Why does a vacuum flask stay hot
for hours?
Pour the same hot water into a mug and a vacuum flask. The mug goes lukewarm in 30 minutes; the flask is still hot half a day later. It's not "sealing the heat in." Heat has three escape routes, and a vacuum flask blocks each one a different way.
Picture a mug of hot tea. As it cools, where does the heat go, and how does it get there?
First. It moves directly, wherever something is touching the cup. That's why your hand gets hot if you hold it.
Second. Air warmed at the surface gets lighter and rises, and cold air moves in to replace it. You can see this as rising steam.
Third. This one is invisible. Hot things give off heat even without touching anything. That's why holding your hand near a campfire feels hot even if you're not downwind.
Those three are the only ways heat escapes. Block all three, and heat has nowhere to go.
To carry heat, you need something to carry it through. A vacuum has nothing in it, so heat can't pass directly through, and there's no air to carry it either. It blocks the first and second routes at once.
The third route, "invisible heat," gets through even a vacuum. So the wall is made mirror-like, reflecting the escaping heat back where it came from.
That's all there is inside a vacuum flask — just these two ideas. A very clever, very simple tool. Let's look at each in turn.
A vacuum stops two routes at once
For heat to pass directly, it needs something to hand the baton to. In metal, atoms vibrate violently and jostle their neighbours. In air, molecules collide and pass it along.
But a vacuum has no one to hand it to. It's like a relay with no runners. So heat can't pass directly.
For the same reason, moving air can't carry heat either — there's no air there to carry it.
One trick blocks two escape routes at once. That's why a vacuum is used.
Winter clothing keeps you warm by trapping air, because air doesn't conduct heat well.
But a vacuum flask removes even that air. Air only conducts "poorly" — it doesn't conduct "not at all." If you want to go all the way, nothing at all is best. Even though both are called "insulation," the answer changes depending on how far you're trying to push it.
The third route passes straight through a vacuum
This is where a vacuum flask gets clever. "Heat given off without contact" passes through a vacuum with no trouble at all.
Remember that the sun's heat reaches Earth. Space is almost a perfect vacuum, yet we feel warmth standing in sunlight. This kind of heat needs nothing to carry it.
Everything gives off heat according to its own temperature. Hot tea is no exception — it's constantly radiating heat outward. A vacuum alone leaves this route wide open.
So a vacuum flask polishes its inner wall like a mirror. The silver shine you see when you look inside isn't decoration.
- A mirror reflects heat that hits it. Heat trying to escape gets sent back
- A mirror also gives off very little heat of its own. A well-polished metal surface can radiate dozens of times less than a black surface
A black, rough surface radiates and absorbs heat readily. A mirror-like surface does the opposite. The silver interior of a vacuum flask exists to exploit exactly this.
A vacuum flask blocks each one separately.
Even so, it cools eventually
A vacuum flask isn't perfect. Leave it half a day to a day, and it will definitely go lukewarm. That's because one spot can't be sealed.
That spot is the mouth and the stopper. It can't be made a vacuum — it's the hole you pour things in and out of, after all.
The inner and outer containers connect near the mouth. Heat gradually creeps out through this narrow neck. Most tricks for improving a flask's performance target this exact spot — making the neck thinner and longer, using materials that conduct heat poorly, hollowing out the inside of the stopper.
Some practical tips follow from this too.
- Preheat it with hot water first — pouring into a cold container wastes heat just warming the container up
- Fill it right to the top — a layer of air inside lets heat move around in there
- Open the lid as little as possible — every time you open it, the biggest leak of all opens up
- Double-glazed windows ― two panes of glass with gas sealed between them. Some also have an invisible, ultra-thin metal coating, playing exactly the same role as the flask's mirror
- Space telescope sunshields ― to hold extremely low temperatures, huge heat shields made of many stacked thin layers are used. More layers means less heat gets through by radiation
- Liquid nitrogen containers ― the original ancestor of the vacuum flask, devised in the late 19th century for experiments turning gases into liquids
- Space suits and emergency blankets ― the silver shine reflects heat radiating from the body back. The same goes for disaster-relief thermal sheets
- Vacuum insulation panels ― used in refrigerator walls and homes. They work despite being thin because of this same principle
Something you can check in your kitchen
- Pour the same amount of the same-temperature hot water into three containers: ① a plain mug ② a mug wrapped in aluminium foil, with a lid too ③ a vacuum flask
- Measure the temperature of each at 30 minutes and 60 minutes (a cooking thermometer gives accurate readings)
- See how much better ② is than ①. The foil and lid cut down radiation and convection
- Even so, ② shouldn't match ③. That gap is the effect of the "vacuum"
Comparing all three lets you see, in numbers, how much of which escape route got blocked. Since ② alone makes quite a difference, it's also a useful everyday trick. Be careful with hot water and watch out for burns. If doing this with children, it's safest to let them just read the thermometer.
