How the Sun Works — Nuclear Fusion Powered Star

The Sun is a G-type main-sequence star (spectral class G2V), an ordinary yellow dwarf by galactic standards, about 4.6 billion years old with roughly 4.5 billion more years of stable hydrogen burning ahead of it. Every second, its core fuses about 600 million metric tons of hydrogen into helium, converting roughly 4 million tons of that mass directly into energy. That energy takes hundreds of thousands of years to fight its way out of the core, then eight minutes to cross 150 million kilometers (93 million miles) of empty space and land on your skin as warmth.

That single fact — light you feel today began its journey before Homo sapiens existed as a species — is the kind of detail that makes the Sun worth understanding in mechanical detail rather than just admiring. This article walks through what's actually happening inside the Sun, layer by layer, reaction by reaction, and why physicists spent nearly a century arguing about the answer before getting it right.

Why the "obvious" explanation for sunlight was wrong for 100 years

The question "what powers the Sun" sounds simple, but 19th-century science genuinely could not answer it. Chemical combustion was ruled out almost immediately: even if the Sun were a solid block of coal burning in pure oxygen, it would exhaust its fuel and go dark in about 6,000 years — laughably short compared to the geological record already being uncovered by Charles Lyell and Charles Darwin's contemporaries (Astronomy.com, "Before Einstein, what did astronomers think powered the Sun?").

The best physics available then came from Hermann von Helmholtz in 1854 and Lord Kelvin (William Thomson) afterward: gravitational contraction. The idea was that the Sun is slowly shrinking, and the gravitational potential energy released by that collapse converts into heat, the same way a falling object heats up on impact. Run the numbers and this "Kelvin-Helmholtz mechanism" gives the Sun a lifespan of roughly 20 to 500 million years. That's a real physical process, and it does happen (protostars are still heated this way during formation) — but it's far too short. Geologists were already finding rock strata and fossil sequences that implied Earth was hundreds of millions to billions of years old, and Kelvin's short sun-lifetime became the subject of a famous public dispute with Darwin, whose theory of natural selection needed much more time to work (ScienceDirect, "Why the Kelvin-Helmholtz timescale is not really their timescale").

The resolution didn't arrive until the 1920s and 1930s, once physicists understood atomic nuclei well enough to realize that fusing light elements together releases enormous energy through mass-to-energy conversion, described by Einstein's E = mc². In 1939, physicist Hans Bethe published "Energy Production in Stars," laying out two nuclear pathways — the proton-proton chain and the CNO cycle — that could keep a star like the Sun shining for billions of years. Bethe won the 1967 Nobel Prize in Physics largely for that paper (American Physical Society, "Landmarks: What Makes the Stars Shine?"; NSF, "The History and Impact of the CNO Cycles in Nuclear Astrophysics"). It's worth sitting with how recent that is: humans have understood how the Sun works for less than a century, out of roughly ten thousand years of recorded civilization built entirely underneath it.

The 19th century's best physics gave the Sun a lifespan of 20 to 500 million years via gravitational contraction alone. Nuclear fusion, understood only since the 1930s, extends that to roughly 10 billion years — more than twenty times longer than Kelvin's estimate.

Structure of the Sun: six layers, one furnace

The Sun isn't a uniform ball of glowing gas. It's a layered structure, and each layer moves energy in a fundamentally different way.

Layer	Approximate location	Temperature	What happens there
Core	Innermost ~25% of radius	~15 million K (27 million °F)	Nuclear fusion; density ~150 g/cm³, about 10× denser than lead (NASA Science, "Sun: Facts")
Radiative zone	25%–70% of radius	~7 million K down to ~2 million K	Energy moves outward as photons, absorbed and re-emitted repeatedly
Convective zone	Outer ~30% of radius	~2 million K down to ~5,800 K	Heat moves via boiling plasma cells (convection)
Photosphere	Visible "surface"	~5,500 °C (~9,900 °F)	The layer we actually see; source of visible sunlight
Chromosphere	Thin layer above photosphere	~6,000–20,000 K	Reddish layer visible briefly during total eclipses
Corona	Outer atmosphere, millions of km deep	1–3 million K	Wispy plasma atmosphere, mysteriously hotter than the surface below it

Table: The Sun's structure from core to corona. Temperature and density figures from NASA Science and NASA's In-Depth solar profile.

