Doubt Cast on Exomoon Candidates

Exomoons, the hypothesized moons of exoplanets, have so far evaded detection. But our own solar system is awash with moons, the largest of which are around 40% of the Earth's diameter. Several are warmed by tidal heating and thought to harbor subsurface oceans beneath their icy crusts. So it seems reasonable to infer that moons are common in other stellar systems, too. That’s a tantalizing prospect, since some of these satellites might be, if not inhabited, then at least habitable.

However, astronomers have only found evidence hinting at a few exomoons so far — and a recent study casts doubt on two of those detections.

When exoplanets transit in front of their host stars, they block some of their light. By measuring these dips in the stellar light, astronomers can build a picture of the objects causing the partial eclipses. Most exoplanets have been found using this technique of transit photometry. Moons could be found this way, too, but because their dips might occur before, during, or after their planet’s, they require some additional statistical vetting.  

Six years ago, a team of astronomers announced the discovery of an exomoon candidate orbiting the Jupiter-size Kepler 1625b, 7,800 light-years away. Based on Kepler data, this discovery received support from a subsequent Hubble observation. Two years ago, another team with some of the same researchers identified another exomoon candidate orbiting a different giant world, Kepler 1708b, 5,600 light-years from Earth. Both times, the researchers urged caution in the interpretation of their results.

---

Read the September 2020 issue’s “The Hunt for the First Exomoons” to learn more about these (and other) initial detections.

---

Now, however, a study in Nature Astronomy casts doubt on those exomoon claims, offering an alternative interpretation of the data. Astrophysicist René Heller (Max Planck Institute for Solar System Research, Germany) and citizen scientist Michael Hippke (Sonneberg Observatory, Germany) re-analyzed previous observations, with the help of a new and improved code called Pandora that they developed to model exomoon transits.

For both exomoon candidates, they made 128 models that had only a planet transit the star as well as another 128 models that included both a planet and its moon. Both models were based on the known parameters of the exoplanets in question.

Comparing these models, they conclude that Kepler would only have been able to detect moons on wide orbits around exoplanets — specifically orbits beyond 30% of the Hill radius, the limiting distance within which a planet’s gravity is stronger than its star’s. Kepler was not designed to detect moons that orbit much more closely than that, like the large moons in our solar system do.

Yet the suspected exomoon of 1625b was orbiting its planet very closely in all the transits that Kepler caught.

The new study proposes that the original team might have overestimated the putative exomoon signal due to stellar limb darkening, the darkening of a star at its edges. This effect gives transiting objects the appearance of having different sizes, shapes, and depths — observed as the “black drop” effect during the transit of Venus — and thus makes it harder to separate a planet’s signal from any additional signals.

Heller and Hippke also found that the high confidence that previous astronomers had attributed to exomoon signals hinged on the method they had used to remove interference, such as that from passing clouds or a telescope’s vibrations. This detrending is challenging, since the light from distant stars has all kinds of blemishes in it anyway, caused by star spots and other random variations. In this case, the detrending might have accidentally injected moons into the light curve where none existed in reality.

“These moons are extremely challenging to detect, since their signature is very subtle,” comments Johanna Vos (Trinity College Dublin), who was not involved in the study. She highlights the importance of each exomoon candidate being “thoroughly investigated, preferably by multiple teams making use of independent techniques.” She also praises studies such as this one for developing “advanced tools for detecting and validating possible signatures of exomoons.”

Brian Jackson (Boise State University) also finds the study “important and impactful.” He adds: “The analysis seems rigorous and comprehensive.” The next step in the exomoon search is to gather more data, he says.

That data may need to come from new instrumentation that can detect exomoons transiting at smaller distances from their host stars. Such visual evidence would end the speculation surrounding exomoons, but Heller and Hippke do not believe it can be obtained in either Kepler’s archive or the upcoming PLATO exoplanet survey, scheduled to launch in 2026.

Jackson admits surprise at the lack of exomoon detections so far but notes, “If there's one thing astronomers have learned to expect, it's to expect surprise.”

For the time being, at any rate, this is where things stand: Despite living in a moon-filled solar system, we remain limited in our ability to detect moons elsewhere in the universe.