Fifteen million degrees, and then it gets hotter

First, the thing that trips everyone up

Three different things are all called "temperature"

If you skip this section, the rest of the article will look like it contradicts itself. It is the single most important idea here.

In everyday life, temperature and heat feel like the same thing. A 200 °C oven cooks a chicken; a 200 °C anything will burn you. That intuition works because everything you normally touch is dense — packed with particles. It stops working completely once you leave the Sun's surface.

Physicists actually mean this: temperature is the average kinetic energy of the particles in a substance — essentially, how fast they are jiggling. A gas whose molecules are moving quickly is hot. That is the whole definition. Note what it does not mention: how many particles there are.

That omission is where the confusion lives. Consider two rooms. One contains air at 1,000 °C — a furnace that would kill you in seconds. The other contains six atoms at 1,000,000 °C. The second room is a thousand times "hotter" by the physicists' definition, and you would not notice it at all. Six atoms, however fast they move, carry almost no total energy. They cannot warm you because there is nothing of them to transfer.

This is exactly the Sun's outer atmosphere. It is genuinely at millions of degrees, and it is genuinely so thin that a spacecraft can fly through it without melting.

One more piece of housekeeping. Scientists measure these things in kelvin, not Celsius. A kelvin is the same size as a Celsius degree; the scale simply starts at absolute zero instead of at the freezing point of water, so K = °C + 273.15. At solar temperatures the difference is irrelevant — 15,700,000 K and 15,699,727 °C are the same claim. It only matters near the cold end, where the difference between 3 K and 3 °C is the difference between deep space and a chilly afternoon. This article gives Celsius, as you asked, and kelvin where the source literature would.

Finally, the Sun has a natural ruler. Its radius — centre to visible surface — is 695,700 km, written R☉ and read "R‑sun". Astronomers describe positions inside the Sun as fractions of it: 0.5 R☉ is halfway out. Distances beyond the Sun use the astronomical unit: 1 AU = 149.6 million km, the average Earth–Sun distance, which is where this journey ends.

The whole journey on two charts

The map

Here is the entire answer before we walk through it. The first chart covers the inside of the Sun; the second covers everything above the surface. They use different horizontal scales because the interior spans 700,000 km and the atmosphere spans a thousand times more.

Notice the shape: down, then sharply up, then slowly down again. Everything from here on is an explanation of one of those three movements.

The Sun is enormous and gravity pulls all of it inward. Nothing in the middle is holding it up structurally — there is no rock, no core of iron. What stops the Sun collapsing is pressure, and to generate enough pressure to support 330,000 Earths' worth of material, the gas at the centre has to be compressed to roughly a hundred and fifty times the density of water and heated to about 15.7 million °C.

Compress a gas and it gets hotter — the same effect that warms a bicycle pump. The Sun's core temperature is, at bottom, the temperature you inevitably reach when you pile that much hydrogen into one place. It is a consequence of the Sun's mass rather than something the Sun chose.

At those temperatures matter is plasma: so hot that electrons are stripped away from their nuclei entirely, leaving a soup of bare protons and free electrons. There are no atoms here in any ordinary sense, and no molecules, and nothing that could be called solid, liquid or gas as you know them. Despite being thirteen times denser than lead, this material is not remotely like a metal — it is a gas whose particles simply have nowhere to go.

Fifteen million degrees happens to be the threshold for the thing the Sun is famous for. Protons repel each other electrically, and hard. Only when they are moving fast enough — which is to say, only when the temperature is high enough — can two of them collide violently enough to stick, beginning a chain of reactions that ultimately fuses four hydrogen nuclei into one helium nucleus and releases energy. This is the proton–proton chain, and it runs at about 3.8 × 10²⁶ watts.

The reaction is exquisitely sensitive to temperature, roughly proportional to T⁴ in this regime, which is why fusion effectively switches off outside the innermost quarter of the Sun. Cross 0.25 R☉, where the temperature has fallen to around 7.6 million °C, and you are outside the furnace: 99% of the Sun's energy is generated inside that boundary, in a volume that is only about 1.5% of the Sun's total.

