For more than three years, the MEarth-South observatory, a cluster of telescopes in Chile, has stared at distant stars in the night sky, waiting for a hint of dimming in their brightness, a sign that something—probably a planet—is passing by. In the fall of 2014, the observatory detected a dip in starlight coming from a red dwarf about 39 light-years away from Earth. Unfortunately, no one noticed. The signal, thought to be an imperfection in the data, was tossed out.
A year later, Jason Dittmann, one of the scientists who worked with MEarth, was experimenting with a new machine-learning algorithm on his own time. Excited to test it out, Dittmann threw the data from the observatory at the algorithm, told it what to look for, and voila. There it was, a tiny signal from a potential new world.
“It’s just one of those scenarios where I had a shiny new hammer and so everything looked like a nail,” Dittmann said.
And he hit the nail on the head. The source of the signal turned out to be a “super-Earth,” an exoplanet slightly larger than ours but with a mass seven times greater, orbiting in the habitable zone of LHS 1140, a red dwarf star in the constellation of Cetus. The findings, released Thursday, are published in the journal Nature.
The planet, known as LHS 1140b, is a prime target in the search for life in the universe. It orbits 10 times closer to its star than Earth does to the sun, but red dwarfs are cooler and dimmer than stars like the one in the center of our solar system. Its distance from the red dwarf suggests it could harbor liquid water on its surface, and its density suggests it’s made of iron and rocky silicate materials, like Earth. The planet is probably tidally locked, with one side perpetually facing its star. Scientists don’t know whether 1140b has an atmosphere, but future technology, like the James Webb Space Telescope that launches next year, could go and look. Dittmann said the Webb team has already committed some time to studying the object once the telescope is ready.
Dittmann used an artificial neural network to scan the data of 300 stars. An algorithm then classified the data as either noise coming from the atmosphere, like wind, or a “real” transit. Dittmann said he had to convince his professors at the Harvard-Smithsonian Center for Astrophysics to consider a closer look at the signal his algorithm found in the data. Machine learning is usually used for things like making Netflix learn which movies you might like based on what you’ve watched or teaching Google to find the faces in your photo collections, not planet-hunting. But the target looked promising, and they decided to go for it.
Dittmann and his team turned to a powerful, exoplanet-hunting instrument at the European South Observatory’s La Silla facility in Chile called the HARPS instrument. HARPS detects wobbles in the motion of stars, looking for a gravitational tug-of-war between a star and a potential planet. Using HARPS and more detections from MEarth, the scientists were able to estimate the mystery object’s orbital period, mass, and density. They predicted that the object’s next transit would be the first day of September in 2016.
Dittmann remembers the “frantic night” they finally confirmed the planet’s existence. Six telescopes were trained on the red dwarf that night. But the telescopes located on Australia’s eastern coast were “weathered out,” which means their view was obscured by clouds and other weather patterns, and the ones in Hawaii were riding out back-to-back hurricanes. The sole telescope that detected the transit that night was run by an amateur astronomer in Australia, at an observatory set up in his backyard.
The discovery of LHS 1140b comes just months after another big exoplanet find. In February, astronomers announced the discovery of seven Earth-sized rocky planets orbiting TRAPPIST-1, a red dwarf located about 40 light-years from Earth. The planets orbit in the system’s habitable zone, where temperatures could allow liquid water to pool on their surfaces.
For Dittman, LHS 1140b also helped him find his next venture. Dittmann will soon start a new job at the Massachusetts Institute of Technology, where he’ll oversee his own research program and work on NASA’s Transiting Exoplanet Survey Satellite (TESS), which will launch next year to look for transits in more than 200,000 stars. The gig is a relief. When he graduated from Harvard last summer with a Ph.D. in astronomy and astrophysics, Dittmann didn’t have any luck getting a full-time job, he said. This year, however, he added his research on LHS 1140b to his applications, which may have helped. How many people can say they discovered a distant planet on their resumé?