Why physics needs such enormous machines to study the smallest things

The Short Version

Seeing smaller things takes more energy, and delivering that energy to a single proton takes a very large machine. Lisa Randall puts the physics in four words: "Higher energy means shorter distances" [3]. Harry Cliff explains why the Large Hadron Collider is a ring 27 kilometres around: protons moving at nearly the speed of light have to be bent into a circle, the magnets that do it could be made only so strong, and so the circle had to be that wide [1]. Don Lincoln explains the rest of the bill. Interesting collisions are so rare that the machine must produce about a billion a second, and the detectors that watch them weigh thousands of tons and discard nearly everything they see [2]. All three say each further step gets harder, and Randall adds that the history of these machines is also a warning about deciding in advance what you will find [1][2][3].

Table of contents

What They Said

Small things need high energy

Cliff, a particle physicist at Cambridge who works on one of the collider's experiments, calls the machine a gigantic microscope. The analogy, he says, is fairly precise, except that "it's a microscope that looks at the structure of the vacuum" [1].

He means something specific. Particle physics, in his view, is "kind of badly named," because particles are not the basic ingredients of the universe. The real building blocks are quantum fields, invisible things that fill all of space in the way a magnetic field surrounds a magnet. Every electron is a ripple in one electron field [1]. To study a field you have to hit it hard enough to make it ripple. The Higgs field was found by colliding protons with enough energy to disturb it, and the disturbance showed up as the Higgs boson [1].

Randall, a theoretical physicist at Harvard, gives the two reasons energy matters. The first is resolution: "Higher energy means shorter distances" [3]. The second is Einstein's relation between energy and mass. "So if you have a heavy particle and you don't have enough energy to make it, you'll never see it" [3].

Lincoln, who works at Fermilab near Chicago, describes the same conversion from the experimenter's side. A collider turns the energy of motion into new particles, always in pairs of matter and antimatter. Below a certain energy a given particle simply cannot be made. Just above it, you make a few. "More is better" [2]. The heaviest known particle, the top quark, weighs about as much as 175 protons, and the carriers of the weak force weigh 80 to 90 protons each [1]. None of these can be produced on a laboratory bench.

High energy needs a big ring

Why does a higher-energy machine have to be bigger? Cliff starts with the two possible designs, a long straight line or a ring. The advantage of a ring is that the particles come round again, and "you can give them a kick every time they go round" [1]. In one section of the collider only about 30 metres long, metal boxes hold oscillating electric fields of around two million volts, timed so that each passing bunch of protons is pulled in and pushed out the far side [1].

He says the acceleration is not the hard part. The limit is steering. The protons travel at more than 99.99 percent of the speed of light, and like a car taking a fast corner they need something to hold them on the curve. Magnets do that, and the collider was "designed with the most powerful magnets that could conceivably be built at the time" [1].

From there the size follows from one trade-off: "The bigger the tunnel, the weaker the magnets can be, the smaller the tunnel, the stronger they've got to be" [1]. A tighter bend at the same energy would need magnets nobody could build. So the designers took the strongest magnets available and made the ring as wide as that required, in a tunnel that had already been dug for an earlier machine [1].

Useful collisions are rare, so there must be billions

Getting two protons to the right energy is half the task. They also have to hit each other, and they almost never do.

Cliff describes the beams as about 2,000 bunches, like swarms of bees. Each bunch holds around 100 billion protons and is about 30 centimetres long and thinner than a human hair. When two bunches cross, perhaps ten protons actually collide [1]. The answer is repetition: 40 million bunch crossings pass a given point every second [1].

Lincoln uses the same image and adds the totals. "You should think of this as like a swarm of bees," he says. Mostly the bees pass through each other, and now and then two hit head on [2]. He counts about 20 collisions per crossing, which with 40 million crossings a second gives roughly a billion collisions every second [2].

Even those are mostly uninteresting. "Most collisions are absolutely boring," Lincoln says, because they show physics that has been tested for decades [2]. The task is "to pick out the cool ones, the weird ones, the ones that nobody's seen before" [2].

His own career shows why the numbers matter. When the top quark was discovered at Fermilab in 1995, his team's paper had 38 candidate events after six months to a year of collecting data, and about half were background. "So we had maybe 19 top quarks," he says [2]. The Large Hadron Collider, with about seven times the energy and a hundred times more collisions per second, may make a top quark every second. "And that's what higher energy and more collisions per second will do for you" [2]. Top quarks have gone from a discovery to a nuisance in the data [2].

