How the Large Hadron Collider Works — And What It's Searching For

On July 4, 2012, two teams at the Large Hadron Collider announced a new particle with a mass between 125 and 127 GeV (gigaelectronvolts, a unit of energy that physicists also use for mass). It turned out to be the Higgs boson, the last missing piece of the Standard Model of particle physics. The machine that found it is a 26,659-meter ring of superconducting magnets near Geneva, where protons circle 11,245 times a second before they collide. This explainer follows a proton from hydrogen gas to detector, then asks what the machine is still hunting.

Key takeaways

• The LHC accelerates protons to a record 6.8 TeV per beam, so head-on collisions reach 13.6 TeV, according to CERN.
• Its main bending magnets are superconducting and run at 1.9 K (−271.3°C), colder than outer space.
• ATLAS and CMS announced the Higgs boson discovery on July 4, 2012. The Higgs appears in only about one in a billion collisions.
• Dozens of new composite particles (79 hadrons by February 2026, per Wikipedia's running tally) have turned up, but no confirmed particle beyond the Standard Model.
• The next step is the High-Luminosity LHC, which aims to raise collision rates by a factor of 10.

What the LHC is: a 27-kilometer racetrack for protons

The Large Hadron Collider (LHC) is a particle accelerator, a machine that uses electric fields to speed charged particles up and magnetic fields to steer them. It sits in a circular tunnel with a circumference of 26,659 meters, typically described as 27 kilometers, near Geneva. CERN, the European laboratory that runs it, says the tunnel lies about 100 meters underground, and the Wikipedia summary of the machine gives a range of 50 to 175 meters along the ring.

"Hadron" is the physicists' word for any particle made of quarks held together by the strong force. Protons and neutrons are hadrons. The LHC uses protons for most of the year and lead ions for part of it, and "collider" means that two beams travel in opposite directions and meet head-on, rather than one beam hitting a fixed target.

The numbers are easier to hold in a table than in a paragraph.

Quantity	Value	Source
Circumference	26,659 m	CERN
Depth underground	About 100 m (50 to 175 m along the ring)	CERN; Wikipedia
Main dipole magnets	1,232, each about 15 m long	CERN
Dipole operating temperature	1.9 K (−271.3°C)	CERN
Energy per proton beam	6.8 TeV (record); 6.5 TeV earlier	CERN
Collision energy (protons)	13.6 TeV	CERN
Bunches per beam	Up to 2,808	Wikipedia
Protons per bunch	About 100 billion	CERN
Laps per second	11,245	CERN
Collisions per second	About 1.5 billion in ATLAS and CMS	CERN
Vacuum in beam pipes	About 10⁻¹³ bar	CERN
Construction cost	4,332 million Swiss francs	CERN
First beam	September 10, 2008	CERN

One tera-electronvolt (TeV) is a trillion electronvolts, and an electronvolt is the energy an electron gains crossing a one-volt battery. One TeV works out to about 1.6 × 10⁻⁷ joules, a minuscule amount in everyday terms. What makes the LHC extraordinary is that each proton carries 6,800 of them, and that the machine pushes billions of protons around the ring at once.

Physicists run the machine in multi-year campaigns called runs. Run 2 lasted from 2015 to 2018 at 13 TeV. Run 3 began its physics season on July 5, 2022 at 13.6 TeV, and it ran until mid-2026, when Long Shutdown 3 started, per Wikipedia's summary of CERN's schedule. CERN's own LHC page says operation continues until Long Shutdown 3, starting in July 2026.

Why smash protons together at all?

Particle physics has a simple recipe. Concentrate a large amount of energy into a tiny volume, and Einstein's E = mc² says some of it can turn into new, heavier particles. The collision does not break a proton open like a nut. It converts motion energy into mass, which then decays into lighter, more stable particles that the detectors can see.

Think of smashing two Swiss watches together to learn what is inside. If you toss them gently, you learn almost nothing. If you slam them at high speed, gears and springs fly out, and the pattern of debris tells you how they were assembled. The catch is that a proton is not a watch with a few gears. It is a seething bag of quarks and gluons, and in any collision only one quark or gluon from each proton does the interesting work. That is why the machine needs so much headroom: the sub-collision between constituents carries only a fraction of the full 13.6 TeV.

