Why does throwing water on burning
deep-fryer oil send flames up to the ceiling?
When something catches fire, you throw water on it. That's usually the right call. But burning kitchen oil is the exception: the moment water hits it, the flames multiply many times over. The reason comes down to a single number: 1700.
It's evening in the kitchen. You've got a pot of oil heating up for tempura, and you glance away for just a moment. White smoke starts rising from the oil. Before you know it, the oil in the pot is on fire, flames licking up over the rim.
Your hand reaches instinctively for the kitchen tap, or a cup sitting nearby. It's a fire, so water should put it out — that's the natural thing to think.
But if you throw a cup of water onto burning oil, the flames become a column of fire reaching the ceiling. Far from going out, it scatters burning oil everywhere, and the fire spreads through the whole room.
Water doesn't stay on top of oil. It pushes through and sinks to the bottom — a place that's already over 300°C.
At the bottom it instantly flashes to steam, blasting all the burning oil above it into the air. That's what the column of flame really is.
In other words, the water isn't attacking the fire — it's exploding at the bottom of the pot. Let's go through it step by step.
Reason 1: water goes "under" the oil, not "on top of" it
Think of a bottle of salad dressing. Shake it and let it sit, and the oil rises to the top, the water settles at the bottom. Oil is lighter.
The same thing happens in a burning pot. The water you throw in doesn't float on the flames — it pushes through the oil and sinks to the bottom of the pot. This is the crucial point. If water stayed on the surface and worked like a lid, blocking oxygen, it would actually help put the fire out. Instead, it dives straight into the hottest part of the pot.
Reason 2: turning to steam makes water expand enormously
When oil is burning, its temperature is above 300°C — more than three times the 100°C at which water boils.
When water reaches that heat, there's no time to warm up gradually. It turns to steam the instant it touches. And when water becomes steam, its volume grows to roughly 1700 times its original size.
Let's translate that number into something familiar.
| Original water | 200 mL (one cup) |
| As steam (≈1700×) | about 340,000 mL = about 340 L |
| What 340 L looks like | 1.7 times a typical home bathtub (≈200 L) |
| Time it takes | almost instantaneous |
※ This estimate is calculated at 100°C and 1 atmosphere. At actual oil temperatures the expansion is even greater.
1.7 bathtubs' worth of gas wells up from the bottom of the pot in an instant, with burning oil sitting right on top of it. That oil gets blasted into the air by steam that has nowhere else to go. Turned into a fine mist in midair, the oil's surface area increases enormously, and it burns fiercely. That's the column of flame.
It goes to explode at the bottom of the pot.
When paper or wood is burning, throwing water on it is correct. In that case, the water stays on the surface of what's burning, and as it evaporates it draws off a large amount of heat, lowering the temperature. At the same time, the resulting steam blocks oxygen.
Water is still an excellent fire-fighting agent. It's only with oil that water's tendency to sink turns that same property into the exact opposite effect.
So what should you actually do?
- Turn off the heat firstSwitch off the stove. Cutting the heat supply comes first. Don't move the pot — carrying it means spilled oil can spread the fire and burn you too.
- Cut off the airUse a residential fire extinguisher rated for oil fires, or a commercial fire-smothering sheet, if you have one. Without those, slide a large pot lid over it from the near side. Lowering it straight down from above lets flames curl around your hand.
- If it's beyond you, get out and call 119If flames spread beyond the pot to the range hood or curtains, give up on fighting it. "Just a bit more and it'll be out" is the most dangerous thought you can have. When you leave, close the door behind you and alert your neighbors.