In summary
A vacuum flask stays hot because it blocks all three of heat's escape routes, each in a different way. A vacuum stops "direct contact" and "carried by air" at the same time, and a mirror-like surface reflects "heat given off without contact." The only remaining leak is the mouth and stopper.
A well-made tool is usually simple.
It's just a matter of counting the routes and blocking them one by one.
How strongly "heat given off without contact" works outdoors is covered in detail in Why does your face burn by a campfire while your back stays cold?
The idea that "heat" is really the jiggling of the particles that make up matter is explored in Why does rubbing your hands together warm them up?, starting from how motion turns into jiggling.
The same vacuum found inside a flask's wall stretches as far as the eye can see in space. Whether that makes it a "cold place" is examined in Is space really cold?
Want to know more? ― Terms, formulas, and textbook connectionsFrom middle-school science to topics still being researched — the level is labelled throughout
- MSCovered in middle-school science
- HSCovered in high-school "Physics Basics"
- HS+Covered in high-school "Physics," or treated as advanced/sidebar material in textbooks
- Univ.Not covered in high school — university-level specialist content (heat transfer engineering)
- ResearchNot even settled fact at university level — something researchers are actively investigating
MSTerms: how heat travels
- Conduction: the main text's "direct contact." Passed particle to particle through a material. Metals are good at it; air is poor at it.
- Convection: the main text's "carried by air." Warmed gas or liquid moves, physically carrying the heat with it.
- Radiation (thermal radiation): the main text's "given off without contact." Travels as electromagnetic waves, so it passes through a vacuum too. This is how the sun's heat reaches us.
- Emissivity: the proportion of heat a surface readily gives off. A black, rough surface is close to 1; a polished metal can be 0.05 or lower.
- Dewar flask: the technical name for a vacuum flask, after its invention in 1892 for low-temperature experiments.
HSChecking with a formula: how many degrees does a water bottle drop in an hour?
Make it a vacuum, and both conduction and convection stop. Only radiation is left. And that radiation can be calculated.
P ＝ ε σ A (T⁴ − T₀⁴)
| P rate of heat loss | in W (J per second) |
| ε emissivity | 0–1. Smaller for a more mirror-like surface |
| σ a fixed constant | 5.67 × 10⁻⁸ |
| A surface area | in m² |
| T, T₀ inside and outside temperature | absolute temperature [K] = Celsius + 273 |
Two things matter here. One is the 4th power. Even a small temperature difference makes a big change in heat loss. The other is that ε is a simple multiplier. Make the surface a mirror and shrink ε, and heat loss shrinks by exactly that fraction.
The silver inside a vacuum flask exists to lower this ε. It isn't decoration.
| 90°C water in absolute temperature | 90 + 273 = 363 K |
| 20°C room in absolute temperature | 20 + 273 = 293 K |
| 363 to the 4th power | 363⁴ ≒ 1.74 × 10¹⁰ |
| 293 to the 4th power | 293⁴ ≒ 0.74 × 10¹⁰ |
| The difference | 1.74 − 0.74 = 1.00 (× 10¹⁰) |
| For an ordinary surface (ε = 1) | 5.67 × 10⁻⁸ × 1.00 × 10¹⁰ ≒ 567 W/m² |
| For a mirror surface (ε = 0.05) | 567 × 0.05 ≒ 28 W/m² |
A mirror alone cuts it to a twentieth. That's the entire reason the inside is silvered.
| Bottle surface area | A = about 0.06 m² |
| Rate of heat loss | 28 × 0.06 ≒ 1.7 W |
| Heat lost over 1 hour (3600 s) | 1.7 × 3600 ≒ 6120 J |
| Heat to drop 500 g of water by 1°C | 4.2 × 500 = 2100 J |
| Temperature drop in 1 hour | 6120 ÷ 2100 ≒ 2.9°C |
About 3°C an hour. That's why 90°C water poured in the morning is still hot by noon. The numbers show the leak is a tiny 1.7 W — about the power of a small torch bulb.
※ Area and emissivity are rough estimates. In practice more heat escapes via the lid and mouth than this (see the next section).