How massive a particle can you make? The Higgs boson weighs about 125 GeV, which is roughly 133 times the mass of a proton (125 GeV divided by 0.938 GeV). Producing it takes a collision energy well above that, and producing it often enough to study takes an enormous number of collisions. Higher energy opens the door to heavier particles, and higher collision rates make rare processes visible.

How is that different from the Big Bang?

The LHC does not recreate the Big Bang, despite the common headline. What it can do is recreate some of the conditions found in the very early universe. CERN says collisions of lead ions in the ALICE experiment produce temperatures more than 100,000 times hotter than the center of the Sun, hot enough to melt protons and neutrons into a soup of quarks and gluons called quark-gluon plasma. That is a laboratory sample of a state of matter the universe is thought to have passed through shortly after the Big Bang, not a universe.

From a bottle of hydrogen to 6.8 TeV: the accelerator chain

The LHC does not start from rest. It is the last stage of a chain of accelerators that CERN has built up over decades, each one feeding the next. The old machines are still in use, which is why the complex is sometimes compared to a relay race: no single runner could cover the distance at the final speed, but each hands off a faster baton.

Stage	What it does	Energy reached
Linac4	Linear accelerator; accelerates negative hydrogen ions	160 MeV
Proton Synchrotron Booster	Strips the electrons so only protons remain	2 GeV
Proton Synchrotron (PS)	Passes the beam on to the SPS	26 GeV
Super Proton Synchrotron (SPS)	Final injector	450 GeV
LHC	Two counter-rotating beams	6.5 TeV design maximum; 6.8 TeV record per beam

Source for the table: CERN's accelerator complex page.

The first step is a bottle of hydrogen gas. Linac4 works with hydrogen ions that carry an extra electron, and at 160 MeV a proton is already moving at a bit more than half the speed of light (my calculation from the proton's rest energy of 0.938 GeV). In the Proton Synchrotron Booster the ions pass through a thin foil that strips off both electrons, leaving bare protons.

Acceleration itself relies on radiofrequency cavities, metal chambers in which an oscillating electric field points forward at the moment a bunch of protons arrives. Each pass gives a small kick, and the protons gain energy a little at a time over many laps. CERN says filling each LHC ring takes 4 minutes 20 seconds, and ramping the beams up to maximum energy takes about 20 minutes.

At full energy a proton has a speed within a hair of light. From the 6.8 TeV beam energy and the proton's rest energy, I calculate a speed of about 99.999999% of the speed of light. As a consistency check, CERN lists 11,245 turns per second for a 26,659-meter ring, which works out to about 299,780 kilometers per second, within a few hundredths of a percent of light speed.

What does "bunch" mean here?

A beam is not a continuous stream. It is a string of tightly packed clumps of protons, called bunches, separated by gaps. The Wikipedia summary of CERN specifications lists up to 2,808 bunches per beam, about 115 billion protons each, spaced 25 nanoseconds apart, so two beams cross 40 million times a second. CERN's overview page puts the typical figure at about 100 billion protons per bunch. Radiofrequency fields force the protons to cluster this way, much as a surfer is carried by a wave rather than spread along the whole beach.

How 1,232 magnets keep protons on a circle

Steering a proton with 6.8 TeV of energy around a 27-kilometer ring takes strong magnetic fields. The LHC uses 1,232 main dipole magnets, each about 15 meters long, to bend the beams, along with 392 main quadrupole magnets to squeeze them, according to CERN. Wikipedia's summary gives the dipole field as 0.54 tesla at injection energy (450 GeV), rising to 7.7 tesla at 6.5 TeV. For scale, CMS's much smaller solenoid magnet, at 4 tesla, is described by CERN as roughly 100,000 times Earth's magnetic field.

Fields that strong need currents that would melt copper. The solution is superconductivity, the phenomenon in which certain materials lose all electrical resistance below a critical temperature. If you want the physics behind it, what superconductivity is and why it works is a good primer, and Onnes's 1911 discovery in cooled mercury tells the origin story. In the LHC, the superconducting windings can carry enormous currents with no resistive loss, but only if they stay cold.

How cold is cold?

