How does LIGO work?
LIGO splits a laser down two 4 km arms and recombines it. A passing gravitational wave changes the arm lengths by far less than a proton's width.
By The Orbital Record editorial teamPublished Last reviewed
LIGO works by measuring tiny changes in the length of two 4-kilometer arms with laser light. It splits a laser beam, sends the halves down arms set at right angles, bounces them off mirrors and recombines them so they cancel out. A passing gravitational wave stretches one arm and squeezes the other by far less than a proton’s width, and light leaks out.
LIGO, the Laser Interferometer Gravitational-Wave Observatory, made the first direct detection of gravitational waves in 2015. This explainer covers the machine itself; for what the waves are and what they have revealed, read our gravitational waves guide.
How does LIGO detect gravitational waves, step by step?
LIGO is a laser interferometer, an instrument that compares two beams of light to measure a difference in distance. Its technical name is a “dual recycled, Fabry-Perot, Michelson interferometer”. In order:
- The laser. A highly stable infrared laser enters the interferometer with about 40 watts of power.
- The beam splitter. A partly reflective mirror splits the beam into two equal halves, one for each arm.
- The arms. Each half travels 4 kilometers through a vacuum tube to a mirror, called a test mass, and back.
- The folding. Extra mirrors near the beam splitter turn each arm into a cavity. The light bounces about 300 times, so the effective path grows from 4 to about 1,200 kilometers.
- The recombination. The returning beams are recombined so that their light waves cancel exactly. With no signal, almost no light reaches the photodetector.
- The signal. A gravitational wave lengthens one arm while shortening the other, then reverses, many times a second. The beams fall slightly out of step and a flickering light pattern reaches the detector. That flicker, sampled over time, is the recorded waveform.
Recycling mirrors build up the stored laser power so that each measurement uses as many photons as possible. During the fourth observing run, O4, the arms held between about 280 and 370 kilowatts of circulating power, and Hanford briefly reached 400 kilowatts (Capote et al., Physical Review D, 2025).
How small is the signal?
The first gravitational wave LIGO detected, GW150914 on September 14, 2015, had a peak strain of 1.0 × 10-21: it changed lengths by one part in a billion trillion (the first detection). On a 4-kilometer arm, that is a few thousandths of the width of a proton. LIGO’s own pages put the change it is built to see at a thousandth of a proton’s width or less, with 1/10,000 given as the ultimate target.
No ruler could measure that. Light can, because the interferometer compares two beams whose waves line up to a tiny fraction of a wavelength.
How LIGO fights noise
Measuring such a small change means removing everything else that moves the mirrors. LIGO’s main defenses:
| Noise source | What LIGO does | Key figure |
|---|---|---|
| Ground vibration | Active platforms sense motion and push back with hydraulics, like noise-canceling headphones | Vibrations cut to at most 2 × 10-13 m before the pendulums |
| Remaining vibration | Each 40 kg test mass hangs from a four-stage pendulum of about 360 kg | Motion at the top is reduced about 100 million times at the mirror |
| Air | Beam tubes kept at about one-trillionth of an atmosphere | 10,000 cubic meters of vacuum per site |
| Mirror imperfections | Fused silica test masses, 34 cm across, with ultra-low-loss coatings | About $2 million per test mass |
| Quantum noise | Squeezed light reduces the random arrival of photons | Up to 65% more detections expected in O4 |
| Local glitches | Two sites far apart must both see the signal | 3,002 km apart, within about 10 ms |
Sources: LIGO Caltech facts, technology and vibration-isolation pages; Ganapathy et al., Physical Review X, 2023.
The 4-kilometer scale brings its own engineering. Over each 4-kilometer arm, Earth’s surface curves away by nearly a meter. The steel tubes are only 3 millimeters thick. Pumping the air out of each one took 40 days, in 1998 at Hanford and 1999 at Livingston.
Squeezed light: beating quantum noise
At high frequencies, LIGO’s sensitivity is limited by the quantum nature of light itself: photons arrive at random, which blurs the measurement. In 2019, during the third observing run, LIGO began injecting “squeezed” light, a quantum state with less uncertainty in the property that matters, improving sensitivity above 50 hertz by up to 3 decibels.
