The LIGO Gravitational Wave Detection: GW150914 Explained

At 5:51 AM Eastern time on September 14, 2015, a signal lasting 0.2 seconds swept through the LIGO detector in Livingston, Louisiana, and arrived 7 milliseconds later at the LIGO detector in Hanford, Washington. Both detectors registered the same characteristic chirp: a vibration growing rapidly in frequency and amplitude, peaking, then ringing down to silence. The source was two black holes — one of 36 solar masses, one of 29 solar masses — colliding and merging approximately 1.3 billion light-years away, converting 3 solar masses of matter into gravitational radiation in a fraction of a second. The event, designated GW150914, was the first direct detection of gravitational waves, confirming the last major untested prediction of Einstein's general theory of relativity.

Einstein's Prediction and Why It Took a Century to Test

Albert Einstein published the general theory of relativity in 1915. The theory replaced Newtonian gravity with a description in which mass and energy curve spacetime, and objects follow the straightest possible paths through that curved spacetime. One of the theory's immediate mathematical consequences — apparent in the linearized equations that describe weak gravitational fields — is that accelerating masses produce waves in the curvature of spacetime itself, propagating outward at the speed of light.

Gravitational waves — oscillations in the curvature of spacetime caused by accelerating masses, propagating outward at the speed of light. They alternately stretch space in one direction while compressing it in the perpendicular direction, and then reverse.

Einstein himself was ambivalent about whether gravitational waves were real or a mathematical artifact for years after their prediction. He and Nathan Rosen published a paper in 1936 claiming to show that exact gravitational wave solutions did not exist. The paper was rejected by Physical Review — one of the few times Einstein's work was rejected by peer review — and the referee (later identified as Howard Robertson) pointed out an error. Einstein and Rosen corrected the mistake and ultimately confirmed that gravitational waves are real physical phenomena. The episode illustrates that even Einstein needed peer review.

The energy emitted as gravitational radiation by ordinary sources — two people orbiting each other, a spinning asteroid — is utterly undetectable with any conceivable instrument. The only sources powerful enough to produce detectable gravitational waves are the most compact, massive objects in the universe moving at substantial fractions of the speed of light: neutron stars orbiting each other, black holes merging, or the collapse of a massive star's core into a supernova. About 20 km wide but containing more mass than the Sun.">neutron star.

Source: LIGO Lab Caltech — "GW150914 Press Release" (ligo.caltech.edu/page/press-release-gw150914)

Indirect Evidence: The Hulse-Taylor Binary Pulsar

The first evidence for gravitational waves came in 1974, not from any detector but from a pulsar. Russell Hulse and Joseph Taylor discovered a pulsar (a rapidly spinning neutron star that emits a radio pulse every rotation) in a close binary orbit with another neutron star. This system — called PSR B1913+16 — was ideal for testing general relativity: two compact, massive objects in a tight orbit, with the pulsar acting as an extremely precise clock.

General relativity predicts that two massive objects in a close orbit radiate energy as gravitational waves, causing their orbit to shrink gradually. The orbital period decreases predictably as energy is lost. Hulse and Taylor measured the orbital period change over years and found that it matched the general relativistic prediction for gravitational wave emission to within 0.2%. The system behaved exactly as if it were radiating gravitational waves at the predicted rate, even though the waves themselves could not be detected directly.

Hulse and Taylor received the 1993 Nobel Prize in Physics for this discovery. The Hulse-Taylor pulsar provided compelling indirect evidence that gravitational waves exist and carry energy at the predicted rate, but the waves themselves still needed to be directly detected.

Source: Physics World — "LIGO detects first ever gravitational waves — from two merging black holes" (physicsworld.com/a/ligo-detects-first-ever-gravitational-waves-from-two-merging-black-holes/)

The Instrument: How LIGO Measures a Change Smaller Than a Proton

LIGO — the Laser Interferometer Gravitational-Wave Observatory — is built on a principle proposed independently by Rainer Weiss at MIT in 1972: a Michelson interferometer with arms long enough to measure the strain produced by a passing gravitational wave.

Strain (h) — the dimensionless amplitude of a gravitational wave, defined as h = ΔL/L, where ΔL is the change in arm length and L is the arm length. For GW150914, h was approximately 10⁻²¹.