What if the same bottle held water at 60°C instead? The temperature difference is about half that at 90°C, but the heat lost isn't halved.
| 60°C in absolute temperature | 60 + 273 = 333 K |
| 333 to the 4th power | 333⁴ ≒ 1.23 × 10¹⁰ |
| Take the difference | 1.23 − 0.74 = 0.49 (× 10¹⁰) |
| Compared with the 90°C case | 0.49 ÷ 1.00 = 0.49× |
In this case it happens to come out to roughly half. But push the temperature higher, and the picture changes. The 4th power kicks in harder the higher the temperature gets. That's also why a flask cools fastest right after you pour in boiling water.
Just reading "proportional to the 4th power" tells you nothing by itself. Plug in numbers twice and compare — that's when it means something.
HS+Why the "neck" becomes the weak point
How poorly heat conducts can be treated the same way as electrical resistance. When there are several escape routes, they act like resistors wired in parallel.
In a parallel circuit, current concentrates in whichever path has the least resistance. The same holds for heat: no matter how much resistance you build into the vacuum gap, if the neck's resistance is low, heat escapes through there.
So designers make the neck thinner (less cross-sectional area), longer (more distance), and use materials that conduct heat poorly. It's a general design lesson showing up here too: the whole system is only as good as its weakest link, unless you strengthen that link.
Univ.Why stacking layers works
Spacecraft insulation uses multi-layer insulation, dozens of thin films stacked with vacuum between them.
The logic is simple. Radiated heat drops with every reflective layer added between the source and the outside. In the ideal case, adding n layers cuts it to roughly 1/(n+1). Stack 10 layers and it's down to a tenth or less.
The same idea appears in house windows. Some double-glazed units have a film on one face that reflects infrared, giving them the selective property of letting visible light through while sending heat rays back. In winter it reflects indoor heat back in; in summer it keeps the sun's heat out. It's essentially the vacuum flask's mirror, made to pick and choose by wavelength.
ResearchOpen questions remain
- A vacuum leaks away little by little. No matter how carefully it's sealed, gas seeps out of materials or in from outside. When vacuum insulation panels are used in buildings, evaluating whether performance holds up over decades is difficult. Getter materials that absorb stray gas are one workaround, but long-term data is still accumulating.
- At extremely small gaps, the usual rules change. Radiated heat transfer has long been thought to have a theoretical ceiling, but once a gap narrows below the wavelength of light, heat transfer has been predicted — and experimentally confirmed — to exceed that ceiling. This happens at the nanometre scale, where the usual vacuum-flask logic no longer applies as-is. Research into using this for power generation and cooling is underway, but practical use is still some way off.
- Devices that let heat flow only one way are also being researched. Electronics have diodes, but doing the same thing for heat is far from simple. If achieved, it could change the whole approach to insulation. Performance so far is limited, and practical use is still a distant goal.
- Holding extremely low temperatures for a long time also remains a hard problem. Space telescopes need to stay at very low temperatures for observation, combining heat shields with cooling systems. Because even a tiny heat leak can compromise observations, the design has to be extremely delicate.
Textbook connections (by level)
| Level | Subject / unit | Where in this article |
|---|---|---|
| MS | Science ― how heat travels (conduction, convection, radiation) | The three escape routes, vacuum, the mirror's role |
| HS | Physics Basics ― heat quantity and heat transfer | The 4th-power law of temperature, the effect of emissivity |
| HS+ | Physics ― thermodynamics (often advanced material) | Thermal resistance in parallel, why the neck is the weak point |
| Univ. | Heat transfer engineering, building environmental engineering | Multi-layer insulation, low-emissivity films, selective radiative properties |
| Research | Nanoscale heat transfer (unresolved) | Vacuum degradation, near-field radiation, thermal diodes |
- Incropera, F. P. et al., Fundamentals of Heat and Mass Transfer (a standard heat-transfer textbook; emissivity, multi-layer insulation).
- Dewar, J., Collected Papers on Low Temperature Research (the invention of the vacuum double-walled vessel).
- Song, B. et al., Near-field radiative thermal transport, Nature Nanotechnology 10, 253–265, 2015 (radiative heat transfer in the near field).
- NASA explanatory material on the James Webb Space Telescope's sunshield.
- Materials from the Building Material Test Center (建材試験センター, Kenzai Shiken Center) and various manufacturers on the performance of vacuum insulation panels and double glazing.
※ Figures for emissivity and insulating performance vary by material, product, and conditions. This article gives commonly used ballpark values.
※ This article is a general-audience science explainer. The figures given are approximations to aid understanding, and vary by product and conditions. When observing hot water, please take great care to avoid burns. Follow the instructions in your product's manual for actual use.