The energy made in the core has to get out, and across this vast middle region it travels as light. Not usefully, though. The plasma here is so dense and so opaque that a photon travels only a fraction of a centimetre before it is absorbed by an electron and re‑emitted in a random new direction.

Repeat that a staggering number of times and you get a random walk: energy staggering outward drunkenly rather than flying straight. Estimates of how long the journey takes vary with the assumptions used, but they run from tens of thousands of years to a few hundred thousand. Sunlight falling on your hand today carries energy that was released by fusion long before there were humans to be warmed by it.

The temperature falls across this zone from about 7.6 million to 2.2 million °C — a huge drop in absolute terms, but a shallow one relative to the distance, which is exactly what lets radiation do the job. Heat only needs to trickle. The layer is stable, unmixed and quiet, and it holds most of the Sun's mass.

This boundary deserves a mention even though nothing dramatic happens to the temperature. Below it, the Sun rotates as a solid body — everything turning together. Above it, rotation becomes latitude‑dependent: the equator completes a turn in about 25 days while the poles take closer to 35.

Two regions rotating at different speeds, sliding past each other in an electrically conducting fluid, is a recipe for stretching and amplifying magnetic fields. The prevailing view is that the Sun's magnetic cycle — sunspots, flares, the eleven‑year rhythm of solar activity — is generated at or near this shear layer. Keep it in mind: almost everything strange that happens to temperature later in this journey is magnetic in origin, and this is plausibly where those magnetic fields are born.

Around 0.71 R☉ something changes. The plasma has cooled enough that electrons begin to recombine with nuclei, forming partial atoms — and partial atoms are much better at absorbing light than bare nuclei are. The material turns opaque. Radiation can no longer carry the energy outward fast enough.

So the Sun switches transport mechanism. Hot plasma physically rises, dumps its heat near the surface, cools, and sinks — the same overturning motion you see in a pan of simmering water or a lava lamp. This is convection, and it is far more efficient than the photon random walk, which is why the temperature now plummets: from 2.2 million °C at the bottom of this zone to 5,500 °C at the top. Ninety‑nine point seven percent of the remaining temperature is lost across the final third of the radius.

You can see the tops of these convection cells directly. A high‑resolution image of the Sun shows granulation — a shifting mosaic of bright cells with dark boundaries, each roughly 1,000 km across, each living about ten minutes. The bright centres are hot plasma arriving from below; the dark lanes are cooler plasma sinking back down. The surface of the Sun is, quite literally, boiling.

The Sun has no surface in the sense that Earth has one. There is no boundary you could stand on, no place where solid becomes gas. What there is instead is a fairly narrow altitude band where the plasma finally becomes thin enough for light to escape without being immediately reabsorbed. Below it, the Sun is opaque; above it, transparent. That band is the photosphere — literally "sphere of light" — and it is what your eye registers as the edge of the Sun.

The famous figure of 5,500 °C (5,772 K) is the effective temperature: the temperature a perfect radiator would need to be in order to emit the amount of energy the Sun actually emits. It is a single number standing in for a layer that is genuinely about 6,300 °C at the bottom and 4,100 °C at the top. The photosphere is only around 500 km thick — 0.07% of the Sun's radius — which is why it looks like a crisp edge from 150 million kilometres away, and why the Sun casts sharp shadows.

This is the temperature that matters for everyday life. It sets the colour of sunlight, and therefore the wavelengths that plants evolved to harvest and that your eyes evolved to see. A cooler star would look redder; a hotter one, bluer.

Roughly 500 km above the visible surface the temperature bottoms out at about 3,830 °C. It is worth pausing on how ordinary that is by solar standards and how absurd it is by any other: the coldest spot in the entire Sun is still hot enough to boil tungsten.

