Investigating the Formation of Planets Interior to in situ Hot Jupiters

TL;DR

This study tests the feasibility of forming hot Jupiters and their interior companions in situ during the gas-free, late-stage phase of planetary systems. Using REBOUND TRACE to run 159 simulations with 30 interior embryos across a range of solid-surface densities, the authors find that higher embryo densities produce more massive interior planets (up to ~$10\ M_\oplus$) and that about $75$–$80\%$ of the initial solid mass remains in planets after 1 Myr. A second suite varying the hot Jupiter's semi-major axis shows that closer-in giants drive more embryo-embryo mergers, while more distant giants lead to less dramatic growth; overall, the giant's position modulates outcomes but disk density dominates. The results imply that in situ formation is a viable pathway for some hot Jupiter systems with interior companions, but the mechanism cannot explain the full diversity of observed architectures and must be integrated with migration dynamics and long-term dynamical evolution in future work.

Abstract

The population of hot Jupiters with adjacent planetary companions is small but growing, and inner companions appear to be a nearly ubiquitous outcome within this subset of the exoplanet census. While most hot Jupiters are believed to form via tidal migration, the presence of adjacent companions is not easily explained by this formation mechanism, requiring consideration of additional formation mechanisms such as disk migration and in situ formation. In this work, we explore the possibility of in situ formation for both hot Jupiters and their interior companions. Using numerical simulations performed with the N-body integrator REBOUND, we investigate the growth of interior companions under various assumptions about disk conditions and hot Jupiter final orbital positions. Our results show that if a sufficiently high density of planetary embryos is transported to short orbital radii, it is feasible for both hot Jupiters and their interior companions to form in situ, providing a viable explanation for a subset of observed planetary architectures.

Figures (5)

• Figure 1: Systems selected from the IPAC Exoplanet Archive Christiansen2025 that contain a hot (orbital period $P\le10$ days) or warm Jupiter (orbital period $10<P<100$ days) and additional planets within $P<100$ days. Marker sizes are scaled by planet radius. Hot/warm Jupiters are denoted by light green circles (and identified as those meeting the above orbital period definitions and with radius $R_p> 0.8 R_J$), and smaller planets or cold Jupiters are denoted by dark green circles.
• Figure 2: A comparison between the planet size expected from a given initial surface density of planetesimals and the orbital location at which a hot Jupiter's core would need to have formed in order to deliver that surface density interior to its final orbit after migration. Bottom panel: The results of our numerical simulations showing the largest planet formed for a population of planetary embryos interior to the hot Jupiter with varying surface densities of solids. Each green dot represents the result from a unique simulation, and we overlay a best-fit model to the results to guide the eye. Top panel: For the range of surface density of solids considered in our simulations, we show the inferred formation radius of the core of the hot Jupiter for four different disk models.
• Figure 3: Mass of the largest planet formed as a function of the hot Jupiter's orbital position. Green points indicate simulation outcomes, and the black curve shows a polynomial fit to guide the eye.
• Figure 4: Frequency distribution of the most massive planet formed in each simulation from our first suite of simulations. Each color represents a bin of disk surface density (in $\text{g}/\text{cm}^2$), and the y-axis indicates the number of simulations that produced a most massive planet with the mass shown on the x-axis. In general, higher surface densities lead to the formation of more massive planets. Planet masses are given in Earth masses ($\text{M}_{\oplus}$).
• Figure 5: The major outcomes of embryo interactions as a function of the hot Jupiter's semi-major axis. Alternate outcomes are significantly less common (across all simulations, $\sim0.7\%$ accrete onto the star, $\sim 0.1\%$ are ejected from the system entirely and $\sim15\%$ are merger products or non-interacting embryos that survive until the end of the integration). Bottom panel: The percentage of embryos that merged into another embryo. Top panel: The percentage of embryos that were accreted onto the hot Jupiter. We find that most solid material remains dynamically bound and redistributed within the inner system rather than being removed entirely.