The dipoles run at 1.9 K, or −271.3°C, which CERN says is colder than outer space. About 96 tonnes of superfluid helium-4 does the cooling, according to Wikipedia's technical summary. Superfluid helium flows without friction and conducts heat extremely well, so it can pull heat out of a magnet the way a very good radiator pulls it out of a car engine.

Inside the magnets, the beams travel in two separate pipes kept at a vacuum of about 10⁻¹³ bar, which CERN says is nearly as low as the pressure on the Moon. A vacuum matters because every stray gas molecule is a target that could knock a proton out of the beam.

What happens when a magnet quenches

A quench is the sudden loss of superconductivity in part of a magnet. Current that was flowing with no resistance suddenly meets resistance, heats the wire, and can spread through the magnet in a runaway. The system is exquisitely sensitive: according to the Wikipedia summary, losing a fraction of just one ten-millionth (10⁻⁷) of the beam is enough to quench a magnet.

The stakes are high because the stored energy is large. According to Wikipedia, each of the two beam dumps must be able to absorb about 362 megajoules. For a sense of scale, a 400-tonne train at 150 kilometers per hour carries about 350 megajoules of motion energy, by my arithmetic. A beam carrying that much energy can damage equipment if it strays, so the machine relies on a beam dump designed to absorb the full beam on demand, along with protection systems that watch for losses.

What went wrong in 2008

Nine days after the first beam circulated on September 10, 2008, the machine had its worst day. On September 19, 2008, a faulty electrical connection between two magnets caused an electric arc and a helium vent, according to the Wikipedia summary. About 6 tonnes of liquid helium escaped, damaging 53 superconducting magnets and their mountings and contaminating the vacuum pipe.

Repairs and added protection systems pushed the restart back about 14 months, to November 20, 2009. The incident is a good reminder that the LHC runs close to the edge of what superconducting engineering can do, and that the safety systems exist for a reason.

Why beams are bunches, and why most collisions are boring

Colliding two beams is much less like two cars crashing and more like two swarms of gnats flying through each other. Each bunch holds around 100 billion protons, but protons are tiny targets, so when two bunches pass through each other only a small number of proton pairs actually interact. CERN reports about 1.5 billion collisions per second in ATLAS and CMS.

Most of those collisions are glancing and uninteresting. The measure that matters is luminosity, a count of how many potential collisions the beams deliver per square centimeter per second. A machine with high luminosity gives the same particles more chances to meet. If a process happens once in a billion collisions, which is what CERN says about the Higgs boson, luminosity decides whether you see it once a decade or once a minute.

By my back-of-envelope arithmetic, a billion-to-one process at 1.5 billion collisions per second occurs on the order of once or twice a second, so the LHC produces a Higgs boson every second or so. The difficulty is not creating it but recognizing it. The Higgs decays in a tiny fraction of a second (a predicted mean lifetime of about 1.56 × 10⁻²² seconds, per Wikipedia), and the same final products can be mimicked by ordinary processes that happen far more often. Physicists therefore look for statistical excesses in thousands of events, not single sightings.

What the four big detectors do with the debris

At four points on the ring the beams cross inside a detector, a layered instrument the size of an apartment building. CERN lists nine experiments installed at the LHC, but four are the main ones.

Detector	Size and weight	Specialty	Collaboration (per CERN pages, 2022)
ATLAS	46 m long, 25 m high, about 7,000 tonnes	General purpose; largest volume	More than 5,500 scientists, 245 institutes, 42 countries
CMS	21 m long, 15 m high, about 14,000 tonnes; 4 T solenoid	General purpose; compact design	About 5,500 people, 241 institutes, 54 countries
ALICE	26 m long, 16 m high, about 10,000 tonnes	Lead-ion collisions, quark-gluon plasma	Almost 2,000 scientists, 174 institutes, 40 countries
LHCb	21 m long, 10 m high, about 5,600 tonnes	Beauty quarks, matter-antimatter differences	About 1,565 people, 20 countries

The general-purpose design, shared by ATLAS and CMS, is an onion. A tracker at the center records the paths of charged particles, and a magnetic field bends those paths so the curvature reveals momentum. Calorimeters outside it absorb particles and measure their energy. Muon chambers at the outer edge catch the one kind of charged particle that passes through everything else.

