Why Do Alkaline Batteries Leak? A Deep Dive Into the Chemistry

Open a remote control that has sat untouched for a year and you may find the familiar white crust caked around the battery terminals. Our Complete guide to why batteries leak explains what that crust is — potassium carbonate, formed when escaped electrolyte meets carbon dioxide in the air — and what to do about it. This Deep dive answers the harder question underneath it: what exactly happens inside the sealed cell, in what order, and how much pressure does it take before the seal gives way? The answers come from a study that built a working AA battery with a transparent wall and filmed it gassing, and from the century-long trade-off that made modern alkaline cells both mercury-free and harder to keep sealed.

A battery is chemically unstable from the day it is made

The uncomfortable starting point is that an alkaline battery is thermodynamically unstable from the moment it leaves the factory. Its anode is powdered zinc sitting in concentrated potassium hydroxide, and zinc reacts spontaneously with water in that electrolyte:

Zn + 2H₂O → Zn(OH)₂ + H₂↑

The reaction releases hydrogen gas, and its driving force is large: the Gibbs free energy change is roughly −90 kJ per mole of zinc, with the hydrogen evolution reaction sitting near −0.92 V against the standard hydrogen electrode. In plain terms, the zinc inside every AA battery “wants” to corrode and release hydrogen, all the time, whether the battery is powering a device or sitting in a drawer. What stops it is kinetics, not thermodynamics. Hydrogen evolves only very slowly on a clean zinc surface because the reaction needs a large overvoltage — extra voltage beyond the thermodynamic minimum — to proceed at any real rate. The cell survives on that kinetic bottleneck. Everything that follows in this guide is about what happens when the bottleneck narrows, widens, or fails.

The slow version of that failure is self-discharge. Even at room temperature, a modern alkaline cell loses roughly 2–3% of its capacity per year to parasitic reactions: Duracell’s own data puts a stored cell at 93–96% of shipped capacity after one year at room temperature, and about 85% after four years at 70 °F (21 °C). Energizer rates its MAX AA, AAA, C and D cells for up to ten years in storage. But “slow” is not “zero” — the parasitic chemistry never stops, and the hydrogen it produces has nowhere to go inside a sealed steel can.

The experiment that filmed a battery gassing

Until recently, the inside of a discharging alkaline cell was largely inferred rather than seen. That changed in 2018, when a team from the University of South Carolina, the University of Connecticut, Duracell and Sandia National Laboratories published a study that built two research versions of an AA cell: one with a transparent wall so the anode could be watched through a microscope during discharge, and one fitted with a pressure transducer that recorded the gas pressure inside the cell every ten seconds. The cells had the same geometry, zinc-to-manganese-dioxide ratio and electrolyte as a commercial AA. The paper — Faegh and colleagues, Journal of The Electrochemical Society, 2018 — is open access, and it remains the clearest window into why alkaline cells rupture. The full paper is available via its DOI.

The team discharged cells at different loads for 96 hours, then let them rest at open circuit for 24 hours, watching the anode and logging pressure throughout. What they found overturns the simple story that a battery leaks because gas builds up steadily while it is used: the gas does not build up steadily, and the most dangerous moment is not during discharge at all — it is after the discharge stops.

The numbers: pressure stays flat, then jumps at rest

At a moderate 400-ohm load — the kind of gentle drain a remote control or a clock puts on a cell — the pressure inside the research cell rose from 35 to 48 pounds per square inch gauge (psig) over 96 hours of discharge, then crept from 48 to 58 psig during the 24-hour rest. Cell voltage fell slowly from 1.60 V to 1.29 V and recovered to 1.36 V at rest. That is the benign case: a modest pressure rise, a seal that holds, a battery that simply goes flat.

At a 40-ohm load — a heavy, deep discharge — the picture was entirely different. During the 96-hour discharge the hydrogen pressure inside the cell stayed almost flat. The cell voltage collapsed from 1.60 V to 0.07 V, but the pressure barely moved. The reason, the authors show, is that while current is flowing the anode potential is high enough to make hydrogen evolution unfavourable: the cell is too busy discharging to gas. Then the load was removed and the cell was allowed to rest back to its open-circuit voltage. Within hours the pressure shot from 36 psig to 193 psig — more than a fivefold increase — as the cell sat doing nothing. The voltage recovered to 1.05 V, but the pressure did not come back down. In a commercial cell, that is the pressure that forces the seal open and pushes electrolyte out.

