In materials development this situation — competition starting before the metric settles — is not unusual. What works then is not chasing other companies' numbers but fixing your own evaluation conditions and varying one factor at a time (our commentary).

11. What is still hard

(1) Complete suppression has not been reached

The review states that every approach reported so far has shown some effect, but none has achieved complete suppression of short circuitsSourced. The two 2026 reports (internal defects as the origin, and brittle needles) advanced the understanding of the mechanism, but at the time of writing (September 2026) no primary source claiming complete suppression of dendrite-induced short circuits could be foundNot yet confirmed.

(2) It cannot even be found

The review explains why elucidating the mechanism has been so hard: the atomic number of Li is only 3, so finding and detecting Li that is threading its way through LLZ along some unknown, minute crack is difficult. Direct evidence for the above hypothesis is therefore hard to obtain, and that is the main factor making elucidation of the short-circuit mechanism so difficult.Sourced The 2026 reports improved that situation greatly, but the same observation cannot be made inside a production cell (our commentary).

(3) No data from production products is published

QuantumScape lists resistance to dendrite formation by its ceramic separator as a feature of its technology, and reports that a 24-layer A0 prototype cell achieved more than 95% energy retention over the equivalent of more than 1,000 charge-discharge cyclesSourced. Its commercial target for energy density is 800 to 1,000 Wh/LNot yet confirmed. But these are prototype results and targets, not a confirmed record for a production product. No quantitative data from production products showing dendrite suppression could be found in published primary sources for this article.

The article in summary
  • The theory that a hard electrolyte would stop it did not hold in practice. Even LLZ, above 30 GPa, shortsSourced
  • The trigger is thought to be the void left by the discharge before, not the charge itselfSourced
  • An hour strips 14.5 µm while vacancies escape only 5.4 µm. The difference stays at the interfaceOur calculation
  • For lithium, room temperature is the creep regime, equivalent to about 920 °C for ironOur calculation, so pressure makes it flow and fill voidsSourced
  • In 2026, the starting point of the branches was reported to be defects inside the electrolyteSourced
  • In the same year, the branches were reported to be brittle needles in an SEI shell rather than a soft metalSourced
  • There is still no unified metric for comparisonSourced

12. Glossary

Dendrite
Lithium metal plating out in branches during charging, also called tree-branch growth.
Short circuit
Cathode and anode in contact. A large current flows and Joule heat is generated.
Shear modulus
A measure of stiffness against shear, in GPa. The quantity used in the theory of dendrite suppression.
LLZ (LLZO)
The garnet-type oxide solid electrolyte Li7La3Zr2O12, with a shear modulus above 30 GPa.
Atomic vacancy
A lattice site with its atom missing. Created when lithium dissolves.
Void
A gap at the interface. It reduces the effective contact area and concentrates current locally.
Critical current density
The current density above which a short is said to occur. Its value depends on the test conditions.
Creep deformation
Strain that grows with time under constant stress, even below the yield stress.
Homologous temperature
Temperature as a fraction of the melting point in absolute terms, used to compare high-temperature behaviour.
SEI
Solid electrolyte interphase. The thin layer formed by reaction between electrolyte and anode, which also becomes the shell of a dendrite.
Anode-free
Building the cell with no anode and forming one in place on the first charge.
Relative density
Actual density as a fraction of theoretical density, describing how tightly a compact is packed.

13. Primary sources

  1. Soshu Motoyama, Takayuki Yamamoto and Yasutoshi Iriyama (Nagoya University) "Elucidating the short-circuit mechanism of solid electrolytes towards higher performance in all-solid-state Li batteries", Oyo Buturi vol. 89 no. 4 (2020) 213-217 (PDF, Japanese-language page) — jstage.jst.go.jp
  2. QuantumScape "Solid State Battery Technology", technology page — quantumscape.com
  3. Nagoya Institute of Technology "A new material for safe, high-capacity all-solid-state lithium batteries: a new chloride solid electrolyte effective at suppressing lithium metal short circuits points the way to safe automotive cells", 23 June 2020 (Japanese-language release) — nitech.ac.jp
  4. SLAC National Accelerator Laboratory "Giving solid-state batteries a squeeze keeps them from short-circuiting", 28 August 2026 — slac.stanford.edu
  5. Rice University "Thorny issue plaguing lithium-ion batteries laid bare in new study", 12 March 2026 — news.rice.edu
  6. University of Houston "UH Research Reveals Lithium Dendrites Cause Battery Safety Risks", 8 April 2026 — uh.edu
  7. Tohoku University "Interface between lithium metal and a garnet-type oxide solid electrolyte formed at room temperature in a short time", 24 March 2026 (Japanese-language release) — tohoku.ac.jp