The core occupies only about 1.6% of the Sun's total volume but, because it's so dense and hot, it generates essentially all of the Sun's energy output. Pressure at the core is estimated at roughly 340 billion times Earth's atmospheric pressure at sea level — pressure so extreme it's what keeps hydrogen nuclei close enough, long enough, for fusion to occur despite their mutual electric repulsion (NASA Science, "Sun: Facts").

Why does energy take so long to leave the radiative zone?

A photon produced by fusion in the core doesn't travel in a straight line to the surface. The radiative zone is so dense that each photon travels only a fraction of a millimeter before being absorbed by a particle and re-emitted in a random new direction — a process called a "random walk." Because each step is essentially undone by the next random direction change, it can take a photon on the order of tens of thousands to a few hundred thousand years (estimates vary by model) to random-walk its way from the core to the base of the convective zone, even though the physical distance is only about 500,000 kilometers (UCF Pressbooks, "The Solar Interior: Theory"). Once it reaches the convective zone, the trip to the surface and out into space takes only minutes by comparison, because convection physically carries hot plasma upward rather than relying on photons bouncing at random.

Why is the corona hotter than the surface below it?

This is the "coronal heating problem," one of solar physics' longest-standing open questions. The photosphere sits at roughly 5,500°C, but the corona above it reaches 1 to 3 million kelvin — hundreds of times hotter, which seems to violate basic thermodynamic intuition about heat flowing from hot to cold. The leading explanations involve magnetic reconnection (sudden snapping and realignment of magnetic field lines that releases energy as heat) and Alfvén waves (oscillations traveling along magnetic field lines that deposit energy into the corona as they dissipate). NASA's Parker Solar Probe, which has flown directly through the corona since 2021, has gathered direct measurements aimed at resolving exactly which mechanism dominates, but the question isn't fully settled.

Nuclear fusion: the proton-proton chain, step by step

About 99% of the Sun's energy comes from the proton-proton chain reaction, with the CNO cycle (which dominates in more massive, hotter stars) contributing only a small fraction in a star as modestly sized as the Sun (Wikipedia, "CNO cycle," citing stellar nucleosynthesis models). The proton-proton chain converts hydrogen into helium in three broad steps:

1. Two protons (hydrogen nuclei) fuse to form deuterium (a heavy hydrogen isotope), releasing a positron and a neutrino. This step is astonishingly rare on a per-proton basis — it also requires a proton to spontaneously convert into a neutron via the weak nuclear force, which happens so infrequently that a given proton in the Sun's core has, on average, billions of years to wait before it succeeds.
2. The deuterium nucleus quickly captures another proton to form helium-3, releasing a gamma ray.
3. Two helium-3 nuclei collide and fuse into helium-4, releasing two protons back into the core to start the process again.

The net reaction is:

4 hydrogen nuclei → 1 helium nucleus + 2 positrons + 2 neutrinos + about 26.7 MeV of energy

That 26.7 MeV per completed chain sounds trivial on its own, but the Sun runs an enormous number of these reactions simultaneously — enough to convert roughly 600 million metric tons of hydrogen into helium every second, with about 4 million tons of that mass converted directly into energy via E = mc² (NASA Science, "Sun: Facts"). Put another way: the Sun is measurably losing mass, roughly 4 billion kilograms per second, and has been doing so for 4.6 billion years. That's still trivial compared to the Sun's total mass of about 2 × 10³⁰ kilograms — the equivalent of a person losing a single eyelash over their entire lifetime, proportionally speaking.

The Sun is, quite literally, getting lighter every second. It has lost an amount of mass equivalent to roughly 100 Earths since it formed — and it still has about 4.5 billion years and the mass of tens of thousands more Earths left to burn.

Why don't hydrogen nuclei just repel each other and never fuse?

