杏吧原创

ASTRONOMERS can barely look up these days without detecting yet another crop of alien worlds circling distant stars. It鈥檚 a thrilling time to have eyes on the skies, and hard to recall that just over two decades ago, the only known planets in the universe were the nine 鈥� back in Pluto鈥檚 glory days 鈥� orbiting the sun.

Many people suspected that our solar system, with its numerous and varied planets, was one of a kind. And there was no evidence to suggest otherwise. Even if there were alien worlds out there, it was thought they would be too small and dark to be detectable, and lost in the glare of the stars they orbit. Anyone with the nerve to explicitly hunt for exoplanets risked not only their reputation but also their funding, so few astronomers tried. Even the man who would bag the first one thought it was far-fetched. 鈥淚t would have been considered a really silly topic for my thesis,鈥� says Didier Queloz, who was a PhD student at the University of Geneva at the time.

So how did Queloz, now primarily at the University of Cambridge, make such a landmark discovery? It began with his PhD advisor, Michel Mayor. Mayor was an authority on analysing the spectrum of light coming from stars. Two decades earlier, he had developed CORAVEL, a type of spectrograph that could detect the shifts in a star鈥檚 light caused by a nearby massive object. For example, in a binary star system, in which two stars orbit around their common centre of gravity, the gravitational pull of one star affects the radial velocity of its partner star, causing it to wobble slightly in its orbit. That wobble creates Doppler shifts in the spectrum of starlight that CORAVEL could detect. Mayor was involved with building a more sensitive instrument. 鈥淢ichel was looking for someone to take on this project, and I liked the style of the guy,鈥� Queloz says. 鈥淚 also really like building stuff. That was how it all started.鈥�

The anomaly

At first, the new instrument, ELODIE, could discern changes in radial velocities as low as 50 metres per second. 鈥淏ut I had built the software and I knew it could do much better,鈥� Queloz says. They improved the sensitivity to 10 metres per second 鈥� able to detect subtler stellar wobbles than ever before. They knew this level of sensitivity was a potential game-changer: it might allow for the detection of giant planets. So they decided to start looking.

Mayor and Queloz picked 100 stars that no one had paid much attention to and settled in for what they thought would be a decade of data gathering. 鈥淢ichel told me to start looking at the radial velocities, then he went on a sabbatical to Hawaii.鈥� That was in May 1994.

It was July when Queloz first looked at 51 Pegasi, a star 50 light years away. By September, he realised the data didn鈥檛 make sense. The star鈥檚 radial velocity kept changing in ways Queloz didn鈥檛 understand: the wobbles were too small for a binary star system, but the value was changing way too fast to be caused by a planet. 鈥淚 was really worried. For a week, I tried to find a bug in the software, but couldn鈥檛 find one,鈥� he says. Further observations continued to throw out inexplicable results. 鈥淚 was in total panic at that point,鈥� Queloz says. But he kept it all to himself: family, friends and Mayor had no idea of his turmoil.

Quietly, doggedly, Queloz kept up a programme of observations, looking at other stars to see whether ELODIE also did something inexplicable to their radial velocities. It didn鈥檛. 鈥淭hat鈥檚 when I started to think, 鈥極K, this must be real鈥�,鈥� Queloz says, but he still couldn鈥檛 bring himself to confide in Mayor. 鈥淚 was worried he would laugh at me.鈥�

By January 1995, Queloz in desperation reached for what he thought was a far-fetched solution. If he was seeing the gravitational effects of a giant planet, it would have to be orbiting disturbingly close to the star to make it wobble so rapidly. This went against three decades of space exploration programmes, billions of dollars of investment in probes sent to Jupiter, Saturn and the other planets. 鈥淭hey all said a Jupiter-scale planet couldn鈥檛 form and orbit so close to a star,鈥� Queloz says. For comparison, Jupiter, the closest gas giant to the sun, takes 12 years to complete an orbit.

He developed a computer program in an attempt to fit the data with a possible orbit, and finally cracked the puzzle in March. 鈥淚 got an observation, and it was right on the curve,鈥� he says. 鈥淎nd then I did another one, and that was on the curve, too.鈥� Queloz decided it must be a giant planet whizzing around 51 Pegasi once every 4.2 days, and it seemed to be orbiting more closely than Mercury orbits the sun. It was a conclusion bordering on scientific heresy.

A thrilled Queloz sent a fax to Mayor. It read: 鈥淚 think I鈥檝e found a Jupiter-mass planet of four days鈥� period.鈥� But Mayor鈥檚 reply was non-committal. 鈥淗e told me to keep it to myself until he got back,鈥� Queloz says. 鈥淗e said afterwards that initially he didn鈥檛 believe it.鈥�

Mayor returned from Hawaii in April, but by then 51 Pegasi wasn鈥檛 in the right part of the sky, so the pair had to wait until July to resume monitoring. 鈥淭hen, for four nights it was exactly on the curve. I said: 鈥榃e鈥檙e Swiss; let鈥檚 be sure.鈥� And on the fifth night, we said, 鈥楾hat鈥檚 it. Wow. It鈥檚 a planet. It really shouldn鈥檛 be, but it is鈥�,鈥� says Queloz.

鈥淭hat鈥檚 it. Wow. It鈥檚 a planet. It really shouldn鈥檛 be, but it is鈥�

Then came paranoia about being scooped. 鈥淲e kept quiet and drafted a paper,鈥� he says. At the end of August, they submitted it to Nature. Queloz remembers the terror: he knew a single mistake in the data would have crippled his career before it had even begun. 鈥淚t was my data, and as a PhD student with no reputation to fall back on, I鈥檇 be dead if we鈥檇 got it wrong.鈥�

In October, Nature still hadn鈥檛 accepted the paper, but regardless the pair presented their results at a conference in Florence, Italy. An audience member phoned astronomer Geoff Marcy at the University of California, Los Angeles, who had been looking for exoplanets. Marcy observed 51 Pegasi for five nights and was able to gather data to support the exoplanet鈥檚 existence. When that November, there were still doubters aplenty. Many astronomers later admitted to Queloz that they didn鈥檛 believe in his discovery for years afterwards. It wasn鈥檛 until 1999, when David Charbonneau, then at Harvard University, saw a star鈥檚 light dim as a planet crossed in front of it that the scientific tide turned. And it turned decisively.

Today, exoplanets are commonplace and include many 鈥渉ot Jupiters鈥� like the one Queloz spotted. The haul of confirmed worlds is more than 3500 and climbing 鈥� . Some, such as the TRAPPIST-1 system with its seven Earth-scale planets, have truly captured the popular imagination. 鈥淲e didn鈥檛 realise that the discovery of other worlds would mean so much to the public,鈥� says Queloz, who remains at the forefront of exoplanet discovery.

It turns out that our solar system is nothing special in planetary terms. But it is unique, at least, in boasting the only known life in the universe. But with hunters now scouring exoplanet atmospheres for signs of alien life, how long before we realise we鈥檙e not so special after all?

This article appeared in print under the headline 鈥淓xoplanet hunter number one鈥