Protons add a complication. They are not simple points. Lincoln says "they're kind of like garbage cans full of stuff," so what actually collides is a piece of one with a piece of another, and the outcome cannot be dialled in the way it can with electrons [2].

The detectors are machines of their own

A collision is useless unless someone records what came out, and the recording apparatus is on the same scale as the accelerator.

Lincoln works on a detector that weighs 14,000 tons and stands five storeys tall [2]. Cliff describes the two general-purpose detectors as barrels built in layers like an onion, each layer measuring a different property of the particles flying out, and each run by about 3,000 people [1]. His own experiment is shaped as a cone pointing along the beam, with a delicate silicon sensor about seven millimetres from a beam that carries as much energy as a jumbo jet at take-off [1].

Then there is the data. Nobody can store a billion collisions a second. Cliff says that keeping everything would fill every computer in the world within days [1]. Fast electronics called triggers decide, 40 million times a second, whether a crossing looks interesting. In Lincoln's account they cut 40 million snapshots a second to about 100,000, and computer farms then keep roughly 1,000 a second for people to study [2]. Cliff sees room for improvement here from machine learning applied to the raw detector signals, which could keep more of the useful data without new hardware [1].

It takes decades, countries and thousands of people

Cliff puts the full cost of the Large Hadron Collider at around 10 to 12 billion euros [1]. It was first discussed in the late 1970s and took about two and a half decades of campaigning before construction began around 2000 [1]. The laboratory that hosts it was founded in the 1950s, with member countries paying a small share of their national income [1].

The organization is as unusual as the hardware. Cliff's experiment has about 800 collaborators who are also rivals racing to publish. Its elected spokesperson cannot fire anyone, so decisions come from consensus and many meetings [1]. He calls it possibly the best example he can think of of "international collaboration that isn't for some unpleasant purpose" [1]. Randall notes the same origin, in an effort after the Second World War to get European physicists working together again, and mentions a later project that brought Palestinian and Israeli scientists together [3].

The scale brings its own risks. About ten days after the collider's launch party in 2008, an explosion badly damaged about half a kilometre of the machine. Engineers at first took the wall of alarms for a software fault, because a failure that large seemed impossible [1].

Each step up is harder than the last

Lincoln gives the rate of progress: collider energy has grown about sevenfold every twenty years [2]. Against that he sets the distance to the energies where a final theory might be tested, which is about a quadrillion times beyond today's machines. He does not expect the answer in the lifetimes of his grandchildren's grandchildren, and he calls extrapolating across that gap "the pinnacle of arrogance" [2].

Cliff makes the same point with a picture. The collider reaches 14 tera-electronvolts. The scale at which the strings of string theory would show themselves is about fourteen orders of magnitude higher. With present technology, he says, that would take an accelerator circling the Milky Way [1].

The nearer future is expensive enough. The planned upgrade raises the collision rate about tenfold, since each further year of running at the present rate adds less and less [1]. Beyond that, the favoured proposal is a new ring 100 kilometres around, which Cliff puts at about 30 billion euros over roughly half a century, with no data before about 2040 [1]. Putting stronger magnets in the existing tunnel is regarded as a dead end, and a compact straight-line machine for electrons is another option [1].

Lincoln adds an illustration of how costly it is to make anything exotic this way. At Fermilab, "we would have to smash 100,000 protons into something to make one antimatter proton," which works out to about a billionth of a gram a year [2].

What the money bought, and Randall's lesson

Randall calls the Large Hadron Collider "a major victory." The Higgs boson was proposed fifty years before it was found, and the mechanism that seemed the only way to explain the masses of elementary particles turned out to be right [3]. Cliff says that by the late 1990s the theory was like a jigsaw with every piece in place except one [1].

Then Randall adds the other half: "it was also a cautionary tale" [3]. When she started in physics the United States had proposed a larger machine, the Superconducting Super Collider. Many physicists, particularly in Europe, argued that the smaller European machine would have the reach to discover what lies beneath the known particles, and many thought they knew what it would be, a theory called supersymmetry [3]. That confidence was one of the factors, she says, in the European collider becoming "the only machine in town" after the American one was cancelled, a machine that would have been a much more robust test [3].

Supersymmetry has not appeared. Cliff says that after a decade of data no new particles of any kind have shown up, and that many physicists now see the chance of a direct discovery soon as slim [1]. Randall says finding the long-favoured dark matter candidate at the collider now looks very unlikely [3]. Her conclusion: "you don't want to assume things before they've been actually found" [3].