Protons carry positive electric charge, and like charges repel — this is the Coulomb barrier, and by classical physics alone, protons in the Sun's core shouldn't be able to get close enough to fuse. Two things solve this. First, the core's temperature (15 million K) and pressure give protons enough kinetic energy to occasionally get very close despite the repulsion. Second, and more decisively, quantum tunneling allows a small fraction of protons to "tunnel through" the repulsive barrier probabilistically rather than needing enough energy to power straight through it classically. Without tunneling, the Sun's core wouldn't actually be hot enough to fuse hydrogen at any meaningful rate — it's a quantum effect keeping the entire star lit.

Neutrinos: the particles that don't wait in line

Every proton-proton chain reaction produces neutrinos — nearly massless particles that interact with matter so weakly that they stream straight out of the Sun in about two seconds, compared to the tens of thousands of years photons spend trapped in the radiative zone. Roughly 60 billion solar neutrinos pass through every square centimeter of your body per second, day and night (they pass through Earth just as easily as through you), without you noticing a thing.

This produced one of the twentieth century's most stubborn puzzles: the solar neutrino problem. Starting in the 1960s, experiments like Raymond Davis Jr.'s Homestake experiment detected only about a third of the neutrinos the standard solar model predicted. For decades, physicists debated whether the shortfall meant the Sun's fusion model was wrong or the detectors were missing something. The answer, confirmed decisively by the Sudbury Neutrino Observatory in 2001, was neutrino oscillation: neutrinos switch between three "flavors" (electron, muon, and tau) as they travel, and early detectors were built to see only the electron flavor. Once all three flavors were accounted for, the numbers matched the Sun's fusion output exactly. The discovery of neutrino oscillation won the 2015 Nobel Prize in Physics, awarded jointly to Takaaki Kajita and Arthur B. McDonald (SNO Collaboration, Physical Review Letters, 2001).

Total solar irradiance: measuring the Sun's output precisely

Scientists express the Sun's total energy output at Earth's distance as the "solar constant" or, more precisely, total solar irradiance (TSI). Space-based instruments — most recently NASA's Total Irradiance Monitor aboard the SORCE satellite — pinned this figure at approximately 1,361 watts per square meter, a downward revision from the older canonical value of about 1,365 watts per square meter established in the 1990s. The correction came from better accounting for scattered light in earlier-generation instruments, and the International Astronomical Union formally adopted the 1,361 W/m² figure in 2015 (LASP, University of Colorado, "TIM Science"). Averaged across the entire globe (day and night, poles and equator), that works out to about 340 watts per square meter reaching the top of Earth's atmosphere — roughly the output of three incandescent light bulbs' worth of power landing on every square meter of the planet, continuously.

The name "solar constant" is a bit misleading. TSI actually varies by about 0.1% across the roughly 11-year solar cycle, rising slightly near solar maximum despite sunspots (which are cooler, darker patches) becoming more numerous — because sunspots are more than offset by bright magnetic regions called faculae that also increase during active periods (arXiv, "Why isn't the solar constant a constant?").

The solar wind, sunspots, and the 11-year cycle

The corona is so hot that some of its particles reach escape velocity and stream outward continuously as the solar wind, typically moving at 400 to 800 kilometers per second. This isn't a gentle breeze — it's a supersonic outflow of charged particles that carves out a bubble around the entire solar system called the heliosphere, deflects incoming cosmic rays, and is responsible for the tails of comets always pointing away from the Sun regardless of the comet's direction of travel.

Sunspots are darker, cooler patches on the photosphere where concentrated magnetic field lines suppress the convection that normally brings heat to the surface, making those patches a few thousand degrees cooler than the surrounding photosphere (they only look black by contrast — in isolation, a sunspot would still glow brightly). Sunspot numbers rise and fall on an approximately 11-year cycle, driven by the Sun's magnetic dynamo — a self-sustaining process, first modeled in detail by Horace Babcock and later refined into the "Babcock-Leighton mechanism," in which differential rotation (the Sun's equator spins faster than its poles) winds up and twists the magnetic field until it becomes unstable and erupts through the surface as sunspot pairs (arXiv, "Recent Developments in the Babcock-Leighton Solar Dynamo Theory").