A Michelson interferometer splits a laser beam into two perpendicular beams, sends each down a long arm, reflects them off mirrors at the far ends, and recombines them at the beamsplitter. If both arms are exactly the same length, the recombined beams cancel perfectly (destructive interference) and no light reaches the detector. If a gravitational wave passes through the interferometer, it stretches one arm while compressing the other, changing their relative lengths, and the interference pattern changes — light reaches the detector in proportion to the arm length difference.

Each LIGO arm is 4 kilometers long. At the strain of GW150914 (h ≈ 10⁻²¹), the arm length change is:

ΔL = h × L = 10⁻²¹ × 4,000 m = 4 × 10⁻¹⁸ m

This is approximately 1/1,000th the diameter of a proton (the proton diameter is about 10⁻¹⁵ m). Measuring a displacement this small requires laser interferometry pushed to its absolute limits.

Source: arXiv — "Properties of the Binary Black Hole Merger GW150914" (arxiv.org/pdf/1602.03840)

Source: LIGO Lab Science Description (labcit.ligo.caltech.edu/~rana/docs/LIGO_Science.pdf)

The Engineering: How LIGO Achieves 10⁻¹⁸ Meter Sensitivity

Achieving sensitivity to displacements of 4 × 10⁻¹⁸ meters required solving noise problems across multiple orders of magnitude. The main noise sources and the solutions developed for each are what make LIGO a technological achievement on the same scale as its scientific one.

Seismic noise — the ground vibrates continuously due to ocean waves, traffic, wind, and microseismic activity. At the frequencies of interest for gravitational waves from compact binary mergers (roughly 10 Hz to 1 kHz), ground vibrations are orders of magnitude larger than the signal. Solution: four-stage pendulum suspension systems hang the mirrors from a quadruple pendulum chain. The pendulum acts as a low-pass filter for vibrations: each pendulum stage suppresses vibrations at frequencies well above its resonance by a large factor. Four stages give isolation of roughly 10¹² at 100 Hz.

Thermal noise — atoms in the mirror and its suspension vibrate due to temperature. This thermal motion shows up as random fluctuations in the mirror position. Solution: use mirror substrates made of ultrapure fused silica, which has very low mechanical loss (vibrations damp slowly, concentrating the thermal noise at specific well-known frequencies rather than spreading it across the band). Operate at room temperature rather than cryogenic temperatures because of other practical constraints, and design the system so that thermal noise peaks fall outside the most sensitive frequency band.

Quantum noise (shot noise and radiation pressure) — at high frequencies, the measurement is limited by the quantum nature of light: detecting fewer photons means more statistical fluctuation in the interference signal. At low frequencies, radiation pressure from the photons themselves pushes the mirrors around randomly. Solution: increase laser power to reduce shot noise (Advanced LIGO uses 100 kilowatts of circulating power inside the interferometer, built up from an input of about 20 W using a power-recycling mirror cavity), and use squeezed light states — quantum optical techniques that reduce uncertainty in one observable at the cost of another.

Scattered light — stray laser reflections off surfaces can return to the beam path and introduce noise. Solution: optical baffles, vibration-isolated beam tubes (the 4-km tubes are evacuated to a pressure of about 10⁻⁹ torr), and careful optical design.

Noise Source	Frequency Range	Primary Solution
Seismic	< 10 Hz	Quadruple pendulum suspension
Suspension thermal	~10–50 Hz	Low-loss silica suspension fibers
Mirror thermal	~40–200 Hz	Low-loss mirror substrate, high laser power
Quantum (shot)	> 100 Hz	High circulating power, squeezed light
Quantum (radiation pressure)	< 50 Hz	Squeezed light (balanced)

Source: Physical Review Letters — "GW150914: The Advanced LIGO Detectors in the Era of First Discoveries" (link.aps.org/pdf/10.1103/PhysRevLett.116.131103)

The History of LIGO: Four Decades from Idea to Detection

LIGO was not built quickly. The gap between Weiss's 1972 proposal and the first detection in 2015 is 43 years — a span that included multiple rounds of funding battles, technological dead ends, leadership transitions, and fundamental redesigns.

Weiss began prototyping small interferometers at MIT in the early 1970s. Kip Thorne at Caltech engaged Ron Drever from Glasgow, a skilled experimentalist, to lead a parallel effort. For years, the MIT and Caltech groups competed as much as they collaborated. In 1984, the two groups formally merged under an NSF grant to form the LIGO project, with Weiss and Drever as co-principal investigators and Rochus Vogt as director.

