A fourth giant planet uncovered in the HR 8799 system poses problems for models of planet formation – neither established theory can explain the in situ formation of all four planets.

Two years ago, three planets were directly imaged at infrared wavelengths using the Keck and Gemini telescopes in Hawaii orbiting the star HR 8799, a 1.5 solar mass main sequence star located 129 light years away in the constellation of Pegasus. The system also possesses a dusty disc much like our own Kuiper Belt, with the giant planets orbiting within the disc's inner edge.

Now, additional analysis reveal a fourth planet orbiting just 14.5 AU (astronomical unit, equivalent to the Earth-Sun separation) from the central star, with a mass around ten Jupiter masses. Its companion planets orbit at distances of 24, 38 and 68 AU and are 10, 10 and 7 Jupiter-masses respectively.

The system is currently the only known example of a wide (greater than 25 AU) system with multiple giant planets, but this presents a problem for theories of planet formation, which occurs by one of two processes. In the core accretion "bottom-up" model, a rocky core builds up through collisions of smaller rocky bodies, until its gravity is great enough to trap a gassy atmosphere. In the gravitational collapse "top-down" model, instabilities in the star's primordial disc leads to fragmentation, i.e. clumps of matter condense out and attract further gas and eventually dusty debris that sinks to the core.

In the case of HR 8799, the core accretion model would allow the innermost planet, HR 8799e, to form within 10 million years before the star's gas disc had dissipated, but HR 8799c would need 20 times longer to grow to its mass at 38 AU, by which point the disc would have lost its mass, and at 68 AU core accretion would exceed the age of the star, an impossible situation. On the other hand, the gravitational instability model cannot explain the formation of the innermost planet, because the star's primordial disc would not have been cold enough or rotating slowly enough to fragment and collapse to form the planet in situ.

"It is not obvious that you can form planets by both scenarios in the same system," lead author of the study Christian Marois of Canada's NRC Herzberg Institute of Astrophysics tells Astronomy Now. "People usually study one or the other, not both at the same time. Given the similar masses and dynamic, we believe it is unlikely that both scenarios occur in this system."

Marois continues by explaining that migration of the planets from an initially different location, could provide a solution. "All four planets could have formed at wider separations and then migrated inward, or formed closer-in and then migrated outward. But this complicates the story and it is not obvious that you can move such massive planets over tens of AU and not eject a planet or two in the process, or find stable orbits for all four planets that last tens of millions of years."

Preliminary calculations suggest that the planets c, d and e are in a 1:2:4 orbital resonance, which means for every one orbit that planet c makes, d makes two and e four, much like the situation with Jupiter's moons, Ganymede, Europa and Io.

"Having a resonance is possibly a sign that the planets have moved, but it doesn't tell you by how much they moved," says Marois. "It is possible the planets form at their current location (or close to their current position) and the resonances is part of their formation process. But this problem requires more research and simulations. We are really just at the beginning of this field – we need more planets and more systems like HR 8799 to learn more about how planets form around stars, but it is clear that HR 8799 is already offering a nice challenge. It shows us how little we know about planet formation and how much diversity is out there."

Marois adds that it is possible that smaller rocky planets, currently below the limits of detection, could be located in the inner part of the system. "This would not change the general picture of planet formation, but would make HR 8799 a really close match to our own system – just younger, bigger and with more massive gas giant planets."