The cycle isn't perfectly regular. Between 1645 and 1715, astronomers recorded a prolonged stretch of almost no sunspots at all, a period now called the Maunder Minimum. It roughly coincided with the coldest part of the "Little Ice Age" in Europe, though the exact causal link between low solar activity and regional cooling is still debated among climate historians. Modern dynamo models suggest grand minima like this occur when fluctuations in the Babcock-Leighton process cause the underlying poloidal magnetic field to collapse to unusually low strength at the end of a cycle (arXiv, "A possible explanation of the Maunder minimum from a flux transport dynamo model").

The most recent cycle, Solar Cycle 25, ramped up faster and stronger than many forecasters predicted, peaking around 2024–2025. It produced the Gannon Storm of May 10, 2024 — the first G5 "extreme" geomagnetic storm to strike Earth in more than two decades, named posthumously for space weather physicist Jennifer Gannon. The storm originated from an active sunspot region roughly 17 times the size of Earth, which released a chain of coronal mass ejections that merged en route and slammed into Earth's magnetosphere together. NOAA issued its first severe geomagnetic storm watch in nearly 20 years ahead of the event. Auroras were visible as far south as Florida and Mexico, and while the storm caused no catastrophic damage, some high-voltage power lines tripped, transformers overheated, and GPS-guided farm equipment in the U.S. Midwest lost precision guidance, costing affected farms an average of roughly $17,000 each during peak planting season (NASA, "What NASA Is Learning From the Biggest Geomagnetic Storm in 20 Years"; Copernicus GC, "The Gannon Storm: citizen science observations").

How bad could a worst-case solar storm actually get?

The historical benchmark is the Carrington Event of September 1859, the most extreme space weather event on the modern instrumental record. It induced currents strong enough to shock telegraph operators, ignite telegraph paper, and — most remarkably — let some stations keep transmitting messages on "auroral current" alone after their batteries were disconnected. A Lloyd's of London risk model estimates that a Carrington-scale storm striking the modern, far more electronics-dependent United States could cost between roughly $0.6 trillion and $2.6 trillion, and a 2008 U.S. National Research Council report warned that full recovery from a comparable long-duration blackout could take four to ten years depending on how many large transformers were destroyed (they are typically custom-built and can take a year or more to replace) (NOAA NESDIS, "When Solar Storms Attack"; Wikipedia, "Carrington Event," citing Lloyd's of London and NRC assessments). Modern power grids, satellite constellations, and GPS-dependent systems are all more exposed to this kind of event than the sparse telegraph network of 1859 was — which is exactly why agencies like NOAA's Space Weather Prediction Center now monitor the Sun continuously rather than treating solar storms as a curiosity.

How scientists actually study the inside of a star

You cannot fly a probe into the Sun's core — no material could survive it, and even the corona pushes spacecraft engineering to its limits. So solar physicists rely on a mix of remote observation and an elegant indirect technique borrowed from geology.

• Helioseismology tracks pressure waves rippling across the Sun's surface, the way seismologists use earthquake waves to map Earth's interior structure without ever drilling to the core. These waves reveal the depth of the convective zone, rotation rates at different depths, and even structures on the far side of the Sun before they rotate into direct view.
• Daniel K. Inouye Solar Telescope (Haleakalā, Hawaii), which achieved first light in 2020, produces the highest-resolution ground-based images of the photosphere ever taken, resolving convection cells down to roughly 30 kilometers across.
• Parker Solar Probe, launched by NASA in 2018, became the first spacecraft to fly directly through the corona, using a carbon-composite heat shield to survive temperatures exceeding 1,300°C during close passes. Its data is central to resolving the coronal heating problem. Read the full mission story in our Parker Solar Probe explainer.
• Solar Orbiter (a joint ESA/NASA mission) captured the first-ever direct images of the Sun's poles, a viewing angle Earth-based telescopes can never achieve because we orbit in the Sun's equatorial plane.
• Standard Solar Model, a decades-refined theoretical framework combining nuclear physics, stellar structure equations, and helioseismology data, is what actually let physicists predict the neutrino flux tested (and eventually confirmed) by SNO in 2001.

