[Paper Review] The New Generation Planetary Population Synthesis (NGPPS) II. Planetary population of solar-like stars and overview of statistical results

Alexandre Emsenhuber, C. Mordasini+4 more|arXiv (Cornell University)|Jul 10, 2020

Astro and Planetary Science16 citations

TL;DR

This study presents the New Generation Planetary Population Synthesis (NGPPS) model to simulate the formation and evolution of planetary systems around solar-like stars using the Generation III Bern model. By simulating five populations with 1–100 embryos per disc, it finds that systems with at least 10 embryos reproduce giant planet demographics, while only the 100-embryo population achieves the giant-impact stage, yielding 8 planets >1 M⊕ per system on average, with 18% hosting giant planets and 1.6% hosting them beyond 10 au.

ABSTRACT

Context. Planetary formation and evolution is a combination of multiple interlinked processes. Constraining the mechanisms observationally requires statistical comparison to a large diversity of planetary systems. Aims. We want to understand global observable consequences of different physical processes (accretion, migration, and interactions) and initial properties (like disc masses and metallicities) on the demographics of the planetary population. We also want to study the convergence of our scheme with respect to one initial condition, the initial number of planetary embryo in each disc. Methods. We selected distributions of initial conditions that are representative of known protoplanetary discs. Then, we used the Generation III Bern model to perform planetary population synthesis. We synthesise five populations with each a different initial number of Moon-mass embryos per disc: 1, 10, 20, 50, and 100. The last is our nominal population consisting of 1000 stars (systems) that was used for an extensive statistical analysis of planetary systems around 1 M⊙ stars. Results. The properties of giant planets do not change much as long as there are at least ten embryos in each system. The study of giants can thus be done with simulations requiring less computational resources. For inner terrestrial planets, only the 100-embryos population is able to attain the giant-impact stage. In that population, each planetary system contains, on average, eight planets more massive than 1 M⊕. The fraction of systems with giants planets at all orbital distances is 18%, but only 1.6% are at >10 au. Systems with giants contain on average 1.6 such planets. The planetary mass function varies as M−2 between 5 and 50 M⊕. Both at lower and higher masses, it follows approximately M−1. The frequency of terrestrial and super-Earth planets peaks at a stellar [Fe/H] of −0.2 and 0.0, respectively, being limited at lower [Fe/H] by a lack of building blocks, and by (for them) detrimental growth of more massive dynamically active planets at higher [Fe/H]. The frequency of more massive planets (Neptunian, giants) increases monotonically with [Fe/H]. The fast migration of planets in the 5–50 M⊕ range is reduced by the presence of multiple lower-mass inner planets in the multi-embryos populations. To assess the impact of parameters and model assumptions, we also study two non-nominal populations: insitu formation without gas-driven migration, and a different initial planetesimal surface density. Conclusions. We present one of the most comprehensive simulations of (exo)planetary system formation and evolution to date. For observations, the syntheses provides a large data set to search for comparison synthetic planetary systems that show how these systems have come into existence. The systems, including their full formation and evolution tracks are available online. For theory, they provide the framework to observationally test the global statistical consequences of theoretical models for specific physical processes. This is an important ingredient towards the development of a standard model of planetary formation and evolution.

Motivation & Objective

• To understand the global observable consequences of physical processes (accretion, migration, interactions) and initial conditions (disc mass, metallicity) on planetary demographics.
• To assess convergence of planetary population synthesis with respect to initial embryo count.
• To identify minimal simulation requirements for capturing key planetary system properties, especially for giant planets.
• To provide a comprehensive, publicly available synthetic data set for observational comparison and model testing.
• To support the development of a standard model of planetary system formation and evolution.

Proposed method

• Simulates five planetary populations with 1, 10, 20, 50, and 100 Moon-mass embryos per protoplanetary disc.
• Uses the Generation III Bern model to perform full N-body and disk-planet interaction simulations over 10 Myr.
• Applies representative initial distributions of disc masses, metallicities, and planetesimal surface densities.
• Analyzes statistical outcomes including planet multiplicity, orbital distribution, mass function, and metallicity dependence.
• Compares results across embryo populations to assess convergence and sensitivity to initial conditions.
• Conducts two non-nominal simulations: in-situ formation without gas-driven migration and a steeper initial planetesimal surface density.

Experimental results

Research questions

• RQ1 Does the planetary population converge with respect to initial embryo count, particularly for giant planets?
• RQ2 What is the minimum number of embryos needed to reproduce observed planetary demographics?
• RQ3 How do dynamical interactions and migration affect the final architecture of planetary systems?
• RQ4 How do planet mass functions and occurrence rates vary with stellar metallicity?
• RQ5 How do model assumptions such as migration and initial planetesimal density affect population outcomes?

Key findings

• The average planetary system with 100 embryos contains 8 planets more massive than 1 M⊕, with 18% hosting giant planets and 1.6% hosting them beyond 10 au.
• The planet mass function follows M−2 between 5 and 50 M⊕, and approximately M−1 at both lower and higher masses.
• Terrestrial and super-Earth planet frequencies peak at stellar [Fe/H] = -0.2 and 0.0, respectively, due to building block scarcity at low metallicity and dynamical disruption at high metallicity.
• Giant planet occurrence increases monotonically with [Fe/H], consistent with observations.
• The presence of multiple low-mass inner planets in multi-embryo systems reduces fast migration in the 5–50 M⊕ range.
• Results for planets ≥10 M⊕ are robust across embryo populations with ≥10 embryos, enabling resource-efficient simulations for gas accretion and giant planet fraction studies.