- Throw water on it — the subject of this article. Even one cup causes a column of flame
- Carry the pot somewhere else — spilled oil creates a trail of fire. Carrying it to the sink is also dangerous
- Throw or fling a wet cloth onto it — the water in the cloth triggers the same effect as pouring water directly. Fire departments do describe using a cloth, but only on the condition that it's well wrung out and laid gently from the near side. If you're not confident, don't risk it — prioritize getting out
- Throw in mayonnaise or similar — a method that spread for a while, but fire departments don't recommend it. It's mostly oil and water, and can make the fire spread or splatter further instead
- Open the lid right after the fire goes out — the oil is still hot enough to ignite. Letting air in can reignite it. Leave it covered until it's fully cooled
- Never leave deep-frying oil unattended. Most kitchen fires start when someone "just stepped away for a moment"
- White smoke is a danger signal. Even if it's not burning yet, the ignition temperature is right around the corner. Turn off the heat
- Use a stove with a built-in overheating sensor for cooking oil — standard on modern household gas stoves
- Keep a residential fire extinguisher, extinguishing spray, or smothering sheet in the kitchen. Choose one labeled for tempura-oil fires
- Check occasionally that your smoke alarm is working
Something you can check in your own kitchen (no fire involved)
- Fill a clear glass halfway with water
- Slowly pour in some cooking oil
- Watch the oil settle on top and the water below, in a clean split. Stir it and it always separates back into the same order
This "oil on top, water below" order is exactly what causes the column of flame. In a burning pot, water heads for the bottom the same way. Never try this with an actual fire. Observing the separation is all you need.
Summary
You shouldn't pour water on burning oil, not because water "doesn't work," but because of water's own nature: it's heavier than oil, so it sinks to the hottest part at the bottom, and there it instantly swells to 1700 times its volume.
Turn off the heat. Put on a lid. If that fails, get out and call 119.
Just remember that order.
A kitchen fire is a problem confined to the pot in front of you, but an outdoor fire can spread to places far away. How far flame heat reaches and how far embers can fly is calculated in the article on wildfires. And fire can start even with no flame in sight at all — how an oil-soaked cloth can ignite on its own is covered in the article on spontaneous combustion of oily rags.
Want to go deeper? ― Terms, formulas, and how this connects to the textbookLabels show which level each part belongs to, from middle-school science to open research questions
- MSCovered in middle-school science
- HSCovered in high-school "basic chemistry" / "basic physics"
- HS+Covered in high-school "chemistry" / "physics", or treated as advanced/sidebar material in textbooks
- UnivContent not covered in high school — a university-level specialty (combustion engineering, heat transfer)
- ResearchNot yet settled even as university-level fact — something researchers are actively investigating
MSTerminology: "flash point" and "fire point" are not the same
- Flash point: the temperature at which vapor coming off a liquid will ignite if a flame is brought near it. For cooking oil, this is generally said to be around 300°C.
- Fire point (autoignition point): the temperature at which the liquid ignites on its own, with no flame needed. For cooking oil, this is generally said to be around 360–370°C. This is the temperature a neglected pot of tempura oil reaches before it bursts into flame.
- White smoke: a sign that the oil has begun breaking down. It isn't on fire yet at this stage, but the fire point is close.
- Oil fires: fire extinguisher classifications distinguish fires involving paper or wood (Class A), oil (Class B), and electrical equipment (Class C). For kitchen use, choose one labeled for tempura-oil fires.
Tempura is normally fried at 170–180°C. There's still some distance to the fire point, but leave the pot heating unattended and that gap closes within a few minutes to ten-odd minutes.
MSHSWhy does water sink? ― Density
Water's density is about 1.0 g/cm³; cooking oil is roughly 0.91–0.92 g/cm³. Oil is lighter, so water settles below it and oil floats above — the middle-school concept of "density" at work.
Incidentally, heated oil becomes even lighter (volume expands with heat). At 300°C, the density gap between oil and water widens further, so water sinks even faster.
HSWorking it out with a formula: how many liters of gas from a single spoonful of water?
Everyone knows "don't pour water on burning oil." But how much actually happens? Put it in numbers and it turns out to be far more violent than you'd imagine.