It is cool enough, though, for a few of the toughest molecules to survive rather than being torn apart — carbon monoxide and water vapour among them, detectable in the Sun's spectrum. There is water in the Sun, in the sense that a small number of intact H₂O molecules exist in this thin cold layer and in sunspots.

This is the hinge of the whole story. Every step outward so far has been a step down in temperature, exactly as intuition demands: you are moving away from the furnace. From this point on, every step outward is a step up. That should be impossible. Heat does not flow from cold to hot on its own — that is the second law of thermodynamics, and it is not a suggestion.

The resolution, revealed over the next two sections, is that heat is not flowing outward from the photosphere to warm the corona. Something else is delivering energy up there and depositing it directly, bypassing the surface entirely. That something is magnetic.

Above the temperature minimum the trend reverses and the temperature climbs steadily to somewhere around 20,000 °C. The chromosphere — "sphere of colour" — is normally invisible, drowned out by the photosphere below it, which is roughly ten thousand times brighter. But in the seconds around totality in a solar eclipse, when the Moon covers the photosphere exactly, a thin crimson rim flashes into view around the black disc. That colour is a single spectral line: hydrogen atoms emitting at 656 nm.

It is not a smooth shell. The chromosphere is a forest of spicules — jets of plasma, some 500 km wide, shooting upward at 20 km per second, rising several thousand kilometres and collapsing within about ten minutes. At any moment there are on the order of a hundred thousand of them. They trace magnetic field lines, and they are one of the routes by which energy and material are fed upward into the layers above.

The density here is already falling off a cliff — the chromosphere is perhaps ten thousand times thinner than the photosphere. Which matters enormously, because thin gas is easy to heat. Pour a modest amount of energy into a small number of particles and each particle ends up moving very fast. That is the mechanism about to run away completely.

This is the vertical cliff in the second chart, and it is the most extreme temperature gradient anywhere in the solar system. The temperature rises by a factor of roughly fifty — from around twenty thousand degrees to over a million — across a layer that in places is less than a hundred kilometres thick. That is the distance from central London to Brighton. Over that span, the plasma goes from something like a very hot flame to something that emits X‑rays.

The reason the layer is so thin is a runaway. Below a certain temperature, hydrogen still has its electron attached, and such atoms are extremely efficient radiators: pump energy in and they promptly dump it back out as light, which stabilises the temperature. Push past a threshold and the hydrogen becomes fully ionised. Its main cooling channel switches off. Now the same energy input has nowhere to go except into particle motion, so the temperature leaps — and the hotter it gets, the less able it is to radiate the heat away. There is no stable state in between, so the Sun does not occupy one. The gas is either chromospheric or coronal, with almost nothing between.

Calling it a "region" flatters it. It is not a smooth shell at a fixed height but a crumpled, shifting boundary that follows magnetic structures up and down by thousands of kilometres and changes shape minute to minute.

The Sun's outer atmosphere sits at one to three million degrees Celsius, some two to five hundred times hotter than the surface beneath it. In active regions above sunspot groups it reaches higher still, and during a large flare, plasma can be heated past ten million degrees — hotter than the Sun's own core, achieved in the thinnest material in the neighbourhood.

And it would not burn you, at least not by conduction. Sea‑level air contains around 2.5 × 10²⁵ particles per cubic metre. The corona contains around 10¹⁵ — ten billion times fewer. This is a harder vacuum than anything produced in a laboratory on Earth. Each of those particles is moving furiously fast, but there are so few of them that the total heat available to transfer is minuscule.

This is why NASA's Parker Solar Probe survives inside the corona at all. Its heat shield does not have to withstand a million degrees; it has to withstand the intensity of sunlight at close range, which is a radiation problem, not a contact‑heating one. On its record passes it flew within 6.1 million km of the surface, roughly 8.8 solar radii from the Sun's centre, while the Sun‑facing surface of its shield reached around 980 °C — hot, but four orders of magnitude below the temperature of the plasma it was flying through. The instruments behind the shield stayed near room temperature.