The Sun compared to other stars

Not every star burns the way the Sun does, and comparing stellar classes clarifies just how unremarkable — in the best way — our star actually is.

Star Type	Mass (Sun = 1)	Approximate Lifetime	End State
Red Dwarf (M)	0.08–0.5	Trillions of years	Helium white dwarf (theoretical; universe isn't old enough yet)
Orange Dwarf (K)	0.5–0.8	~15–30 billion years	Carbon/oxygen white dwarf
Sun (G2V)	1.0	~10 billion years total	Carbon/oxygen white dwarf
White Star (A)	1.4–2.1	~1–3 billion years	Carbon/oxygen white dwarf
Blue Giant (O)	>16	~3–10 million years	Supernova → neutron star or black hole

Table: Stellar mass, lifetime, and fate by spectral class. Lifetimes scale inversely with mass because more massive stars burn fuel far faster despite having more of it.

Mass, somewhat counterintuitively, is destiny in the opposite direction you'd expect: heavier stars burn through their hydrogen much faster because fusion rates rise steeply with core temperature and pressure, both of which scale with mass. A blue giant twenty times the Sun's mass has roughly twenty times the fuel but burns through it so much faster that it lives only a few million years instead of billions. The Sun will never go supernova — that fate is reserved for stars above roughly 8 solar masses, and the Sun sits at exactly 1.

What people get wrong about the Sun

"The Sun is on fire." It isn't. Fire is a chemical combustion reaction that requires fuel plus an oxidizer, typically oxygen, and produces smoke and combustion byproducts. The Sun's core has no oxygen involved in the process and doesn't need any — it produces light and heat through nuclear fusion, and its visible glow at the surface is blackbody radiation, the same physical phenomenon that makes a metal rod glow orange-hot in a forge, just at a much larger scale and higher temperature.

"Sunlight reaches us instantly, and it's the same light produced by fusion right now." Neither is true. Sunlight takes about 8 minutes and 20 seconds to cross the 150-million-kilometer gap to Earth once it leaves the photosphere — but the photons themselves were generated in the core hundreds of thousands of years earlier and spent nearly all of that time slowly random-walking through the radiative zone. The light warming your face today began as fusion energy while woolly mammoths still walked the Earth.

"The Sun is shrinking, so eventually it'll just fizzle out." Some fringe claims argue the Sun is measurably shrinking in radius year over year in a way that threatens its stability. Mainstream helioseismology and decades of precise solar-radius measurements show no such alarming trend — normal fusion-driven equilibrium, not gravitational collapse, governs the Sun's size on human timescales (Answers in Genesis critique aside, mainstream astrophysical measurement shows solar radius stability). The Sun's eventual changes are real, but they play out over billions of years, not years or decades.

"More sunspots mean the Sun is 'cooling down' and putting out less energy." Backwards. Total solar irradiance actually rises slightly during solar maximum, when sunspots are most numerous, because bright magnetic faculae surrounding active regions more than compensate for the darker sunspots themselves.

The Sun's far future

The Sun's luminosity has been rising gradually for its entire life and will continue to do so — roughly 10% per billion years, a consequence of the core's composition slowly shifting from hydrogen to helium, which changes the core's density and fusion rate over very long timescales. In roughly 1 billion years, that increase alone will likely be enough to boil away Earth's oceans, regardless of anything humans do to the climate in the meantime. In roughly 5 billion years, the Sun will exhaust the hydrogen in its core, and the core will contract while the outer layers expand dramatically as the Sun becomes a red giant, likely engulfing Mercury and Venus and possibly reaching Earth's current orbit. After that, the outer layers will drift away as a planetary nebula, leaving behind a dense, slowly cooling white dwarf roughly the size of Earth but containing most of the Sun's original mass — a corpse that will keep glowing faintly for trillions of years after every human civilization's memory of the Sun as it looks today has vanished.

For the mechanics of that transition across different stellar masses, and for a look at whether humanity might one day replicate fusion here on Earth for power generation, see our nuclear fusion energy explainer and our Parker Solar Probe deep dive. You can also browse more solar system and stellar physics coverage on the space category page.

