Matthew A. Kenworthy’s Research

A paper by Peters-Limbach and Turner in 2013 titled “On the Direct Imaging of Tidally Heated Exomoons” (THEMs) suggested that, for nearby exoplanet systems with multiple moons in resonance, you can have tidally heated exomoons (THEMs) that can be directly imaged at thermal infrared wavelengths - think Io, the tidally heated moon of Jupiter with its sulphur volcanoes.

Building on the original paper, Elina Kleisioti wrote her PhD thesis on investigating the tidal heating properties of exomoons using more realistic models of the internal structure and heat flow within them.

In order to see the thermal emission from a THEM, the exoplanet system has to be very close to the Solar system, and you also need to make sure that the gas giant exoplanet that the exomoon is orbiting around has cooled off: it’s entirely possible that for a gigayear old system, the exoplanet can be fainter than the tidally heated exomoon at some wavelengths.

Since multiple moons are present around all the gas giant exoplanets in our system, it seems reasonable that there are exomoons elsewhere in the Universe.

We then chose a suitably nearby exoplanet that was orbiting far enough from its parent star that we had a chance of separating the light from the exoplanet/moon system, and ϵ Eridani fits the bill quite nicely. Elina carried out a study for what a THEM around this exoplanet would look like (Kleisioti et al. 2023).

Elina worked together with then Masters studen Quirijn van Woerkom to look at the spectroastrometric detectability of nearby Solar System-like exomoons (Woerkom & Kleisioti 2024). By using a spectrograph they showed that an additional astrometric shift should be possible to detect.

THEMs have to be close to their parent exoplanet in order to have strong tidal heating, and there is a non-negligible probability that the exomoon might transit the exoplanet, providing distinctive eclipses that could be seen by photometric monitoring of the exoplanet…. but then Elina pointed out that we don’t need to have the exomoon transit at all! Instead, if there is an active volcano on the tidally locked side of the exomoon facing the exoplanet, then the volcano appearing and disappearing on modestly inclined orbits provides an additional detection opportunity for exomoons (Kleisioti et al. 2024).

We took 11 epochs of radial veloctiy measurements of the planet Beta Pic b and performed a periodogram analysis on the resulting time series to provide the best upper limits on the presence of ecomoons around the planet to date. Together with astrometric limits from our coauthors, these rule out anything larger than two Saturns, a good start for finding exomoons around gas giant exomoons (Kenworthy et al. 2026).

J1407 is a nearby young star that underwent a complex series of eclipses that lasted for almost two months in 2007. Our best explanation is that an (as yet) unseen massive substellar companion with a giant ring system passed in front of the star, causing the eclipse (Kenworthy et al. 2015). We haven’t seen another eclipse towards J1407 (or its full name, 1SWASP J140747.93-394542.6), but data from the ALMA telescope array has detected what might be the dust making up the rings themselves.

We now give a brief history of what we’ve done since the discovery was announced in 2012.

Mark Pecaut, a then graduate student under the supervision of Eric Mamajek at the University of Rochester, discovered the unusual light curve towards the end of 2011. Mark was looking at the light curves of young stars as part of his thesis research. It is difficult to determine the ages of most stars, so several different pieces of evidence are needed to confirm a young star. Looking through the public light curve database of the SuperWASP collaboration, he found 1SWASP J140747.93-394542.6, a star almost the same mass as the Sun, with a light curve that showed a 3% variability that repeated every 3.2 days, almost certianly due to fainter star spots on the surface of the star rotating in and out of view as the star rotated.

J1407 showed a very long lasting eclipse event from April through May 2007, where the star’s changed brightness on a nightly basis, sometimes as much as 95% of the flux was blocked by something moving in front of the star. The discovery paper (Mamajek et al. 2012) contained a simple model including four large rings about 0.6 au in diameter surrounding an unseen substellar companion, called J1407b, which orbited around the parent star.

Follow up observations were carried out to try and detect J1407b, which must be several tens of Jupiters in mass in order to keep such a large ring system held together and not ripped apart by the gravity of the star. These included looking for another eclipse in other photometric data (no other eclipses were found), looking for the reflex motion of the star as the gravity of the unseen companion tugged on it (the star is young and is very active, preventing the method from being sensitive enough), and directly detecting the thermal glow of the companion itself (no - even with the largest telescopes and an optical trick called sparse aperture masking we did not see anything).

All these non-detections were still useful, since they allowed us to rule out possible orbits and masses for J1407b, detailed in Kenworthy et al. (2015).

One interesting conclusion we had was that there were no circular orbits for J1407b since we needed to explain why the ring system seemed to be either very big or moving very fast in front of the star, implying that it was in an orbit close to the star. This presented one big problem: our ring system seemed to now be very large indeed, so large that the rings would spill away from J1407b and fall towards the star. Putting J1407b on an elliptical orbit seemed to make the problem even worse.