The corona is not uniform. Where magnetic field lines loop back to the surface, plasma is trapped in bright arcs — the glowing loops in ultraviolet images of the Sun. Where field lines open into space, the plasma streams away, leaving darker, cooler, less dense regions called coronal holes. Those holes are the launch points for the fast solar wind, which is where this journey goes next.

The open question

Why is the corona hot? Nobody is sure.

This has been an unsolved problem for roughly eighty years, and it is worth being honest that it remains one.

The constraint is thermodynamic. Heat cannot flow spontaneously from the 5,500 °C photosphere into the million‑degree corona, so the corona cannot simply be warmed from below by conduction or by ordinary radiation. Energy has to be transported through the cooler layers without depositing much there, and then released high up. The one mechanism capable of doing this is the Sun's magnetic field, which threads all these layers and is anchored in the churning convective plasma below. The argument for a magnetic cause is strong. The argument about which magnetic mechanism dominates is not settled.

Two families of explanation have competed for decades, and current thinking is that both contribute:

• Nanoflares. Proposed by Eugene Parker in 1988. Convection continually shuffles the footpoints of magnetic field lines, tangling and braiding them. Tangled field lines store energy; when they snap and reconnect into simpler configurations, they release it as heat. Individually these events are tiny — a billionth of a large flare, hence the name — but if they happen constantly and everywhere, the aggregate could heat the corona. The difficulty is that individual nanoflares are near or below the detection limit of current instruments, which makes the theory awkward to test directly.
• Wave heating. The convecting surface shakes magnetic field lines like plucked strings, launching waves — in particular Alfvén waves — that travel upward along the field. Because the density falls so steeply with height, a wave carrying a fixed amount of energy into ever‑thinner material produces ever‑larger motions, and eventually dissipates into heat. Such waves have been directly observed. Whether they carry enough energy, and whether they deposit it in the right places, is still argued.

Recent missions have narrowed the field without closing it. ESA's Solar Orbiter has imaged ubiquitous small‑scale brightenings nicknamed "campfires" — plausible relatives of the nanoflare, though their energy budget is still being assessed. Parker Solar Probe has flown inside the corona and sampled the plasma and fields in place rather than inferring them from a distance. The measurements are far better than they were a decade ago. The answer is not yet in hand.

The corona is too hot for the Sun's gravity to hold onto it. Beyond a few solar radii the plasma simply escapes, accelerating outward into a continuous flow that fills the solar system. Eugene Parker predicted this in 1958, was widely disbelieved, and was confirmed by Soviet and American spacecraft within a few years. The probe now flying through the corona carries his name.

As the wind expands into an ever‑larger volume it cools — this time in the ordinary way, the same physics as a gas cooling when it expands out of an aerosol can. From roughly a million degrees near the Sun, the proton temperature falls to somewhere around 100,000 °C by the time it reaches Earth's orbit. Electrons run a little hotter, and fast wind runs hotter than slow wind; measured values at 1 AU spread across a range of roughly 30,000 to 200,000 °C depending on stream and conditions.

Interestingly, the wind cools more slowly than simple expansion alone would predict, which means something is still depositing energy into it far from the Sun. That is another live research question, and probably a relative of the coronal heating problem.

Density is the headline, though. By 1 AU the solar wind is down to about five particles per cubic centimetre. A cubic centimetre of the air you are breathing contains around 2.5 × 10¹⁹ molecules. The plasma streaming past Earth at a hundred thousand degrees is something like ten million million million times thinner than air.

Ask what the temperature is at Earth's distance from the Sun and you get three different correct answers, which is a fitting place to end.

The plasma is at about 100,000 °C. True, and irrelevant to anything you could feel, for the reason we established at the start: five particles per cubic centimetre. This number matters for spacecraft charging and for space weather physics. It does not matter for temperature in any sense you'd recognise.