14. Claim-to-source audit

Claim in the textBasisLabel
That there would be no problem if Li plated smoothly but in practice it grows as protrusions (dendrite growth); that with repeated cycling the Li grows towards the cathode, eventually touches it and causes a short; that a short passes a large current and generates Joule heat; that Li metal anodes are at present used only in primary batteries; that the LIB ceiling is about 250 Wh kg⁻¹ and that 300 to 500 Wh kg⁻¹ can be expected with a Li metal anode; that Li metal has a theoretical capacity an order of magnitude larger than graphite, the most negative redox potential of any single element, and can be called the ultimate anode active material; that an inorganic solid electrolyte is hard enough to act as a separator and was expected to prevent Li growth; the Monroe and Newman prediction about twice the shear modulus of Li (about 4 GPa); the shear moduli of sulfide glass (6 to 8 GPa) and LLZ (over 30 GPa); the observation of columnar Li penetrating a Li2S-P2S5 glass electrolyte; that many researchers report even LLZ shorting and that the Monroe and Newman theory is coming to be seen as inapplicable; that understanding is vague as to why such a hard material is pierced by such a soft metal; the hypothesised process from dissolution to vacancies, voids, current concentration and piercing; the self-diffusion coefficient of 8×10⁻¹¹ cm²/s; the melting point of 181 °C and the account of creep; that raising cell pressure or temperature suppresses short circuits; the alloy layer and damaged-layer removal countermeasures; the statement that no approach has achieved complete suppression; the statement that there is no unified metric and that the factors are mutually dependent; and the difficulty of detecting Li because its atomic number is 3Review, Oyo Buturi vol. 89 no. 4 (2020)[Source 1] https://www.jstage.jst.go.jp/article/oubutsu/89/4/89_213/_pdfSourced
That the QuantumScape ceramic separator is described as meeting high conductivity, stability to lithium metal, resistance to dendrite formation and low interfacial impedance; that the design is anode-free, with the battery manufactured anode free in a discharged state and the anode forming in situ on the first charge; and that a 24-layer A0 prototype cell achieved more than 1,000 full charge-discharge cycle equivalents with more than 95% energy retentionQuantumScape technology page[Source 2] https://www.quantumscape.com/battery-technology/Sourced
That QuantumScape's commercial target for energy density is 800 to 1,000 Wh/L, this being a target rather than a production recordQuantumScape technology page (a value the company presents as a commercial target)[Source 2] https://www.quantumscape.com/battery-technology/Not yet confirmed
That with monoclinic LiAlCl4 a simple, low-environmental-impact process of cold pressing alone suppressed the short-circuit phenomenon that is the problem with lithium metal anodes; that it has a high relative density of 94% with an almost negligible (7.5%) solid-to-solid resistance, against 63% and 99.9% for the garnet-type oxide; that stable cycling was achieved over 70 cycles; that the chloride ion has a low charge density with weak Coulomb interaction with the lithium ion and high polarisability so that particles can be expected to deform under pressure; and the announcement date of 23 June 2020 with publication in ACS Materials LettersNagoya Institute of Technology press release, 23 June 2020[Source 3] https://www.nitech.ac.jp/news/press/2020/8406.htmlSourced
That dendrites form at internal defects within the electrolyte; that a shape-memory alloy ring heated to 170 degrees Celsius shrank down to compress the battery; that under compression dendrites still formed during charging but spread horizontally instead of vertically; that after thousands of charge cycles and the formation of many internal dendrites the batteries still worked; and that the paper appeared in NatureSLAC National Accelerator Laboratory news, 28 August 2026[Source 4] https://www6.slac.stanford.edu/news/2026-08-28-giving-solid-state-batteries-squeeze-keeps-them-short-circuitingSourced
That lithium dendrites exhibit unexpectedly high strength and brittle behavior under mechanical stress; that the SEI coating enhances their structural rigidity and prevents the dendrites' lithium core from deforming plastically; that bulk lithium is soft and ductile; that snapping produces dead lithium; and that the paper appeared in Science on 12 March 2026Rice University news, 12 March 2026[Source 5] https://news.rice.edu/news/2026/thorny-issue-plaguing-lithium-ion-batteries-laid-bare-new-studySourced
