Searching for exomoons - moons the orbit around another planet - was one of the most exciting capabilities expected of the James Webb Space Telescope (JWST) when it launched in late 2021. So, after four years of operation, why hasn’t it found one yet? Turns out it’s really, really hard to find a moon around a planet light-years away. A new paper available in pre-print on arXiv from David Kipping of Columbia University (and Cool Worlds YouTube Channel fame) shows why. They used 60 hours of time on JWST’s NIRSpec instrument and weren't able to definitively confirm the existence of a possible exomoon.
That has not been for lack of trying. JWST has found two potential exomoon candidates, but neither has yet been confirmed. WASP-39b is a “hot Saturn” that showed a lot of variability in its sodium and sulfur dioxide levels, which led the researchers looking at it to posit that an extremely volcanic tidally heated exomoon is spewing gas onto the planet. While plausible, its definitely based on indirect evidence at best. Another finding about W1935, as brown dwarf, is similarly based on indirect evidence. In this case, it has some unexplained methane emissions that could be caused by an undiscovered exomoon - but yet again it has yet to be directly observed.
That brings us to Kepler-167e, the focus of the recent paper by Dr. Kipping and his co-authors. This planet seemed ideal to search for a moon, as it's one of the closest Jupiter analogues we’ve yet found. It’s about .91 times the mass of Jupiter, with an orbit of about 1.88 AU, which is in between Mars and Jupiter, around a star that is around 1,119 light years away in the constellation Cygnus. Given that Jupiter itself has upwards of 70 moons, including Ganymede, the largest moon in our solar system, it seems likely that an exoplanet so similar to it would have some as well. The system also holds three other “Super-Earth” exoplanets, which had to be accounted for as confounding factors in the light curves JWST was creating.
Dr, David Kipping, the lead author on the paper, describes how the search happened. Credit - Cool Worlds YouTube ChannelTo create those light curves, the authors received 60 hours of observational time on JWST - an absurd amount when considering how many demands on its time JWST has. These observational blocks were broken into six separate 10 hour segments, with a short break in between each. One thing the authors noted was that, over the ten hour observational window, the intensity of the light curves they were looking at would gradually decrease. They chalked it up to “detector effects”, but noted that it occurred on a similar time-scale to the effect an exomoon would have on the light of its planet, making it difficult to distinguish between an exomoon transit and a detector artifact.
Despite the struggles with the detector, the authors did their best to eliminate any bias it might have caused, and ran the data JWST collected through several “pipelines” as they’re called in astronomy jargon. One was custom built just for this observational run, while two more, known as ExoTiC-JEDI and katahdin, had been previously developed. Once the data was run through the pipelines, it was compared against one of four separate models, ranging in complexity between a simple quadratic fit to a Gaussian process with a Matérn-3/2 kernel - anyone interested in learning what that means can look forward to a lot of signal processing math.
Of the twelve different possible combinations of data-pipeline and model, seven actually showed a possible exomoon in the JWST data. But astronomy is never that easy - other explanations must be ruled out to ensure the signal is real. In this case, the authors noted that there was another explanation that could have caused the data to match so well with what’s called a syzygy-like event - where an exomoon, planet, and star are all aligned. By their reckoning, it looked like the moon, planet, and star were lined up, but in fact it was just the planet passing over a sunspot.
Fraser interviews Dr. Kipping about the search for exomoons.Kepler-167 is known to be relatively quiet - or “quiescent” in astronomy jargon. But, previous studies of it from Kepler and Spitzer indicated that it was possible that it creates spots that are large enough to account for the dip in lightcurve seen in the JWST data. Also, when the authors attempted to calculate the size of the moon they had found, they calculated that it would have to be 30% bigger than the models suggested would be possible in order to account for the dip in light.
Given those two factors, they concluded that a sunspot was the most likely explanation for their findings. While it might be hard to admit that, after spending so much time and effort to get all that observational time on the most sought-after telescope in the world, the authors decided they came up empty-handed. But that is how the scientific process goes - sometimes a negative result is what you get.
But that won’t stop them from trying again. They’re suggesting another observational campaign when Kepler-167e transits again in October 2027. That’s still a few observational cycles from now for JWST, so it’s unclear whether they’ll be granted the additional time. But even if they aren’t, there are plenty of other dedicated exomoon observational programs in the current and planned cycles over the next few years. If these elusive worlds do exist, it's likely only a matter of time before we find one - and this new paper provides a roadmap to the kind of intense data processing, model fitting, and in some cases soul searching, required to make sure that, when we do, we know its real.
Learn More:
D. Kipping et al - A JWST Transit of a Jupiter Analog: II. A Search for Exomoons
UT - Tentative Exomoon Signal in HD 206893 B
UT - The Search for Exomoons is On
UT - These are the Best Places to Search for Habitable Exomoons