For O4, the A+ upgrade added a 300-meter filter cavity to each site so that the squeezing changes with frequency, cutting noise at high frequencies without adding it at low ones. The team reported that this boosted the detection range by 15 to 18% and the expected detection rate by up to 65%.
Why LIGO needs partners: Virgo, KAGRA and LIGO-India
Two detectors can confirm a signal, but they locate it poorly on the sky. More detectors, spread across the globe, let astronomers triangulate a source from arrival times, which is what made the GW170817 neutron star merger findable with telescopes. The network today:
| Detector | Location | Arm length | Notes |
|---|---|---|---|
| LIGO Hanford | Washington State, USA | 4 km | Operating |
| LIGO Livingston | Louisiana, USA | 4 km | Operating |
| Virgo | Cascina, near Pisa, Italy | 3 km | Hosted by the European Gravitational Observatory |
| KAGRA | Kamioka, Gifu, Japan | 3 km | Underground, with cryogenically cooled mirrors |
| LIGO-India | Aundha, Maharashtra, India | 4 km | Broke ground April 23, 2026; first observations targeted for 2030 |
Together the operating detectors form the LIGO-Virgo-KAGRA network. LIGO-India will copy the US design and, far from the other sites, sharpen sky positions (LIGO-India breaks ground).
What LIGO has found with this machine
The detectors started the O4 run on May 24, 2023, with a typical reach of 152 to 160 megaparsecs for neutron star mergers, about 500 million light-years. As of the GWTC-5.0 catalog (May 2026), the network has confirmed 390 gravitational-wave events (GWTC-5.0). The clearest so far, GW250114, was strong enough to test Stephen Hawking’s area theorem (GW250114).
Ground-based interferometers hear waves from tens to thousands of hertz. Slower waves need other instruments: pulsar timing arrays for waves lasting years, and the planned LISA mission, with arms 2.5 million kilometers long in space, for the hum of merging supermassive black holes. For the most spectacular events LIGO catches, see our explainer on neutron star mergers.
Questions people ask
How does LIGO detect gravitational waves?
LIGO splits a laser beam in two and sends the halves down two 4-kilometer arms at right angles. Mirrors bounce them back and they recombine so their light waves cancel. A passing gravitational wave stretches one arm and squeezes the other by a tiny amount, the beams fall out of step, and light leaks to a detector.
How sensitive is LIGO?
Extremely. The first signal LIGO detected, in 2015, changed distances by about one part in a billion trillion at its peak, so each 4-kilometer arm changed by a few thousandths of a proton's width. LIGO's own pages describe the target sensitivity as a change of 1/1,000 to 1/10,000 of the width of a proton.
Why does LIGO have two detectors?
To tell real signals from local noise. A truck, an earthquake or a power glitch affects one site only, while a gravitational wave reaches both. Hanford and Livingston are 3,002 kilometers apart, so a real wave arrives at both within about 10 milliseconds. Signals farther apart in time are rejected.
Why are LIGO's arms 4 kilometers long?
Because a gravitational wave changes lengths by a fixed fraction, so a longer arm means a larger, more measurable change. LIGO also folds the light inside each arm about 300 times between mirrors, which makes the effective path about 1,200 kilometers long without building a 1,200-kilometer tunnel.
What is squeezed light in LIGO?
Squeezed light is a quantum trick that reduces the random flicker of photons, which limits LIGO's sensitivity. LIGO first used it in its third observing run in 2019. For the fourth run, a 300-meter filter cavity made the squeezing frequency dependent, raising the expected detection rate by up to 65%.
Where are the LIGO detectors?
LIGO has two observatories in the United States: Hanford, in Washington State, and Livingston, in Louisiana. They work with Virgo near Pisa, Italy, and KAGRA in Kamioka, Japan. A third LIGO detector, LIGO-India, broke ground in Maharashtra on April 23, 2026, with first observations targeted for 2030.