This is the paper’s most useful finding for anyone who has found a leak: the gas that breaks the seal is mostly made after the battery is already dead, while it rests. It explains why the worst leaks turn up in devices nobody has touched for months. The damage happens in the quiet period, not during use.

Black zinc: the corrosion accelerator that discharge creates

What drives that sudden pressure spike? Through the transparent wall, the researchers watched something discharge creates and rest destroys: a ring of black zinc.

During discharge, zinc does not simply dissolve evenly. The team showed that electrons in the zinc slurry anode are transported by what they call a redox electrolyte mechanism: zincate ions — Zn(OH)₄²⁻ — migrate in the opposite direction to the current flow, and deposit fresh zinc onto reacting particles near the separator. Hydroxide moves the other way, four ions per zincate, to balance the charge. The result is steep concentration gradients across the anode: zones rich in zincate and poor in hydroxide and water, where ion mobility falls. In those zones, the deposited zinc grows as a mossy, high-surface-area black deposit instead of a smooth metal surface.

Under the microscope, the sequence at a 40-ohm load was precise. After about 18 hours of discharge, a dark black ring appeared in the anode column, migrating from the separator interface towards the centre of the cell. Around it, white zinc oxide formed — first a loose, porous Type I ZnO, then a coherent, passivating Type II ZnO layer that choked off the cell and sent the voltage into its steep decline. The black ring persisted to the end of the 96-hour discharge. Then, the moment the cell was allowed to rest, the ring vanished — corroded away in hours — and the pressure jumped from 36 to 193 psig.

The numbers explain why. Scanning electron microscopy showed the black zinc deposits have roughly 10,000 times the surface area of the original undischarged zinc powder. Surface area is what corrosion feeds on: ten thousand times the area means ten thousand times the sites where zinc can react with the electrolyte and release hydrogen. The black zinc is, in effect, a corrosion accelerator manufactured by the discharge itself, which then detonates — chemically speaking — the moment the cell rests and the anode potential falls back to open circuit.

Black zinc formation and cell passivation are locked together: across loads from 4 to 400 ohms, black zinc appeared after consistently about 79% of the charge passed before passivation. Passivation arrived at 43% depth of discharge after 8 hours at a 4-ohm load, 72% after 22 hours at 40 ohms, and only after 220 hours at 400 ohms. Slower discharge gives the zincate gradients time to relax by diffusion, delaying both — one more reason a gently used battery lasts longer than the same cell hammered by a high-drain device.

Why mercury left, and what replaced it

There is a second, older story behind the leak, and it explains why modern batteries are in some ways more leak-prone than the ones your grandparents used. For decades, battery makers protected zinc by amalgamating it with mercury. Mercury coats the zinc particles and raises the hydrogen overvoltage still further, choking off the parasitic corrosion reaction. Historical formulations used mercury at up to around 10% of the weight of the amalgamated zinc — an extremely effective corrosion inhibitor.

It was also an environmental disaster in waiting. Mercury from discarded batteries leaches into soil and water, and by the 1980s the pressure to remove it was decisive. Manufacturers cut mercury content to 250 parts per million, then to zero, over roughly a decade from the 1980s to the 1990s. The replacement was a package of substitutes: zinc alloyed with small amounts of indium, bismuth and aluminium, plus inhibitors added to the electrolyte such as bismuth oxide, and surfactants that coat the zinc surface. A 1996 US patent for a mercury-free alkaline cell describes the core of the approach — bismuth compounds in the electrolyte that capture impurity metals such as iron and nickel, which would otherwise set up tiny galvanic cells on the zinc surface and accelerate corrosion.

The trade-off is real and documented in the patent literature itself: in a mercury-free cell, hydrogen generation from trace impurities is “rapidly accelerated”, and leak resistance after partial discharge is harder to secure than with mercury. The modern alkaline battery is therefore a compromise — far cleaner to throw away than its mercury-bearing ancestor, but protected by a more delicate package of alloys and inhibitors that deep discharge, heat or age can overwhelm. European law now caps mercury in batteries at 0.0005% by weight under Directive 2006/66/EC, so there is no going back; the problem is managed, not solved.

