To Catch a Ripple in Space, Hang a Mirror and Keep It Very Still

A gravitational-wave detector rests on glass mirrors hung at the far ends of long laser arms. Its real trick isn't sensing motion. It's stopping almost every other kind of motion first.

Imagine an L laid out on the ground, each leg 4 kilometres long.

At the far end of each leg, and at the corner, hangs a mirror. Not on a hook, and not on a bracket bolted to the floor. It hangs at the bottom of a chain of weights, swinging free, in a structure built so that the ground it stands on barely reaches it.

That's the heart of a gravitational-wave detector. Everything else (the laser, the vacuum, the computers) exists to read out what those mirrors do. And what the mirrors are meant to do, nearly all the time, is nothing at all.

What actually hangs

Each of the big mirrors is a solid cylinder of pure silica glass weighing 40 kilograms. They're called test masses, and the name is the job description: they test for a passing wave. When one rolls through, they're what responds.

Each test mass hangs at the bottom of a pendulum with four stages, known in the trade as a quad. Think of a mobile above a baby's cot, except with four weights in a line, one below the other, and no cheerful animals. The top two weights are metal, 20 kilograms each. The bottom two are silica glass, 40 kilograms each. The lowest of those is the mirror.

The mirror's last link isn't wire. It hangs from glass threads just 0.4 millimetres thick, fixed to the glass with glass welds. Those threads don't swell or shrink as the temperature changes, which keeps the mirror clear of thermal noise: the faint jitter that heat itself sets off in a material.

Beside the main chain hangs a second chain of four, called the reaction chain. Its job is to steady the main one without touching it. Small motors made of magnets and electrical coils, which work like loudspeakers, push on its upper weights. At the mirror itself, the push comes from electrostatic forces, the same pull that makes your hair cling to a comb on a dry day.

The aim, in the builders' own phrase, is to make the mirrors "'feel' like they are floating in space". Not literally. But as close as a mirror on Earth can get.

Why stillness is the hard part

Here's what a gravitational wave does. As it passes, it stretches space in one direction and squeezes it in the other, then swaps. One arm of the L gets slightly longer while the other gets slightly shorter, and then the reverse.

So the mirrors move. Just not by much. The usual way to describe it is strain: the change in an arm's length as a fraction of its whole length. For a wave typical of what these detectors can pick up, the expected strain is likened to about one ten-thousandth the width of a proton.

The difficulty isn't spotting a mirror that moves. It's that everything moves (floors, buildings, the earth itself) and nearly all of it moves far more than a passing wave ever will. A detector this sensitive can feel road traffic close by, weather on the far side of a continent, people driving along the arms, the surge of the ocean, and almost every sizeable earthquake on the planet.

So the instrument's real work is subtraction. Strip out every shake that isn't the wave, and whatever's left might be.

A ruler made of light

To read the mirrors, a laser beam is split in two and half goes down each arm. At the far end, the beam reflects off a test mass, comes back, and bounces between the mirrors about 300 times before the two halves are brought together again.

When they meet, they're arranged to cancel each other out. Light plus light, lined up just so, gives darkness. A photodetector, an electronic eye that measures light, watches that dark spot.

Now let a wave pass. One arm lengthens, the other shortens. The two beams have travelled slightly different distances, so they come back out of step, no longer cancel cleanly, and a flicker of light reaches the photodetector.

That's why it works as a ruler. It isn't measuring how long each arm is; it's comparing the two, and only the difference shows up. Picture two people pacing out the same distance side by side. If they stay in step, there's nothing to see. If one stride stretches, the mismatch is obvious. Here, though, the stride is a wavelength of light, which is why the laser has to be so steady.

The arms sit inside a very clean vacuum, with the air pumped out, because stray molecules would scatter the beam and make false flickers. If even the tiniest speck of dust landed on a mirror and the laser struck it, the dust would burn up and permanently damage the mirror's surface.

Beating down the ground

The quad is a passive defence: it works by its shape, with nothing switched on. Each link in a pendulum soaks up some of the shaking that arrives from above and passes less to the link below. Heavy weights help too, because heavier things are harder to shake. Between the top of the chain and the mirror, vibration is cut to a hundred-millionth of what it was.

On its own, that still isn't enough. So the quad is mounted inside a bigger system that fights back. Seismometers, instruments that feel ground motion, pick up the shaking around the building. A computer works out the opposite motion, and hydraulic pushers move the platform to cancel the vibration before it reaches the pendulum. Noise-cancelling headphones, only for floors.

The active stage goes first. The passive chain cleans up whatever gets past it.

The day it worked

On 14 September 2015, the two detectors, thousands of kilometres apart, picked up the same signal. It was named GW150914, after its date.

The signal swept sharply upward in pitch over about two tenths of a second. It reached one detector first and the other about seven thousandths of a second later, a gap that fits the time light, or a gravitational wave, takes to travel between the two sites. A truck rumbling past one building couldn't fake that. Only a real wave turns up in both.

Nobody saw anything, mind you. There was no picture of black holes. What the detectors recorded was a changing difference in length between hanging mirrors, read out as flickers of light. Everything else came from matching that pattern against what theory predicts. The match fitted two black holes spiralling in and merging, in a distant galaxy more than one billion light years away. A noise event that strong, mimicking the signal by chance, would be expected only about once in 200,000 years of such data.

The result was published on 11 February 2016, as the first direct detection of gravitational waves and the first observation of two black holes merging.

A ruler with loose ends

Most rulers work by being rigid. This one works because its ends are loose: glass hanging on glass threads, four pendulum stages down, above a floor that's being cancelled out from underneath.

Get that right and, somewhere a billion light years away, two black holes can collide and nudge a mirror in an L-shaped building on Earth, one arm stretched, the other squeezed. The light notices.

SOURCES

1. LIGO Lab (Caltech), "LIGO Technology." https://www.ligo.caltech.edu/page/ligo-technology (accessed 24 Oct 2026, via web.archive.org)

2. LIGO Lab (Caltech), "Seismic Isolation." https://www.ligo.caltech.edu/page/vibration-isolation (accessed 24 Oct 2026, via web.archive.org)

3. LIGO Scientific Collaboration, "Observation of gravitational waves from a binary black hole merger" (science summary). https://ligo.org/science-summaries/gw150914/ (accessed 24 Oct 2026)

4. B. P. Abbott et al. (LIGO Scientific Collaboration and Virgo Collaboration), "Observation of Gravitational Waves from a Binary Black Hole Merger," Physical Review Letters 116, 061102 (2016). https://doi.org/10.1103/PhysRevLett.116.061102 (accessed 24 Oct 2026, via web.archive.org)