This inspired Steven Reider to simulate a ring system around a planet that itself was in an elliptical orbit around the star J1407 (Rieder et al. 2016). The new trick was one of a simple spin - if the rings were added so that they span around J1407b in the same direction that the planet orbited around the star (called prograde motion) then the rings fell apart within a few orbits when the planet passed close to the star.

However, if the rings were spinning in the opposite direction (a retrograde motion) then the rings held together for a much longer time, so much so that we could find possible masses for J1407b and orbital periods that would work!

Even if these orbits worked, then another problem presented itself, in that the rings were tilted with respect to the orbital plane of the planet. Most planets rotate, and the centripetal force generated by this rotation causes the planets to bulge out along their equators, distorting them from ideal spheres into oblate spheroids (M&M or Smarties if you like candy/sweeties). This deviation from an ideal sphere means that any ring system orbiting around the planet will tend to align with the plane of the equator, as is the case with Saturn. Since J1407b is younger and more massive than Jupiter, the bulge around the equator is proportionally much bigger, marshalling the rings into this plane.

The rings of J1407b are about 200 times larger than Saturn’s rings, and for the rings a long way away from the planet, the tidal forces of the star J1407 will force them down into the orbital plane of J1407b, some 20 degrees away. The rings should no longer be flat, but have a large systematic warp to them, as calculated in Zanazzi et al. (2017).

Photographs of the night sky have been taken for over a century, and one of the largest collections is at Harvard University. These photographic plates are being scanned in digitally and published onto the internet through the DASCH project. The process is slow and time consuming, especially as you work towards the Galactic plane, where there are far more stars crowded together and disentangling the signals is challenging, but they reached the location in the Galaxy where J1407 is located around 2017. This enabled the baseline for the J1407 lightcurve to be extended back to the 1900’s, but still no eclipses were detected. The coverage of the photographic plates from DASCH were combined with others from Sonneborg and other more recent all sky surveys that looked towards J1407, but still a null result (Mentel et al. 2018).

Meanwhile, the star is being observed every clear night by the dedicated observers of the AAVSO. They watch J1407 for when the next eclipse will start. The constraints they provide rule out successively more obital periods, leaving J1407b with fewer orbital periods to hide in.

If the planet J1407b could not be detected, then how about looking for the rings directly?

The rings are probably made of small rock or icy particles, ranging from one micron in size up to rocks and boulders. The surface area of the dust in the rings is vast, so much so that they radiate thermal energy at millimetere wavelengths that can be detected with radio telescopes on the Earth. The most sensitive sub-millimetre telescope is the Atacama Large Millimetre Array (ALMA) in the Atacama desert in Chile. We were able to get observing time on this array of telescopes in the Summer of 2018, and to our surprise we detected a sub-mm source with the brightness we predicted for J1407b’s rings (Kenworthy et al. 2020).

After our initial joy, we then became confused again - because the ALMA source wasn’t exactly in the same position as the location of the star. At these very long wavelengths, the star J1407 is not visible in ALMA, so we had to calculated the expected location of the star and compare it to the ALMA source. If the source is the same physical object that caused the eclipse in 2007, we run into a big problem - the source cannot be in a gravitaionally bound orbit around J1407…. but there’s an very unlikely alternative explanation - that the object we called J1407b is floating freely through the Galaxy, and just happened to pass in front of a very young star.

This is very, very, very unlikely to happen, but there’s a first time for everything…. and to confirm this tantalizing idea, we needed another image from the ALMA array. And in 2023, we got our wish granted! Thanks to a successful ALMA application and a suggestion by the ALLEGRO team, we obtained a second ALMA image looking towards J1407 in 2024. Unfortunately, we didn’t see the ALMA source again, despite having the sensitivity to do so (Klaassen et al. 2025).

This does not mean that the rings aren’t there though - looking for the rings in ALMA was always a long shot, because we’d only be able to detect the most massive (and flared) ring models due to the distance of the system. The rings are still there, just not detectable with these ALMA observations. Our best guess for the 2018 detection is that it was a noise spike or some distant transient source. That’s pretty a unsatisfying explanation, but without more data we’re going to leave it at that.

So that is the latest and greatest information about J1407 and the ring system around J1407b. We still think that the eclipse in 2007 was caused by something orbiting J1407, and that one day we will see another eclipse that will confirm the incomplete light curve from 2007. Once we see it, all the geometry of the ring system is solved almost immediately. Until then, we tide ourselves over and patiently wait for a triggering signal from an observer who says “This star has started fading again!”

I’ve been looking at the light curves that the ASAS-SN survey releases on their transients page - these are light curves that show curious dimmings caused by stellar variability or by something moving in front of the star. One of these light curves, named ASASSN-21js, showed a sudden dimming in 2021, and over the past three years, has faded and then brightened up again. Theo worked on analyzing this light curve with me as a Masters thesis project here in Leiden: after some experimentation, the light curve is exactly matched by a highly tilted and inclined disk with two rings with different transmissions. What’s particularly impressive is that the residuals after the model fit are consistent with noise. This model explains the data almost perfectly.

