I have been working on the formation of large satellites (moons) around gas planets, Jupiter and Saturn. The satellites formed in circumplanetary gas disks (CPDs), byproducts of the gas accretion of gas planets during their formation process. Solid materials supplied from the parental protoplanetary gas disks (PPDs) grow large to satellites in CPDs like the planet formation in PPDs. There are mainly two formation scenarios; satellitesimal accretion and pebble accretion. In satellitesimal accretion scenarios, satellites form by collisional growth of satellitesimals (km-sized). The satellitesimals form by coagulation of dust particles (mm-sized) in CPDs or planetesimals captured by CPDs are directly used as the satellitesimals. In pebble accretion scenarios, pebbles (cm-m sized) drifting toward the central gas planets in CPDs are accreted to the seeds of satellites (100 km sized). The captured planetesimals with relatively large sizes could be used as the seeds of the satellites. Both scenarios have advantages and disadvantages.
Our review of the formation of large satellites: Shibaike & Alibert (2025), 芝池 (2019) (Japanese), and Blanc, ..., Shibaike et al. (2025)
Image of the Galilean satellites and Jupiter (NASA)
The question of how satellitesimals form from dust particles has not been addressed so far. We calculated the evolution of dust particles grow by their mutual collisions in CPDs and drift toward the central planets owing to the headwind from the gas disks with continuous supply of gas and dust from the parental PPDs. As a result, we found that the conditions for satellitesimal formation is that the dust-to-gas mass ratio of the inflow is larger than unity. This condition is difficult to fulfill so that an alternative way to form satellitesimals is necessary to be considered (Shibaike et al. 2017). According to our non-ideal MHD simulations, however, CPDs could be magnetic wind-driven accretion disks. In that case, the disks are laminar, and the settled down dust particles on the mid-plane can grow efficiently, resulting in the satellitesimal formation with a much lower dust-to-gas mass ratio of the inflow (Shibaike & Mori 2023).
We proposed a new scenario for the origin of the Jovian (Galilean) satellites. In our scenario, two kinds of solid materials are supplied to the CPD of Jupiter; dust particles and planetesimals. The captured planetesimals accrete drifting pebbles formed by the small particles thanks to the strong gravitational focusing and aerodynamic drag so that the planetesimals can grow large to the satellites. Our scenario can reproduce most of the physical, dynamical, compositional, and structural characteristics of the Galilean satellites (Shibaike et al. 2019).
I also robustly demonstrated that Callisto's partial differentiation is not achievable through satellitesimal accretion, which inevitably leads to significant differentiation, but can be maintained with pebble accretion. Pebbles can release their impact energy at the surface of the satellite, allowing efficient radiative cooling, and their impact velocities can be reduced by aerodynamic drag from the CPD. If JUpiter ICy moons Explorer (JUICE) confirms that Callisto is only partially differentiated, it could provide an indirect but the first observed evidence for the pebble accretion mechanism—not only in the context of satellite formation but also in the broader framework of planet formation! (Shibaike 2025).
Large satellites form in CPDs as explained above. Recently, the material of the satellites, the dust in CPDs, have been detected by (sub)millimeter continuum observations by Atacama Large Millimeter/submillimeter Array (ALMA), a state-of-the-art radio telescope. We built a dust emission model considering the dust evolution in CPDs by updating the model developed for the satellite formation and predicted the intensity of the dust thermal emission from the disks. We applied the model to the (potential) CPDs of the two confirmed gas accreting planets, PDS 70 b and c, and succeeded in obtaining constraints on the planet mass and the gas accretion rates (Shibaike & Mordasini 2024). We also discussed the detectability of the CPD by ngVLA (Shibaike et al. 2025b)
We have also applied the model to the planet candidate AB Aur b, and found that emission from its potential CPD should already have been detected by ALMA, although it has not been observed yet (Shibaike et al. 2025a).
Our review about dust evolution and emission in CPDs : 芝池 (2024) (Japanese)
Submillimeter continuum emission from PDS 70 observed by ALMA. The small dot inside the bright outer ring is considered as the CPD of PDS 70 c.
(ALMA (ESO/NAOJ/NRAO)/Benisty et al.)
I have also been working on the planetesimal formation. We proposed a new scenario in which planetesimals can form in broad areas of PPDs. Pebbles drifted from the outer region of a PPD pile up at the gas pressure bump created by a planet and form planetesimals. The formation region then spreads inward in the disk as the planet migrates (Shibaike & Alibert 2020). We found that planetesimals form in a belt like region in the PPD, if the planet forms while enough amount of pebbles left at the outer region (Shibaike & Alibert 2023). We also found that the profile of the planetesimal surface density and its slope can be estimated by a simple expression .
I also studied the effects of the Late Heavy Bombardment, the concentration of impacts on the early Earth (about 4.0 Ga). We estimated the degree of impacts from the craters on the Moon and found that over 70% of Earth’s surface area could have been covered by the melts (Shibaike et al. 2016).