A mass curves the spacetime around it. When masses move and their arrangement changes, the curvature changes too, and the change spreads outward at the speed of light. This is a gravitational wave.
Einstein derived gravitational waves from his equations in , but for a long time afterwards it was disputed whether they were real at all.
Only motion that changes the arrangement gives off gravitational waves. A star that is symmetric about its axis gives none by simply spinning; it takes something like two stars orbiting each other.
The two stars return to the same arrangement every half orbit, so the gravitational wave repeats with half the orbital period.
Lay out marks floating in space in a ring, and while a gravitational wave passes the ring stretches and shrinks. When it is wide across it is short up and down, and a moment later the two directions swap.
What is gained in one direction is lost in the other, so the area of the ring does not change.
There are two ways for the stretching to be oriented. The one along the vertical and horizontal is called plus, the one turned degrees to stretch along the diagonals is called cross, and any gravitational wave can be written as a superposition of the two.
The strength of a gravitational wave is given by the fraction by which a length stretches and shrinks. Even for the wave first caught in , that fraction was about at its largest.
On an arm km long the stretch is m, about a four-hundredth of the m width of a proton.
The coefficient that ties matter to the curvature of spacetime in the Einstein equation is only about in SI units. However violently stars move, the tremor of spacetime that reaches far away is tiny, and Einstein himself thought gravitational waves would never be measured.
The stretch is measured with a laser interferometer. One beam is split in two, sent down two arms set at right angles, and brought back together on return.
While the lengths of the arms stay the same, the brightness of the recombined light does not change. A passing gravitational wave stretches one arm and shrinks the other, so the crests and troughs of the two returning beams slip against each other and the brightness changes.
An interferometer of the same form is what Michelson and Morley used in to look for a difference in the speed of light with direction.
To read a stretch of m, the mirrors are kg pieces of glass hung on fine threads of glass. The light makes some round trips in each arm, and its path is kept in a vacuum of about a trillionth of an atmosphere.
On that day the two interferometers of LIGO in the United States recorded the same shape of tremor, milliseconds apart. The two are about km apart, and noise inside the instruments would hardly take the same shape at both sites.
The millisecond gap is shorter than the milliseconds light takes to cover km, as it must be for a wave travelling at the speed of light.
The waveform began slowly, grew faster and larger toward the end, and died away abruptly where it was strongest. The whole recorded tremor lasted less than seconds.
billion light years away, black holes of and solar masses had merged. The black hole left behind had solar masses, and the difference of solar masses went out as the energy of gravitational waves. By that is J.
Just before the merger, the energy going out each second reached W, more than the light of all the stars in the visible universe put together.