Where the pressure goes: the seal

Every alkaline cell is built with a deliberate weak point: the seal at the negative end, where a plastic gasket is crimped into the steel can. It is the weakest point by design, so that pressure escapes there rather than bursting the can itself. When the internal pressure exceeds what the gasket can hold, the seal opens and potassium hydroxide oozes out with the escaping gas. Duracell’s safety data sheet puts the expected leakage at 2–20 millilitres depending on battery size — a small volume, but potassium hydroxide is caustic enough that even a few millilitres will eat into contacts and circuit boards.

Once outside, the chemistry is quick and visible. Potassium hydroxide reacts with carbon dioxide in the air:

2KOH + CO₂ → K₂CO₃ + H₂O

The product, potassium carbonate, is the white, feathery crust — a stable, crystalline solid one battery expert described to Consumer Reports as “essentially like rock salt”. It is less caustic than the potassium hydroxide it came from, but it is hygroscopic: it draws moisture from the air, stays damp, and keeps corroding the metal underneath for weeks after the leak has stopped. Green or blue tints in the crust are copper corrosion products, where leaked electrolyte has attacked copper terminals or wiring.

The same failure in other chemistries

Alkaline cells are not the only ones that leak, but the mechanism changes with the chemistry, and the differences are instructive.

Zinc–carbon. The cheapest household cells are the most leak-prone of all, for a structural reason: the zinc outer can is itself the anode, so the casing is consumed as the battery delivers power. After a few years, or once the cell is fully discharged, the wall thins enough to develop pinholes, and the ammonium chloride and zinc chloride electrolyte seeps out through the metal itself. There is no seal to fail because the container is the fuel.

Lithium-ion. The batteries in phones and laptops fail differently, and more seriously. Their electrolyte is a flammable organic solvent, not a water-based alkali, and their gas comes from the breakdown of a protective layer called the solid electrolyte interphase (SEI) on the graphite anode. When a lithium-ion cell is deeply discharged for a long time, overheated or overcharged, the SEI breaks down and the electrolyte decomposes into gases — mostly carbon monoxide and carbon dioxide, with hydrogen and light alkanes. Because the cell is sealed, the pouch or casing swells. A 2025 review in the journal Batteries found that swelling in lithium-ion cells ranges from a few percent during normal cycling to more than 45% in severely aged cells. The review, by Maddipatla and colleagues, is open access. Unlike an alkaline leak, a swollen lithium-ion battery is a fire risk, not a cleaning job: stop using it, do not charge or puncture it, and take it to a battery collection point. The contrast is the point: alkaline cells leak because zinc corrodes; lithium-ion cells swell because their protective layer decomposes. Same symptom family, different chemistry, very different danger.

Temperature, time and the odds of a leak

Three numbers govern whether a given battery leaks, and all three point the same way.

Temperature. Reaction rates roughly double for every 10 °C rise — the Arrhenius rule of thumb — so a battery stored at 30 °C ages about twice as fast as one at 20 °C, and one left in a hot car in summer several times faster. Shelf life roughly halves for every 10 °C above room temperature; cool storage is the cheapest leak insurance there is.

Time. A new alkaline cell self-discharges at 2–3% of capacity per year at room temperature, so after five years in a drawer it has quietly lost a tenth or more of its charge to parasitic reactions — and produced the corresponding hydrogen. The risk curve is not linear: it rises steeply after the expiry date and after the cell is spent, because a discharged cell has less zinc left, more zinc oxide, and the conditions that favour black zinc formation.

Depth of discharge. The 2018 study’s central result is that deep discharge is what creates the corrosion accelerator. A cell discharged gently to 1.29 V and rested saw its pressure rise by 10 psig; a cell driven to 0.07 V and rested saw its pressure rise by 157 psig. The practical translation is blunt: the single most dangerous thing you can do to an alkaline battery is to leave it in a device until it is completely dead, and then leave it there. A device switched off but still drawing a trickle current — a remote control, a clock, a seasonal decoration — is the classic case, because it drives the cell past empty slowly and then leaves it to rest.

What the research still cannot tell you

An honest Deep dive has to say where the evidence stops. Three limits matter.

First, the 2018 study used research cells, not shop-bought AAs. They were built to the same geometry and chemistry, but they had a transparent wall and a pressure transducer, and they were discharged under laboratory loads. Commercial cells add manufacturing variation — seal quality, impurity levels, alloy composition — that the study deliberately controlled away. The mechanism is almost certainly the same in your remote control; the exact pressures are not.