Some particles, such as neutrinos and any hypothetical dark matter particles, pass through the whole detector without leaving a signal. Physicists find them by what is missing. In every collision, momentum perpendicular to the beam must balance to zero, so if the visible particles do not sum to zero, something invisible took the rest. CERN's dark matter page says exactly this: dark matter particles would carry away energy and momentum, and their presence could be inferred from the "missing" energy.

The detectors cannot record every crossing; the data would overwhelm any storage system. Fast electronics decide in a tiny fraction of a second which events look interesting and discard the rest. What survives is shared through the Worldwide LHC Computing Grid, which CERN describes as spanning over 170 sites in 42 countries. The web itself was invented at CERN to help researchers share information, a story told in the profile of Tim Berners-Lee.

Modern analysis also leans on machine learning to pick signals out of the noise, a thread picked up in AI at CERN and the Higgs detection.

ATLAS and CMS: two teams, one check

The two general-purpose detectors were built by independent teams on purpose. When both announced a signal on July 4, 2012, each had effectively reproduced the other's result with different hardware and different analysis methods. That independent cross-check carried a great deal of weight with the wider community.

ALICE and LHCb: the specialists

ALICE studies what happens when lead nuclei collide and form quark-gluon plasma. CERN says the LHC runs lead-ion collisions for part of each year. LHCb focuses on the beauty quark, a heavy cousin of the quarks inside protons, and on slight differences between matter and antimatter, which may bear on why the universe is made almost entirely of matter. Unlike ATLAS and CMS, which surround the collision point, LHCb detects mainly forward-going particles over a length of about 20 meters.

What the Higgs discovery proved, and what it did not

The Higgs boson is the signature particle of the Higgs field, an energy field that fills all of space. The field was proposed in 1964, and CERN credits Robert Brout, François Englert, and Peter Higgs as the theorists behind the mass-giving mechanism. Peter Higgs and François Englert shared the 2013 Nobel Prize in Physics for the theory.

One helpful picture: imagine a party where guests are spread evenly across the room. A celebrity walks in and gets mobbed, so she moves slowly. A person nobody recognizes crosses the room with no resistance. In this analogy, the field is the crowd, and a particle's mass reflects how strongly it is mobbed. CERN puts it plainly: the stronger a particle's interaction with the Higgs field, the heavier it is, and photons do not interact with the field at all, so they have no mass. The Higgs boson itself is a wave in the crowd, a localized ripple, rather than the crowd.

For a deeper treatment, see the Higgs boson explained and the story of the CERN discovery.

How sure were they?

Announcing a particle in July 2012 and confirming it as the Higgs boson were two different steps. CERN says physicists then checked the particle's spin, which must be zero for a Higgs, and after analyzing about two and a half times more data they concluded in March 2013 that it was a Higgs boson. Later results showed the particle decaying to tau leptons and, in 2018, to bottom quarks.

The mass itself has been refined. According to Wikipedia's summary, CMS measured 125.35 ± 0.15 GeV in 2022, and ATLAS reported 125.11 ± 0.11 GeV in 2023. The two numbers differ by a margin that physicists are still watching, but both are consistent with the original 125 to 127 GeV announcement.

Does the Higgs give everything its mass?

Not quite, and this is a common slip. The Higgs mechanism gives mass to elementary particles such as electrons and quarks. But a proton is a composite object, and the same Wikipedia page notes that about 91% of the proton's mass comes from quark and gluon fields and the strong-force energy that binds them, not from the Higgs interaction. So the Higgs explains why electrons have mass and why the weak force is short-ranged, but most of the mass of everyday matter comes from the strong force.

The nickname "God particle" has a simple origin. It comes from the title of Leon Lederman's 1993 book, whose author wanted to call it "the Goddamn Particle" before the publisher changed it. Many physicists dislike the name, and Higgs himself reportedly found it embarrassing.

What 15 years of data have, and have not, found

The popular summary is that the LHC found the Higgs and nothing else. That is not quite right, and it is not quite wrong either.

Dozens of new composite particles

By March 2021 the LHC experiments had found 59 new hadrons, and by February 2026 the count reached 79 across the first three runs, according to Wikipedia. These are not new fundamental particles. They are new combinations of known quarks, such as the pentaquark that LHCb reported on July 5, 2022 (made of a charm quark, a charm antiquark, and up, down, and strange quarks) and the first pair of tetraquarks announced the same day. They are useful because they test how the strong force assembles matter.