Progress was slow and contentious. In 1994, NSF administrator Barry Barish took over as director, reorganizing LIGO from a small academic collaboration of roughly 40 people into the LIGO Scientific Collaboration (LSC) — an international consortium that would eventually include over 1,000 researchers from 18 countries. Barish's management restructuring is widely credited as the institutional change that made the project viable.

Construction of the two observatories — Hanford, Washington, and Livingston, Louisiana — began in the mid-1990s. The first generation of LIGO detectors (Initial LIGO) ran from 2002 to 2010 without detection, at sensitivities insufficient to reach the distances where black hole mergers were expected to be frequent enough to observe. The Advanced LIGO upgrade, which replaced essentially every component of the detectors, ran from 2010 to 2015 and achieved roughly 10 times better strain sensitivity — covering a volume of space about 1,000 times larger than Initial LIGO.

Source: Nobel Prize — "Press release: The 2017 Nobel Prize in Physics" (nobelprize.org/prizes/physics/2017/press-release/)

GW150914: The Signal Itself

At 5:51 AM Eastern time on September 14, 2015, with Advanced LIGO running in "engineering mode" (final testing before official science operations), the signal arrived. The Livingston detector registered it first; Hanford registered the same pattern 7 milliseconds later — consistent with a signal traveling at the speed of light between the two sites, 3,002 km apart.

The signal had a characteristic shape: amplitude and frequency both increasing rapidly over about 0.2 seconds, reaching a peak, then damping away in a "ringdown" as the newly formed merged black hole settled into its final state. This is the gravitational wave equivalent of a struck bell — the event creates the wave, the wave carries energy, the source loses energy and settles into a stable configuration.

The measured strain at peak was h ≈ 10⁻²¹. The frequency swept from about 35 Hz at the start to about 150 Hz at the peak — the characteristic "chirp" sound when the data is converted to audio (the signal is in the human hearing range).

The two black holes had masses of 36+5/-4 solar masses and 29+4/-4 solar masses respectively, from fitting the signal to general relativistic waveform templates. They merged to form a black hole of 62 solar masses. The difference — 3 solar masses — was emitted as gravitational radiation. In SI units, 3 solar masses × c² ≈ 5.3 × 10⁴⁷ joules of energy, radiated in about 0.1 seconds. At peak, the gravitational wave power output of the merger was approximately 3.6 × 10⁴⁹ watts — more than the combined electromagnetic luminosity of all observable stars in the universe at that moment.

The merger happened approximately 1.3 billion light-years away — the signal had been traveling since the Ediacaran period on Earth, when multi-cellular life was just beginning to evolve in the oceans.

Source: arXiv — "Properties of the Binary Black Hole Merger GW150914" (arxiv.org/pdf/1602.03840)

Source: LIGO Scientific Collaboration — "Observing gravitational-wave transient GW150914 with minimal assumptions" (dcc.ligo.org/ligo-p1500229/public)

A Timeline of Gravitational Wave Physics

Year	Event
1915	Einstein publishes general theory of relativity
1916	Einstein derives gravitational wave solutions from linearized GR
1936	Einstein and Rosen submit paper claiming no exact GW solutions exist; referee finds error; paper corrected
1957	Felix Pirani gives rigorous analysis of gravitational wave detectability; Conference at Chapel Hill establishes GW reality
1972	Rainer Weiss proposes laser interferometric GW detector; estimates noise sources
1974	Hulse and Taylor discover PSR B1913+16 binary pulsar
1984	MIT and Caltech groups formally merge to form LIGO project
1993	Hulse and Taylor receive Nobel Prize for pulsar evidence of GW emission
1994	Barry Barish restructures LIGO as large international collaboration
2002	Initial LIGO detectors begin science runs
2010	Initial LIGO concludes; Advanced LIGO upgrade begins
Sep 14, 2015	GW150914 detected at 5:51 AM Eastern time
Feb 11, 2016	LIGO and Virgo collaborations announce GW150914 publicly
Jun 2016	Second detection: GW151226 from smaller-mass black hole merger
2017	GW170817: first neutron star merger detected; simultaneous gamma-ray burst observed
Oct 2017	Nobel Prize in Physics awarded to Weiss, Barish, and Thorne
2023	Pulsar timing arrays report evidence for gravitational wave background

Source: AAS Nova — "LIGO Discovers the Merger of Two Black Holes" (aasnova.org/2016/02/11/ligo-discovers-the-merger-of-two-black-holes/)

Confirming That No Natural Phenomenon Could Mimic the Signal

The most important verification was not that both detectors saw the signal, but that the signal matched a theoretical prediction so precisely.