Where They Agree, and Where They Don't

Energy is the price of resolution. Randall states it, Cliff explains it through fields, and Lincoln through the making of heavy particles [1][2][3].

Rate matters as much as energy. Cliff and Lincoln both describe the swarms of bees and the need for vast numbers of crossings [1][2].

The Higgs discovery was real and expected. All three treat it as a confirmed success [1][2][3].

The next step is far harder. Lincoln's sevenfold-per-twenty-years and Cliff's galaxy-sized accelerator make the same point [1][2].

The collaboration is an achievement. Cliff and Randall both say so [1][3].

They differ in emphasis on three matters.

How much the Higgs changed. Randall calls it a major victory [3]. Lincoln calls it the last unvalidated piece of the existing theory, important and not a shift in worldview [2]. Cliff calls it both a triumph and a puzzle, because the value of the Higgs field appears finely tuned in a way he describes as "deeply fishy" [1].

Was one machine enough? Randall regrets the loss of the larger American collider [3]. Cliff describes the European machine's long struggle for approval and its present plans without that comparison [1]. Both are consistent with the fact that the most hoped-for discoveries did not come.

Bigger, or more careful? Randall names two routes to new physics, higher energy and higher precision: "If it's been tested 1%, you want to look at a tenth of a percent" [3]. Cliff's experiment takes the second route. He compares it to tracking a rare elephant by its footprints, and says of his group, "We are the footprint people." The measurements of one kind of quark are, he says, the one place at the collider with fairly compelling hints of something beyond the standard theory [1]. Lincoln favours chasing anomalies in general, and notes that his own laboratory now concentrates on intense beams for studying neutrinos instead of the highest energy [2]. Randall also points to detectors deep underground that search for dark matter directly [3]. Cliff adds that if the hints from precision work hold up, a bigger machine would still be needed to produce the new particle itself [1].

Why It Works

Energy probes distance. A higher-energy collision resolves a shorter length [3].

Energy becomes mass. A particle can be made only if the collision carries at least its mass in energy [2][3].

A ring reuses its accelerator. Each lap adds another push [1].

Magnets set the radius. For a given magnet strength, more energy means a wider turn [1].

Probability demands volume. Bunches mostly miss, and the events of interest are a tiny fraction of the collisions that occur [1][2].

Filtering replaces storage. Triggers throw away almost everything at once, because nothing else is possible [1][2].

These are explanations from working physicists. The figures they give for sizes, rates and costs are their own statements in conversation, and Cliff notes that he is not an engineer [1][2].

Where It Applies, and Where It Doesn't

The argument is about colliders. Not all progress needs one. Precision measurement, underground detectors, neutrino beams and telescopes are other routes named in these conversations [1][2][3].

Size does not guarantee discovery. The collider found the particle it was expected to find and none of the ones many hoped for [1][3].

Some questions are out of reach. The energies at which string theory could be tested are beyond any machine that can be imagined with current technology [1][2].

Costs are spread over decades and many countries. Cliff's figure for the proposed new ring covers about half a century [1].

Estimates differ in detail. Cliff's cost figures were given before the pandemic, and the conversations were recorded years apart [1].

What To Do With It

1. Ask what energy a claim requires. When you read that a theory will be tested, check whether the test needs energies a real machine can reach [1][2].

2. Ask about rate as well as energy. A discovery needs many events, so the number of collisions per second matters as much as their force [2].

3. Look at precision experiments. The results that currently puzzle physicists come from careful measurement of known particles [1].

4. Treat predictions of what a machine will find with caution. Randall's lesson from supersymmetry is to design a search broad enough to be surprised [3].

5. Judge the cost over its lifetime. Divide a headline figure by the decades and the countries involved before comparing it with other spending [1].

6. Count what is learned by not finding something. Ruling out a favoured idea, as with the expected dark matter particle, redirects the field [3].

7. Notice the engineering and the cooperation. Both Cliff and Randall count the technology and the long international commitment among the results [1][3].

Sources

1. Harry Cliff on Particle Physics, the Higgs Field and the Large Hadron Collider · Lex Fridman Podcast
2. Don Lincoln on the Biggest Mysteries in Physics: Unification, the Higgs, Antimatter, Dark Energy, and Dark Matter · Lex Fridman Podcast
3. Lisa Randall on Dark Matter, the Dinosaur Extinction, and the Edges of Physics · Lex Fridman Podcast