Artist’s impression of the exomoon candidate Kepler-1625b-i, the planet it is orbiting and the star. (NASA/ESA/L. Hustak, STScI)

“Moons are where planets were in the 1990s,” predicted René Heller from the Max Planck Institute for Solar System Research a few years ago. “We’re on the brink.”

Heller was predicting that we were close to the first discoveries of exomoons: moons that orbit extrasolar planets outside our solar system. When a possible exomoon detection was announced in 2017, Heller’s prediction was proved correct. Not only had we found a candidate moon, but its properties defied our formation theories just as with the discoveries of the first exoplanets.

However, a paper published in Science this month has proposed a method for building this most unusual of moons.

As we move away from the sun, the planets of our solar system become mobbed with moons. How these small worlds formed is attributed to three different processes:

Moons in our solar system are thought to have formed through three different mechanisms (E. Tasker / Many Worlds)

The most extensive moon real estate orbits our gas giants, Jupiter, Saturn, Uranus and Neptune. The majority of these moons are thought to have been born during the planets’ own formation, forming in disks of gas, dust and ice that circled the young worlds. These circumplanetary disks are like miniaturised versions of the protoplanetary disks that circle young stars and give rise to planets.

One exception to this is Neptune’s moon, Triton, which orbits in the opposite direction to the planet’s rotation. This retrograde path would not be expected to arise if Triton has formed out of a circumplanetary disk around Neptune, which always rotate the same direction as the forming planet. Instead, Triton was likely a dwarf planet that was snagged by Neptune’s gravity during a chance encounter.

The capture scenario has also been proposed for the two moons of Mars. The lumpy satellites resemble asteroids and may have been born in the asteroid belt that sits between Mars and Jupiter. However, both moons orbit the red planet in circular orbits that sit in the same plane, pointing to a more disk-like formation method. Although Mars is too small to have had a substantial circumplanetary disk during formation, a giant impact later in its history could have thrown debris into orbit. This debris disk could then have coalesced into the two moons.

Such a violent start to Mars’s moons would mimic the beginnings of our own moon. The Earth’s moon is believed to have formed when a Mars-sized protoplanet collided with the young Earth and ejected enough debris from the Earth’s mantle to form our large natural satellite.

But while these three formation mechanism successfully explain moons from the tiny Aegaeon to the large Ganymede, none of them quite fitted the moon found orbiting Kepler-1624b.

This animation represents the Hubble data and their best-fit light curve for the transit of exoplanet Kepler-1625b and its possible moon. (NASA GSFC / J. Koynock

Evidence for the exomoon was announced in 2017 and 2018 in two papers led by Alex Teachey and David Kipping at Columbia University. Small variations in the orbital period of Kepler-1624b suggested the presence of an unseen body that was tugging on the planet. This was accompanied by a double dip in light as the planet passed in front of its star, indicating a near-by second body was also blocking out light. It was the best indication yet of a moon outside our solar system.

Then estimates of the size of the exomoon brought everyone up short. Denoted Kepler-1625b-i, the moon is on a long orbit about the planet and has a measured mass similar to that of Neptune. Such a large moon defies ready formation by any of the above three moon-y recipes.

Models of moon formation in circumplanetary discs struggle to build such a large moon, even allowing for the sizeable bulk of Kepler-1625b which is a super-sized Jupiter. A giant impact that ejected a Neptune-worth of debris would have to involve two gas giants. This seems unlikely to have been efficient at throwing out rocky material given their thick atmospheres.

Even Triton’s capture is difficult to envisage due to fly-bys between planets usually occurring at speeds too fast for capture. In the case of Triton, the most popular scenario is the moon was once part of a binary of dwarf planets, similar to Pluto and its giant moon, Charon. The forwards-and-backwards motion as the binary circled one another allowed one of the pair to be moving slowly enough to be captured by Neptune, whose gravity ripped the binary apart and allowed Triton’s sibling to escape into space. But for such a mechanism to work for Kepler-1625b-i, the moon must have been in a binary with another planet of similar mass. No binary of such large planets has even been seen.

Pluto and its giant moon, Charon, are sometimes described as a binary planet system. Triton might have been in a similar configuration before it was captured by Neptune. (image taken by New Horizons, NASA / JHUAPL / SWRI)

This left us with a possible moon that seemed too big to be real. That was until the paper this month authored by Bradley Hansen from the Mani Bhaumik Institute for Theoretical Physics at UCLA. Hansen proposes an alternative moon recipe which begins with Kepler-1625b forming close to a number of other young planets.

While we normally think of planet orbits as fixed, this is not true in their early days when the new worlds are forming in the protoplanetary disk. The gaseous disk tugs on the growing worlds, dragging them into new orbits that are often (but not always) closer to the star. Such migration can be halted by sudden changes in disk conditions, such as a sharp rise in density or temperature change. One such example is the ice line, beyond which is it cold enough to freeze water vapour into ice. This produces a sharp boost in the density of solid material as ice joins the dust particles and alters the drag forces on a planet, potentially stopping it in its migrating tracks.

These planetary stop signs are known as planet traps and can cause a pile-up where several young planets end up in close orbits.

The planets grow by gathering dust and gas. This accretion increases with mass, due to the boost in the planet’s gravity. The result is that the first planet to grow a little larger than its neighbours will begin to grow more rapidly, stealing away building materials from the other planets. This will leave one larger planet surrounded by one or more smaller planets that were unable to gather enough material to reach the same size.

During this growth spirt, the gravitational pull of the larger planet on its neighbour will change. The increasing tug on the orbit of the smaller planet will gradually draw it closer, allowing an easier capture to form a moon than during a single fly-by.

Jupiter’s humungous atmosphere is thought to surround a solid core, but this has never been seen (NASA).

This technique is known as a pull-down capture. It was first proposed in the late 1970s as a way to capture small moons around Jupiter, but had not been previously shown to work for planets of similar size. In simulations of pull-down capture, a growing Kepler-1625b was able to drag on a neighbouring planet with the same mass as Kepler-1625b-i until the second world was unable to escape the larger planet’s gravity. It then began to orbit as a giant moon.

By exploring simulations with different mass planets, Hansen also found that the smaller planet could sometimes be flung outwards, rather than be captured. This could have happened in our own solar system, with the ejected planets becoming Neptune and Uranus whose large bulk is difficult to form in their current far-away orbits.

If Kepler-1625b-i is an exomoon formed through pull-down capture, then other giant moons may yet be discovered. This class of moon would be a failed gas giant, with a body similar to whatever resides in the core of Jupiter. Observing these giant exomoons would therefore be like looking at a naked gas giant, giving us a glimpse we never thought possible at what might reside under the throttling, turbulent mass of a gas giants’ huge atmospheres.