Frequently Asked Questions

Q: Is the Sun on fire? A: No. Fire is a chemical combustion reaction that needs an oxidizer like oxygen. The Sun shines because of nuclear fusion in its core, converting hydrogen into helium, plus blackbody radiation from its roughly 5,500°C surface.

Q: How long does it take sunlight to reach Earth? A: About 8 minutes and 20 seconds to cross the 150-million-kilometer gap. But the fusion energy in that photon was actually generated in the core hundreds of thousands of years earlier — it just spent nearly all that time bouncing randomly through the dense radiative zone before escaping.

Q: Why is the Sun's corona hotter than its surface? A: This is the unsolved "coronal heating problem." The photosphere sits around 5,500°C, but the corona above it reaches 1 to 3 million kelvin. Magnetic reconnection and Alfvén wave dissipation are the leading explanations, actively being tested with Parker Solar Probe data, but the exact balance between mechanisms isn't fully settled.

Q: Will the Sun ever explode as a supernova? A: No. Supernovae are reserved for stars with more than roughly 8 times the Sun's mass. The Sun will instead swell into a red giant in about 5 billion years, shed its outer layers, and settle into a slowly cooling white dwarf.

Q: How many fusion reactions happen in the Sun per second? A: An extremely large number, converting roughly 600 million metric tons of hydrogen into helium every second and turning about 4 million tons of that mass directly into energy via E = mc².

Q: What is the solar constant, and is it really constant? A: It's the Sun's total energy output measured at Earth's distance, about 1,361 watts per square meter according to NASA's SORCE/TIM instrument. It isn't perfectly constant — it varies by roughly 0.1% across the 11-year solar cycle, rising slightly during solar maximum.

Q: What are neutrinos, and why did they cause a decades-long scientific puzzle? A: Neutrinos are nearly massless particles produced by fusion that pass through matter almost undisturbed, reaching Earth from the Sun's core in about eight minutes rather than the hundreds of thousands of years photons take. Early detectors found only about a third of the predicted count; the 2001 discovery that neutrinos oscillate between three "flavors" resolved the mismatch and won the 2015 Nobel Prize in Physics.

Q: What caused the Gannon Storm of May 2024, and how bad was it? A: A cluster of coronal mass ejections from an active sunspot region roughly 17 times Earth's size merged together and struck Earth's magnetosphere as a single G5 "extreme" geomagnetic storm, the strongest in over two decades. It caused visible auroras as far south as Florida, some grid disturbances, and roughly $17,000 in average losses for affected U.S. Midwest farms relying on GPS-guided equipment, but no catastrophic damage.

Q: Could a solar storm like the 1859 Carrington Event happen again, and what would it cost? A: Yes — nothing prevents a Carrington-scale event from recurring, and modern infrastructure is far more electronics-dependent than in 1859. A Lloyd's of London risk model estimates a comparable storm today could cost the United States between roughly $0.6 trillion and $2.6 trillion, with a 2008 U.S. National Research Council report warning that full grid recovery could take four to ten years depending on transformer losses.

Q: When did scientists actually figure out that fusion powers the Sun? A: Not until 1939, when physicist Hans Bethe published the theoretical framework for the proton-proton chain and CNO cycle, work that earned him the 1967 Nobel Prize in Physics. Before that, 19th-century physics could only explain a Sun that would be a few hundred million years old at most, which conflicted directly with the geological evidence of Earth's much greater age.

The takeaway

The Sun's basic trick is simple to state and staggering in scale: it fuses hydrogen into helium under pressure and temperature so extreme that a tiny fraction of the reacting mass converts directly into energy, and it has run that reaction reliably for 4.6 billion years without much variation. Nearly everything else about the Sun's structure, from the centuries-long photon traffic jam in the radiative zone to the still-unexplained heat of the corona, follows from that one core process and the physics of moving its energy outward. Understanding it took humanity until the 1930s, and pieces of the puzzle, like exactly what heats the corona, remain open questions that missions like Parker Solar Probe are actively working to close.

---

Read next:

• Nuclear Fusion Energy: How Close Are We Really?
• Parker Solar Probe — Touching the Sun
• Space category — more solar system and stellar physics