When water becomes steam, its volume grows about 1700-fold
| Volume of the original water | 1 g = 1 cm³ (unit: cm³, same as 1 mL) |
| Volume after turning to steam | about 1700 cm³ (unit: cm³, 1.7 L) |
| Ratio | about 1700-fold (a plain ratio, no unit) |
In liquid water, molecules are packed tightly together, sticking to each other. As steam, those bonds come apart and the molecules fly freely, so the same amount of matter occupies vastly more space.
What matters here is that this expansion happens while pushing back against whatever is above it — in this case, all the oil, which gets shoved upward together.
| Water that slipped in by accident | 1 tablespoon = 15 mL |
| Volume once turned to steam | 15 × 1700 = 25500 mL |
| Converted to liters | 25500 ÷ 1000 = 25.5 L |
One tablespoon of water produces 25 liters of gas. That's the equivalent of twelve 2-liter bottles, generated all at once at the bottom of the pot.
All of that surges up from underneath the oil. The oil is pushed upward and scattered into the air as fine droplets. Scattered oil in midair meets plenty of oxygen, so it burns fiercely all at once. That's the column of flame. The oil hasn't "exploded" — the water has launched it.
If water sank and then warmed up slowly, there might still be time to react. In reality it's instant. Looking at how much heat is needed shows why.
| Heat needed to turn 1 g of water to steam | about 2260 J |
| Heat released as 1 g of oil cools from 300°C to 100°C | 2.0 × 200 = 400 J |
| Amount of oil needed to vaporize 1 g of water | 2260 ÷ 400 ≒ 5.7 g |
Just cooling down roughly 6 grams of oil around it is enough. The whole potful of oil doesn't need to cool. The moment they touch, the heat is supplied entirely by the oil right next to it.
So there's no time to "wait a second and think it over" after pouring it in. Sinking, boiling, and blasting upward all happen as one continuous event.
| Density of water | about 1.0 g/cm³ |
| Density of cooking oil | about 0.92 g/cm³ |
| Difference | 1.0 − 0.92 = 0.08 g/cm³ |
Being heavier, water always sinks to the bottom of the oil. And hot oil expands and becomes even lighter, so at 300°C the gap widens further and water sinks even faster.
In other words, water heads straight for the hottest spot — the bottom of the pot — and that's where ①②③ unfold. The conditions couldn't be more perfectly aligned.
If a fire breaks out, turn off the heat first and step away. Never pour water on it. What this calculation shows isn't that water "doesn't work" — it's that water makes things worse.
HSWhere does "1700-fold" come from?
This number isn't something to memorize — it's something you can calculate, using the gas concepts taught in high-school chemistry.
| Volume of 1 mol of liquid water | 18 g ÷ 1.0 g/cm³ = 18 cm³ |
| Volume of 1 mol of gas (0°C, 1 atm) | 22.4 L = 22,400 cm³ |
| Converted to 100°C (Charles's law) | 22,400 × 373 ÷ 273 ≒ 30,600 cm³ |
| Volume ratio | 30,600 ÷ 18 ≒ 1,700-fold |
In other words, "1700-fold" comes out of just two numbers: water's molar mass and the volume of one mole of gas. Calculate it at the oil's actual temperature of 300°C and the figure gets even bigger (about 2,600-fold).
HS+Why does it turn to steam "instantly"? ― Latent heat and temperature difference
Vaporizing water requires a large amount of additional heat even after it reaches 100°C (the latent heat of vaporization, about 2,260 kJ/kg). Normally this would buy some time.
But here the opponent is a large volume of oil above 300°C. With a temperature gap of over 200°C and the oil's large heat capacity, the required heat is supplied almost instantly. There's no time-buying effect at all.
What's more, a droplet of water inside the oil is completely surrounded by oil. With nowhere to escape, the steam that forms turns directly into a force pushing the oil above it upward. The same amount of water behaves very differently spilled out in the open versus boiling at the bottom of oil.