General relativity gives exact predictions for the waveform shape produced by two black holes of given masses at a given distance, spiraling inward and merging. The computation of these waveforms required decades of work on numerical relativity — solving Einstein's field equations computationally, including the strongly nonlinear regime near merger where perturbative approximations fail. By 2015, the collaboration had waveform templates for millions of possible source configurations pre-computed.

The GW150914 signal matched a template from a binary black hole merger with masses of approximately 36 and 29 solar masses with a signal-to-noise ratio of 24 — an extremely high confidence level. The probability of a noise fluctuation producing a signal this significant has been calculated at about 1 in 3.5 million (a 5.1-sigma detection, equivalent to a p-value of about 2.9 × 10⁻⁷).

The 7-millisecond time delay between the two detectors is consistent with a signal traveling at the speed of light from the direction on the sky where the analysis placed the source. No known terrestrial source — earthquake, industrial activity, atmospheric disturbance — would produce a matched waveform at both detectors simultaneously with a light-travel-time delay.

Source: Einstein-Online — "Observation of Gravitational Waves from a Binary Black Hole Merger" (einstein-online.info/en/spotlight/observation-of-gravitational-waves-from-a-binary-black-hole-merger/)

GW170817: The Neutron Star Merger That Opened Multi-Messenger Astronomy

The detection of GW150914 opened gravitational wave astronomy. The detection of GW170817 on August 17, 2017 — a merger of two neutron stars — established a new field: multi-messenger astronomy, where the same event is observed through multiple physical signals simultaneously.

The neutron star merger produced gravitational waves (detected by LIGO and the European Virgo detector) and a short gamma-ray burst detected 1.7 seconds later by the Fermi and INTEGRAL gamma-ray telescopes. The 1.7-second delay is consistent with the gamma-ray burst being produced by material ejected during the merger as it collides with surrounding matter, slightly after the gravitational wave emission from the merger itself.

The event was subsequently observed across the electromagnetic spectrum — from gamma rays to radio waves — by over 70 observatories worldwide. Spectroscopic analysis identified heavy elements (gold, platinum, strontium, cesium) in the explosion, confirming that neutron star mergers are a primary site of r-process nucleosynthesis — the process that creates elements heavier than iron. The golden ring you may own likely contains gold forged in a neutron star merger billions of years ago, and GW170817 provided the first direct observational confirmation of this.

Source: Nobel Prize Popular Information — "2017 Nobel Prize in Physics: Cosmic Chirps" (nobelprize.org/prizes/physics/2017/popular-information/)

What People Get Wrong About LIGO

The most common misconception is about scale. People hear "a change of 10⁻¹⁸ meters" and think of it as a measurement of a mirror moving by that amount. More precisely, LIGO measures the differential length change between two 4-km arms. The arms are 4 km long — far larger than a proton — and the relevant quantity is the ratio ΔL/L = 10⁻²¹. The absolute displacement ΔL is about 4 × 10⁻¹⁸ m, but it is the ratio that the instrument is engineered to measure.

A second misconception is that LIGO measures the position of the mirrors directly. It does not. It measures the phase difference in laser light returning from two arms. The relationship between phase difference and arm length difference is derived from the wavelength of the laser (1064 nm), the number of bounces the light makes in the arm cavity (about 300), and the calibration of the photodetector. The mirrors' positions are inferred from the light, not measured directly.

A third common error involves what "space being stretched" means physically. The arm length changes because spacetime itself is distorted by the passing wave — not because the mirrors are pushed by a force in the ordinary mechanical sense. The light travel time changes as the wave passes, and it is that change in travel time that produces the interference signal. In an important sense, the laser wavelength is also stretched by the gravitational wave, but the stretching of the arms wins because of the way the interferometer is calibrated and because the laser wavelength is not affected by the local spacetime in the same way as the arm length at the frequencies involved. The detailed physics of why this works requires care with general relativistic optics that popular accounts generally skip.

A fourth misconception: LIGO detected gravitational waves, not gravity waves. Gravity waves (no "ational") are a different phenomenon — oscillations in fluid density driven by buoyancy, like the waves in Earth's atmosphere that cause weather. The two terms are not interchangeable, and conflating them produces genuine confusion in Earth sciences and fluid dynamics.