Second, nobody publishes the leak rate of modern alkaline cells. Manufacturers do not release figures for what fraction leak in storage or in devices, so online statistics should be treated with suspicion. One widely repeated claim — that up to 20% of unused batteries leak after five years — traces to content-farm articles quoting experts whose affiliations do not check out; we do not repeat it because we cannot verify it. What can be verified is the direction: risk rises with age, heat and depth of discharge, and steeply after expiry.

Third, the corrosion-inhibitor package in a modern cell is proprietary. Alloy composition, inhibitor concentrations and seal design differ between brands and are trade secrets — which is why independent testing consistently finds premium brands outlast economy brands in storage. The science explains the failure; it cannot tell you which brand’s package will hold longest in your drawer, because that information is not public.

None of this makes the mechanism uncertain. The thermodynamic instability of zinc in potassium hydroxide, the black zinc accelerator, and the rest-period pressure spike are all measured, published and replicated. What is uncertain is only the exact odds for a particular cell on a particular shelf — and that is precisely why the prevention advice is the same regardless of brand: take batteries out of devices you are not using, respect the expiry date, store them cool and dry, and never let a cell sit dead in a drawer.

The short version

• An alkaline battery is thermodynamically unstable from the day it is made: zinc reacts spontaneously with water in the electrolyte, releasing hydrogen. Only slow kinetics keeps it alive.
• Discharge creates the problem it later dies from. Steep zincate gradients deposit mossy “black zinc” with roughly 10,000 times the surface area of the original zinc powder.
• The pressure spike happens at rest, not during use. In the 2018 operando study, a deeply discharged cell’s internal pressure jumped from 36 to 193 psig after the load was removed, as the black zinc corroded away.
• Mercury once suppressed this corrosion almost completely, at up to ~10% of the amalgamated zinc’s weight. Its removal in the 1980s–1990s was an environmental necessity, replaced by indium, bismuth and aluminium alloys — a more delicate package that deep discharge can overwhelm.
• The seal at the negative end is the designed weak point. When it gives way, potassium hydroxide escapes and reacts with carbon dioxide in the air to form the white potassium carbonate crust.
• Heat, time and depth of discharge set the odds: reaction rates roughly double per 10 °C, self-discharge runs at 2–3% a year, and a cell left dead in a device is the highest-risk case of all.

For the practical side — whether the crust is dangerous, how to clean it off contacts, how to store batteries, and how to dispose of a leaking cell — see our Complete guide: Why Do Batteries Leak? The Chemistry Behind the White Crust.

Sources

1. Faegh, E., Omasta, T., Hull, M., Ferrin, S., Shrestha, S., Lechman, J., Bolintineanu, D., Zuraw, M., & Mustain, W. E. (2018). Understanding the Dynamics of Primary Zn-MnO₂ Alkaline Battery Gassing with Operando Visualization and Pressure Cells. Journal of The Electrochemical Society, 165(11), A2528–A2535. doi.org/10.1149/2.0321811jes (open access, CC BY 4.0). Also available as PDF from IOP Publishing and indexed at OSTI (ID 1529287).
2. US Patent 5,541,021 — Alkaline cell (mercury-free, bismuth compounds in the electrolyte), issued 30 July 1996. patents.google.com/patent/US5541021A.
3. Directive 2006/66/EC of the European Parliament and of the Council on batteries and accumulators (mercury limit 0.0005% by weight). eur-lex.europa.eu.
4. Maddipatla, S., Rauf, H., Osterman, M., Arshad, N., & Pecht, M. (2025). Swelling Mechanisms, Diagnostic Applications, and Mitigation Strategies in Lithium-Ion Batteries. Batteries, 11(10), 356. doi.org/10.3390/batteries11100356 (open access).
5. iFixit. “Why Is Your Old Phone Battery Swollen?” ifixit.com — SEI breakdown and deep-discharge swelling in lithium-ion pouch cells.
6. US Environmental Protection Agency. “Used Household Batteries.” epa.gov — disposal guidance by chemistry.
7. Duracell. Safety Data Sheet, alkaline manganese dioxide batteries (composition ~35–40% manganese dioxide, 10–25% zinc, 5–10% potassium hydroxide; expected leakage 2–20 mL by size). Cited as a manufacturer document; see also the Complete guide.
8. Energizer. MAX alkaline battery storage claims (AA/AAA/C/D up to 10 years in storage; 9V up to 5 years). Manufacturer product data.