Using a MCMC fitting package, he was able to determine the confidence intervals on the parameters, with the outermost ring about 1.12 au in radius. Due to the geometry of the eclipse, we only see these two rings - there may be more material interior to this radius, but the star’s motion will not reveal this. What’s more amazing is that the eclipse is still ongoing, finishing only in May 2027!

Okay, so what are the challenges here? First, the star itself is very large star with a mass of 4.5 times that of the Sun. Those are relatively rare, but are seen long distances across the Galaxy. Second, the transverse velocity is less than one kilometre per second! If this is orbiting around this star, then it’s at a distance of over 13000 au from the star! Deep into Oort Cloud territory. This means that the chance of seeing such an eclipse is tiny, around one in ten thousand for all the solar type stars in the ASAS-SN survey size and duration. That’s an uncomfortable problem and we don’t know the answer as to why we’re seeing this now.

What’s tantalizing is that the MCMC fitting also found solutions with a cleared ring gap between the two main rings (squint at around 59200 days to see this additional gap in the rings). It’s compelling, but not confirmed. There is also multi band data coming in as well, so in future research we can see if the material is sub-micron sized or not. More of these unusual systems are being discovered in ASAS-SN, and with other large survey telescopes coming online soon (think the Vera C. Rubin observatory) many more of these systems might be discovered.

Looking through the first set of second epoch images, Alex discovered our first exoplanet companion, YSES 1b, a 14 Jupiter mass companion around a 1.0 Solar mass star (Bohn et al. 2020), and then whilst taking data to characterise YSES 1b, Alex discovered a more distant companion from YSES 1, and this is a 6 Jupiter mass exoplanet some 320au away from the star (Bohn et al. 2020b).

The YSES survey announced the discovery of a second directly imaged exoplanet system (Bohn et al. 2021), this time with a 6 Jupiter mass exoplanet with a projected orbit of 115au. What’s interesting is that this was the second discovery when only a third of the stars have been observed for their second epoch observations, suggesting that we may well find more exoplanet systems in the rest of the survey.

2025 update: HOWEVER… sometimes the universe throws you a curveball - after working on characterising the orbit of YSES 2b, we discovered that it is not an exoplanet, but a distant M dwarf star some two kiloparsecs behind YSES 2. The background star has proper motion that looks like Keplerian orbital motion, and only several epochs of observations with the VLT interferometer GRAVITY were able to spot the parallax in the background star.

We break the bad news in Kenworthy et al. (2025).

One of the first discoveries from YSES was a circumstellar disk around the star Wray 15-788 (Bohn et al. 2019), showing detailed circumstellar structure and shadowing from the inner rings onto the outer rings. There are at least two gaps in the disk, and the inner disk is tilted with respect to the outer parts of the disk, such that the outer disks are not completely illuminated by the light of the star. There may well be one or more planets in the gaps in the disk, but these are always challenging to image directly. Deeper observations will reveal the detail and fainter structures, if they are present in the outer regions of the disk. Future observations will constrain the nature of the ring material and the geometry of inner structures closer to the star.

WISPIT 1 is a stellar binary with a primary star of around one solar mass, with two comoving companions identified as gas giant exoplanets (van Capelleveen et al. 2025a). These companions have photometric masses of 10 and 5 Jupiter masses, orbiting at projected distances of 338 and 840 au respectively, presenting a challenge for formation mechanisms as to how these planets ended up so far from their parent stars. Maybe something to do with the stellar binary? Or some other process that’s switched on in this particular system?

Future observations will narrow down their origin, and identify what molecules there are in the system. It is interesting to note that this is another multi-planet system.

We detected a spectacular circumstellar disk which has several cleared rings within it. The widest ring has a protoplanet of about 5 Jupiter masses (van Capelleveen et al. 2025b), and is detected as a point source in the optical Hydrogen alpha line indicating that is is undergoing acitve accretion within this cavity (Close et al. 2025).

The two papers detail the orbital dynamics of the planet, which is at about 56 au from the central star with a photometric mass of 5 Jupiter masses. Orbital motion is seen over two years of observations, confirming it is comoving with the star, and the astrometry is consistent with Keplerian orbital motion within the dust gap opened up by the planet. Imaging from Hydrogen alpha in the optical through to L’ band at 3.6 microns confirm that the planet is accreting gas, and has a thermal spectral energy distribution with a young planet consistent with the age of the disk and star.

Follow up observations with the GRAVITY+ interferometer confirms that there is a second planet, WISPIT 2c, at around 16 au from the star with a mass of around 10 Juptiers (Lawlor et al. 2026), and to add even more intrigue, the star has been resolved to be a spectroscopic binary with a lower mass secondary on a 4 day orbit around the one solar mass primary star (Burgy et al. 2026).

Since the star is close to the celestial equator of the Earth, it is visible from all the large observatories. The width of the gap is an important test of planet formation theories.