For the theory that predicted gravitational waves, see how Einstein's general relativity works. For how the detected black holes compare to other black holes in the universe, see types of black holes.

FAQs

Q: How can LIGO measure a displacement smaller than a proton?
A: By using laser interferometry with very high precision and very long arms. The key is that LIGO measures the difference in round-trip light travel time between two 4-km arms, each of which bounces the laser about 300 times (building up an effective path length of about 1,200 km). A change of 4 × 10⁻¹⁸ m in arm length shifts the phase of the light by a detectable amount when accumulated over this effective path length. Quantum noise, seismic noise, and thermal noise are suppressed through engineering solutions developed over 40 years.

Q: Why are there two LIGO detectors rather than one?
A: A single detector cannot distinguish a genuine gravitational wave signal from local noise events (trucks driving past, an earthquake microseism, a power glitch). Two detectors at widely separated locations (Hanford, Washington and Livingston, Louisiana, about 3,000 km apart) can be compared: a gravitational wave signal arrives at both with a time delay consistent with light-speed propagation, while noise at one detector is uncorrelated with noise at the other. The coincidence requirement dramatically reduces the false alarm rate.

Q: Why did it take 40 years to build a detector after the idea was proposed?
A: Several factors. The required sensitivity (h ~ 10⁻²¹) was far beyond what any technology could achieve in 1972. Developing the necessary laser stability, vibration isolation, mirror surface quality, and vacuum systems required decades of incremental advances. Funding for such a long-horizon, high-risk project was chronically difficult to secure. The project also went through significant management difficulties in the late 1980s and early 1990s before Barish reorganized it. The 40-year gap reflects both genuine technological difficulty and the organizational complexity of building a billion-dollar instrument with no guarantee of detection in a given time frame.

Q: What happens to the energy emitted as gravitational waves?
A: It propagates outward from the source as a wave in spacetime, spreading over an ever-larger sphere and decreasing in amplitude as 1/distance. The 3 solar masses of energy emitted by GW150914 are now spread over a sphere of radius 1.3 billion light-years and expanding at the speed of light. The energy density at our location when the wave passed Earth was extraordinarily low — equivalent to the energy output of a small light bulb spread over the area of a football field. This is why detecting the displacement requires a 4-km interferometer; the energy density is small but the spacetime distortion is measurable with a long enough baseline.

Q: What does a gravitational wave actually sound like?
A: The GW150914 signal, when the frequency data is converted to audio (stretched in time to bring it into the audible range), sounds like a brief chirp — a rising tone that peaks and then fades quickly. The frequency swept from about 35 Hz to about 150 Hz during the 0.2-second signal. Human hearing covers roughly 20 Hz to 20 kHz, so the signal falls in the low hearing range. The LIGO collaboration released audio files of the signal; it is often described as sounding like a brief "blip" or "zap."

Q: Can LIGO detect gravitational waves from sources other than black hole and neutron star mergers?
A: Yes, in principle, though other sources are weaker. LIGO could detect gravitational waves from a supernovae core collapse in or near the Milky Way, from rapidly spinning neutron stars with surface asymmetries (continuous wave sources), or from the stochastic background of gravitational waves from many overlapping sources. None of these have been detected by LIGO as of 2025. In 2023, pulsar timing arrays (which use millisecond pulsars as precision clocks to measure gravitational wave effects over years-long timescales) reported evidence for a gravitational wave background likely from supermassive black hole mergers.

The Takeaway

GW150914 confirmed Einstein's last major untested prediction from 1916 — that accelerating masses produce waves in spacetime — and opened an entirely new observational window on the universe. In the century before 2015, all astronomy was done in electromagnetic radiation: light, radio waves, X-rays. Gravitational wave astronomy hears phenomena that are invisible in every wavelength of light — black hole mergers, the innermost cores of collapsing stars, the dynamics of neutron star interiors. The universe is now observable in a mode it had never been observed in before. LIGO built that instrument over 40 years, with thousands of scientists, at a cost of about $1.1 billion — and on September 14, 2015, it heard two black holes collide a billion light-years away, in a signal that lasted 0.2 seconds and displaced its mirrors by a distance a thousand times smaller than a proton.

---

Read next:

• The Cosmic Microwave Background Discovery
• The Hubble 1929 Redshift Observation
• The Bohr-Einstein Debates Thought Experiment