That dendrites are unexpectedly strong and brittle and have been proven to be actually brittle and snap like glass; that operando SEM observed lithium dendrites actually snapping in real time within operating solid-state cells; and that they measure just hundreds of nanometersUniversity of Houston news, 8 April 2026[Source 6] https://www.uh.edu/news-events/stories/2026/april/04082026-lithium-battery-weakness.phpSourced
That an insulating lithium carbonate layer (Li2CO3) forming on the LLZ surface is a cause of high interfacial resistanceTohoku University press release, 24 March 2026[Source 7] https://www.tohoku.ac.jp/japanese/2026/03/press20260324-02-Lithium.htmlSourced
Putting 500 divided by 250 at 2.0 times and 300 divided by 250 at 1.2 times; converting a 500 km range into 600 to 1,000 km; putting the distance a vacancy moves in an hour at √(8×10⁻¹¹ × 3,600) = about 5.4 µm and in ten minutes at about 2.2 µm; putting the thickness of 3 mAh/cm² of lithium at about 14.5 µm; computing 298 K divided by 454 K = about 0.66 and 1,811 K × 0.66 = about 920 °C; and stating that LLZ has more than seven times the shear modulus of LiOur calculation. The areal capacity of 3 mAh/cm², lithium density 0.534 g/cm³, atomic mass 6.94, Faraday constant 96,485 C/mol, room temperature 25 °C and melting point of iron 1,538 °C are all assumptions set by this article, being general property values and assumptions without source linksOur calculation
Specific values of strength or elastic modulus for the dendrites themselvesThe qualitative statements of unexpectedly high strength and brittleness were confirmed, but no specific figures could be found in published primary sources within the scope of this article, so none are givenCommentary
Any result claiming complete suppression of dendrite-induced short circuits, and quantitative data on dendrite suppression in production productsNeither could be found in published primary sources at the time of writing (September 2026) by this article, so neither is stated. The QuantumScape figures are prototype results and targetsNot yet confirmed
Detailed conditions such as current density and temperature in the Nagoya Institute of Technology 70-cycle testNo specific charge-discharge conditions could be confirmed in the announcement, so no conditions are stated hereCommentary
Grouping the countermeasures into five; the reading that bulk property values do not explain what happens locally; the framing that the anode disappears from the bill of materials; the observation about designing for creep at room temperature and the reading that clamping also serves to keep the voids filled; reading the SEI shell as a composite of a slender core with a hard coating; the framing that deformability beats hardness in recurring situations; and the suggestion to fix the evaluation conditions and vary one factor at a timeOur summary and commentary based on published content. Not views expressed by the institutionsCommentary
That Figs. 1, 2, 3, 4, 6 and 7 are explanatory concepts rather than real observations or design drawings, and that the hero image and Figs. 1, 3 and 5 are AI-generated, Fig. 5 being a metaphor for brittle snapping rather than a dendrite itselfOur noteCommentary

Last updated 21 September 2026. Sources are limited to primary material (a review in a peer-reviewed journal, official announcements from universities and national laboratories, and company technology pages). Because the article includes structural readings and materials-design interpretations, those are marked as Commentary and kept separate from sourced fact. Specific strength values for dendrites, any result achieving complete suppression of short circuits, and quantitative data from production products are not stated here because no published primary source could be confirmed. The process in Section 5 is a hypothesis that the source itself describes as what is thought to happen, not an established mechanism. All figures are explanatory concept graphics. Figs. 2, 4, 6 and 7 are vector drawings; the hero image and Figs. 1, 3 and 5 are AI-generated images, and none of them shows a real observation, micrograph or physical product.

🌐 Japanese