Supersymmetry has not appeared

The most prominent hope for new physics was supersymmetry. According to CERN, supersymmetry extends the Standard Model by giving every known particle a partner whose spin differs by half a unit. CERN notes that the partners could cancel out contributions that would otherwise drive the Higgs mass very high, and that the lightest partner could be stable, electrically neutral, and weakly interacting, which are the properties dark matter needs.

If the theory is right, CERN says, supersymmetric particles should appear in LHC collisions. So far no confirmed supersymmetric particle has been reported from the collisions analyzed. That rules out many of the simplest and most "natural" versions of the theory, but it does not rule out supersymmetry entirely. Particles may simply be heavier than the LHC can reach, or hidden in more complicated signatures.

Why a null result still counts

A null result in physics is a measurement. When ATLAS or CMS searches for a particle and finds nothing, it excludes a range of masses and couplings, and theorists must revise their models. That process has narrowed the options, which is progress, even if it feels like the opposite.

What the Standard Model still cannot explain

The Standard Model covers quarks, leptons, and three of the four fundamental forces: the electromagnetic, strong, and weak forces, carried by photons, gluons, and the W and Z bosons. CERN is blunt about its limits: gravity is not part of it, and it does not say what dark matter is. For a quick tour of the forces, see the four fundamental forces of nature.

According to CERN, ordinary matter makes up about 5% of the universe, dark matter about 27%, and dark energy about 68%. That means the Standard Model describes only the smallest slice. Dark matter in 2026 surveys the evidence and the candidates, and CERN's page points to candidates in theories such as supersymmetry and extra dimensions, including a "Hidden Valley" sector with little in common with ordinary matter.

Matter and antimatter are the other loose thread. The Big Bang should have produced equal amounts of both, yet the universe today is almost entirely matter. LHCb's job is to look for tiny differences in how matter and antimatter behave, and what antimatter is and why it matters explains the puzzle from the other end.

What comes next: a brighter LHC and a bigger ring

The LHC is not retiring so much as being rebuilt. After Run 3, Long Shutdown 3 begins the installation of the High-Luminosity LHC (HL-LHC), an upgrade that started in June 2018 and aims to raise luminosity by a factor of 10, to 10³⁵ per square centimeter per second, according to the Wikipedia summary. The goal is to collect far more collisions so that rare processes, such as pairs of Higgs bosons, can be measured. CERN says the upgraded machine will run for at least another decade. Wikipedia lists physics starting around 2030, though its own sections differ on the restart date, so treat that year as approximate.

Beyond that is the Future Circular Collider (FCC), a proposed ring of about 91 kilometers (90.7 km), roughly 3.4 times the LHC's circumference (my calculation: 90.7 divided by 26.7). CERN's FCC page says it would first run as an electron-positron "Higgs factory" and might later become a proton-proton machine with collision energies of up to 100 TeV. It estimates the cost at 15 billion Swiss francs over about 12 years, with the CERN Council due to decide in 2028 and start-up around 2045.

Machine	Size	Key goal	Timeline (per CERN unless noted)
LHC (Run 3)	26,659 m ring	13.6 TeV proton collisions	Physics from July 2022; ended mid-2026 (Wikipedia)
HL-LHC	Same ring	About 10 times the luminosity	Physics from about 2030 (Wikipedia); at least a decade of running
FCC	About 91 km	Higgs precision, then up to 100 TeV	Decision 2028; start-up around 2045

Other colliders exist or have existed, but none matches the LHC's energy; the full list is in the world's most powerful particle accelerators. The Wikipedia page also notes that preliminary FCC designs have ranged from about €9 billion to €21 billion depending on type, so the cost remains an open question.

Common confusions about the LHC

Several ideas about the collider keep circulating. Most come from a kernel of truth and a leap.

"It could make a black hole that eats Earth." CERN states that "according to Einstein's theory of relativity, it is impossible for LHC to produce black holes." Some speculative theories predict microscopic black holes, but even those would decay immediately. CERN adds that the LHC "can only reproduce phenomena that already happen naturally all around us," and the safety reviews note that cosmic rays hit Earth with energies far above anything the LHC reaches. The "Oh-My-God particle" carried about 320 million TeV (Wikipedia, citing the safety review). Readers interested in those natural collisions can start with cosmic rays and particle physics.