UnivThe name for this phenomenon ― slopover and boilover
In fire-fighting and combustion engineering, the violent eruption that occurs when water enters a layer of burning oil is called slopover. A related and long-studied major hazard is boilover, where water pooled at the bottom of a petroleum storage tank heats up and triggers the same kind of event in a tank fire.
Related specialties include film boiling and the Leidenfrost effect, where water droplets boil unstably on a very hot surface, and spray combustion, where the ejected oil burns as a mist — all topics covered in university-level heat transfer and combustion engineering. The finer the scattered oil droplets, the greater their surface area and the fiercer the burning.
ResearchWhat's still unresolved ― the paradox that "misted water can put out an oil fire"
Up to this point we've said "never put water on oil, full stop." But in fact, spraying water as an extremely fine mist (water mist) can extinguish an oil fire. Commercial kitchens and engine rooms do use water-based fire-suppression systems. Why does the same substance, water, produce the opposite result?
- Mist doesn't sink into the oil. When droplets are small enough, they evaporate completely within the flame and never reach the bottom. Turning to steam "before it can sink" is thought to be why the explosive upward blast doesn't occur.
- The steam dilutes the flame. A fine mist has a large surface area, draws heat away rapidly, and the resulting water vapor displaces oxygen.
- But where exactly the line falls isn't simple. Droplet size, spray force, oil depth, pot shape — combinations of these factors flip the outcome between "extinguished" and "exploded." It's impossible for an ordinary person to judge where that line sits under home-kitchen conditions.
- That's why the safety advice always stays the same: "never put water on oil." This simple rule isn't adopted because it's scientifically precise — it's adopted because the cost of misjudging the boundary is too high.
Some safety rules are deliberately simplified like this. Knowing what's behind them also helps you understand why the rule is worth following.
Connections to the textbook (by level)
| Level | Subject / unit | Where in this article |
|---|---|---|
| MS | Science — density / states of matter / combustion and oxygen | Why water sinks, why volume grows on becoming steam, why a lid puts the fire out |
| HS | Basic chemistry — moles / gas volume | Calculating the "1700-fold" figure yourself |
| HS | Basic physics — heat quantity and specific heat | How a temperature difference speeds up heat transfer |
| HS+ | Chemistry — gas laws / physics — latent heat | Temperature correction via Charles's law, latent heat of vaporization of 2,260 kJ/kg |
| Univ | Combustion engineering / heat transfer engineering | Slopover, boilover, film boiling, spray combustion |
| Research | Fire-suppression research (unresolved) | The conditions under which water mist extinguishes vs. triggers an explosion |
| ― | Disaster prevention / safety education | Turn off heat → cover it → evacuate and call 119, types of fire extinguisher |
- Public-awareness materials on "tempura oil fires" and residential fire prevention from Japan's Fire and Disaster Management Agency (総務省消防庁, Ministry of Internal Affairs and Communications Fire and Disaster Management Agency) and local fire departments.
- Reproduction experiments and advisories on overheating and ignition of cooking oil by the National Institute of Technology and Evaluation (NITE, 独立行政法人 製品評価技術基盤機構) and the National Consumer Affairs Center of Japan (国民生活センター).
- Fire extinguisher classification under Japanese fire service law (Class A = ordinary combustibles, Class B = oil fires, Class C = electrical fires).
- Explanatory materials on cooking-oil overheat prevention devices (Si sensor stoves) from the Japan Gas and Oil Appliances Inspection Association (日本ガス石油機器工業会) and related bodies.
- Standard combustion-engineering textbook descriptions of boilover/slopover, and NFPA standards on water-mist fire-suppression systems.
※ Figures for fire point, flash point, latent heat of vaporization, and similar values vary with the type of oil and the measurement conditions. This article gives commonly used approximate values.
※ This article is a general-audience science explainer. In an actual fire, follow the instructions of your local fire department and the manuals for any fire extinguisher or suppression equipment you use. Keep any experiments described here strictly fire-free. The figures given are approximate values meant to help illustrate the underlying mechanism.