Somewhat more than 4000 planets orbiting stars outside our solar system are known so far. Of these so-called exoplanets, about 96 percent are significantly larger than our Earth, most of them more comparable with the dimensions of the gas giants Neptune or Jupiter. This percentage likely does not reflect the real conditions in space, however, since small planets are much harder to track down than big ones. Moreover, small worlds are fascinating targets in the search for Earth-like, potentially habitable planets outside the solar system.

The 18 newly discovered worlds fall into the category of Earth-sized planets. The smallest of them is only 69 percent of the size of the Earth; the largest is barely more than twice the Earth's radius. And they have another thing in common: all 18 planets could not be detected in the data from the Kepler Space Telescope so far. Common search algorithms were not sensitive enough.

In their search for distant worlds, scientists often use the so-called transit method to look for stars with periodically recurring drops in brightness. If a star happens to have a planet whose orbital plane is aligned with the line of sight from Earth, the planet occults a small fraction of the stellar light as it passes in front of the star once per orbit.

"Standard search algorithms attempt to identify sudden drops in brightness," explains Dr. René Heller from MPS, first author of the current publications. "In reality, however, a stellar disk appears slightly darker at the edge than in the center. When a planet moves in front of a star, it therefore initially blocks less starlight than at the mid-time of the transit. The maximum dimming of the star occurs in the center of the transit just before the star becomes gradually brighter again," he explains.

Large planets tend to produce deep and clear brightness variations of their host stars so that the subtle center-to-limb brightness variation on the star hardly plays a role in their discovery. Small planets, however, present scientists with immense challenges. Their effect on the stellar brightness is so small that it is extremely hard to distinguish from the natural brightness fluctuations of the star and from the noise that necessarily comes with any kind of observation. René Heller's team has now been able to show that the sensitivity of the transit method can be significantly improved, if a more realistic light curve is assumed in the search algorithm.

"Our new algorithm helps to draw a more realistic picture of the exoplanet population in space," summarizes Michael Hippke of Sonneberg Observatory. "This method constitutes a significant step forward, especially in the search for Earth-like planets."

The researchers used data from NASA’s Kepler space telescope as a test bed for their new algorithm. In the first mission phase from 2009 to 2013, Kepler recorded the light curves of more than 100,000 stars, resulting in the discovery of over 2300 planets. After a technical defect, the telescope had to be used in an alternative observing mode, called the K2 mission, but it nevertheless monitored more than another 100,000 stars by the end of the mission in 2018. As a first test sample for their new algorithm, the researchers decided to re-analyze all 517 stars from K2 that were already known to host at least one transiting planet.