2007ApJ...662.1282M

The core-accretion and disk instability models have so far been used to explain planetary formation. These models have different conditions, such as planet mass, disk mass, and metallicity for formation of gas giants. The core-accretion model has a metallicity condition ([Fe/H]>-1.17 in the case of G-type stars), and the mass of planets formed is less than 6 times that of the Jupiter mass M_J. On the other hand, the disk instability model does not have the metallicity condition, but requires the disk to be 15 times more massive than the minimum mass solar nebulae model. The mass of planets formed is more than 2 M_J. These results are compared to the 161 detected planets for each spectral type of the central stars. The results show that 90% of the detected planets are consistent with the core-accretion model regardless of the spectral type. The remaining 10% are not in the region explained by the core-accretion model, but are explained by the disk instability model. We derived the metallicity dependence of the formation probability of gas giants for the core-accretion model. Comparing the result with the observed fraction having gas giants, they are found to be consistent. On the other hand, the observation cannot be explained by the disk instability model, because the condition for gas giant formation is independent of the metallicity. Consequently, most of planets detected so far are thought to have been formed by the core-accretion process, and the rest by the disk instability process.