An object out there settles at around +5 °C. This is the one that actually governs experience. Sunlight arrives at Earth's distance carrying about 1,361 W/m² — the solar constant. An object absorbs that energy and re‑emits it as infrared, and it warms until the two balance. For a perfectly black, uniformly heated sphere at 1 AU the balance point is 278.6 K, or about +5 °C. Space, in the only sense that matters for a thermometer, is roughly the temperature of a cool spring morning.

Real objects deviate from this considerably, because the balance depends on how reflective they are and whether they rotate. Earth reflects about 30% of incoming sunlight, dropping its equilibrium temperature to roughly −18 °C; the atmosphere's greenhouse effect brings the actual surface back up to about +15 °C. A satellite that does not rotate is far worse off: its sunlit face and its shadowed face have no way to share heat, and surfaces on the International Space Station swing between roughly +120 °C and −150 °C depending on whether they are in sunlight. Spacecraft thermal design is mostly the art of managing that split with radiators, insulation and slow rotation.

Empty space bottoms out at −270.4 °C. Take the Sun away entirely, shield an object from every star and every galaxy, and it still cannot cool below 2.725 K. The universe is filled with faint microwave radiation left over from about 380,000 years after the Big Bang, and everything sits in that bath. It is the coldest temperature the universe supplies for free, and it is the floor beneath every number in this article.

Evidence

How anyone knows any of this

Reasonable question: no instrument has ever been inside the Sun. Four independent lines of evidence converge, which is why the numbers are trusted.

Spectroscopy — reading the light

Atoms absorb and emit light at wavelengths that depend on their element and on how many electrons have been stripped away. Since stripping electrons takes energy, the ionisation state of a gas reports its temperature. Line widths add a second, independent check: hotter atoms move faster, and their motion smears the wavelength through the Doppler effect, so a line's width measures the temperature directly.

This is how the coronal temperature was discovered, and the story is a good one. During eclipses in the nineteenth century, astronomers found a bright green emission line in the corona at 530.3 nm that matched no known element. They proposed a new one and called it coronium. It took until the early 1940s for Bengt Edlén, building on work by Walter Grotrian, to identify it: not a new element but ordinary iron, stripped of thirteen of its twenty‑six electrons. Doing that to iron requires temperatures of order a million degrees. The corona had been telling us it was extraordinarily hot for eighty years before anyone could read the message.

Helioseismology — listening to the Sun ring

The Sun oscillates. Convection generates sound waves that bounce around inside it, making the whole star vibrate in millions of overlapping modes, which show up as tiny rhythmic Doppler shifts across the visible surface. Because the speed of sound in a gas depends on its temperature and composition, and because different modes penetrate to different depths, measuring millions of these frequencies lets you invert the problem and reconstruct the interior — much as seismologists map Earth's core using earthquakes. This is how the base of the convection zone was pinned to 0.713 R☉, and it is the strongest constraint on the interior temperature profile.

Neutrinos — a direct look at the core

Fusion produces neutrinos that escape the Sun immediately and reach Earth in eight minutes. Detecting them is extraordinarily difficult, but experiments have now measured the flux from the main proton–proton reactions. The count depends steeply on the core temperature, so it independently confirms the central value to within about one percent. Historically, the measured flux came out too low — the "solar neutrino problem" — and the resolution turned out to be that neutrinos change type in flight, a discovery about particle physics rather than about the Sun. The solar model had been right all along.

Going there

Since 2018, Parker Solar Probe has flown repeatedly through the corona, measuring plasma temperature, density and magnetic fields in place. Solar Orbiter images the Sun from close range and out of the ecliptic plane. Older missions — Ulysses, Wind, ACE, SOHO — have sampled the solar wind for decades. Everything from the transition region outward now has direct in‑situ measurement behind it, not just inference.

Reference

Every layer, in one table

Going deeper

Where to read more

Grouped by how far you want to go. Everything in the first group is written for non‑specialists.