"It will create a mini Big Bang." It produces tiny, brief, hot regions that resemble early-universe conditions in some respects, and that is all.

"The Higgs is the God particle that explains everything." As above, it explains the mass of elementary particles, not the whole story of mass, and it says nothing about gravity or dark matter.

"The LHC failed because it did not find supersymmetry." The machine was built to explore a new energy range, and it was never promised to find a specific theory. Discovering the Higgs and excluding other ideas are both results.

"Nothing new is coming." The accumulating data may yet hold surprises, since many analyses are still being refined and the High-Luminosity upgrade will multiply the data set.

The takeaway

The LHC is best understood as a microscope that works by brute force. It turns a bottle of hydrogen into beams with a combined 13.6 TeV, steers them with 1,232 superconducting dipoles that sit at 1.9 K, and sorts a billion-odd collisions a second with detectors that are among the largest scientific instruments ever built. Its headline result, the Higgs boson, confirmed a mechanism proposed in 1964 and completed the Standard Model.

The harder part is what comes after. Dark matter, the matter-antimatter imbalance, and gravity all sit outside the theory that the LHC just completed, and the collider has not yet shown which direction leads past it. That uncertainty is the real return on 4.3 billion Swiss francs of construction: the field now knows where several attractive ideas are not, and the next decade of data will test whether the answer is more precision, a higher energy, or a different kind of experiment altogether.

FAQ

Q: Could the LHC create a black hole that destroys Earth? A: No. CERN says Einstein's relativity makes it impossible for the LHC to produce black holes, and that even speculative theories predicting microscopic ones say they would decay immediately. Two safety reviews found no danger, and nature already produces far higher-energy collisions when cosmic rays strike the atmosphere.

Q: How big is the team behind the LHC? A: ATLAS and CMS each list more than 5,000 members, ALICE almost 2,000, and LHCb about 1,565, according to CERN's pages from 2022. ATLAS alone lists 245 institutes in 42 countries.

Q: Is the LHC the most powerful collider ever built? A: Yes, in collision energy. Its 13.6 TeV proton collisions are the highest produced in a laboratory. CERN's proposed Future Circular Collider, a ring of about 91 kilometers, could reach up to 100 TeV in a later stage, but its construction would depend on a decision expected in 2028.

Q: Why does the LHC need superfluid helium? A: The magnets must stay at 1.9 K to remain superconducting. Superfluid helium flows without friction and carries heat away efficiently, so it keeps the windings cold even when beams and currents generate heat. The Wikipedia technical summary puts the amount at about 96 tonnes of helium-4.

Q: How do scientists find new particles in all that data? A: By looking for patterns. Most heavy particles decay almost instantly, so detectors record the energies, directions, and types of the decay products. Physicists then ask whether the pattern matches a known particle or shows an excess that would need a new one, and they require very strong statistical significance before announcing a discovery.

Q: What would finding supersymmetry prove? A: It would show that every known particle has a heavier partner, which could stabilize the Higgs mass and supply a dark matter candidate in the lightest partner. CERN notes the lightest partner would be stable and electrically neutral. It would not by itself complete the theory of everything, since gravity would still need a quantum description.

Q: Has the LHC found only the Higgs? A: No. It has measured the Higgs in several decay channels and discovered 79 new composite particles (hadrons) by February 2026, per Wikipedia. What it has not found is a confirmed particle beyond the Standard Model.

Sources

• CERN, Large Hadron Collider (key facts, safety, schedule)
• CERN, The accelerator complex
• CERN, The Higgs boson
• CERN, Future Circular Collider
• CERN, Dark matter, Supersymmetry, and The Standard Model
• CERN experiment pages: ATLAS, CMS, ALICE, LHCb
• Wikipedia, Large Hadron Collider and Higgs boson (2008 incident, beam parameters, Run 3, hadron counts, mass measurements)

---

Read next:

• What Is the Speed of Dark — And Other Things That Move Faster Than Light
• What Is Entropy — and Why the Universe Keeps Running Down
• What Is the Largest Known Prime Number — And Why Do Mathematicians Keep Searching?
• What Is the Biggest Number — And Why Infinity Breaks Mathematics