Untangling Stellar Components of Galaxies: Evaluation of Dynamical Decomposition Methods in Simulated Galaxies with GalaxyChop

TL;DR

This paper introduces JEHistogram, a new dynamical decomposition method for simulated galaxies, and evaluates its performance against existing methods, demonstrating improved accuracy especially in complex inner regions.

Contribution

The paper presents JEHistogram, a novel method for galaxy component separation, and provides a comprehensive evaluation against five established techniques using simulated galaxy data.

Findings

JEHistogram outperforms existing methods in accuracy, especially in galaxy cores.

All methods produce approximate properties of galaxy components, with JEHistogram showing superior precision.

The methods are made publicly available via the GalaxyChop Python package.

Abstract

Galaxy formation is intrinsically connected to the distinct evolutionary processes of disk and spheroidal systems, which are the fundamental stellar components of galaxies. Understanding the mutual dynamical interplay and co-evolution of these components requires a detailed dynamical analysis to allow for a disentanglement between them. We introduce JEHistogram, a new method for the dynamical decomposition of simulated galaxies into disk and spheroidal stellar components, utilizing the angular momentum and energy of star particles. We evaluate its performance against five previously established methods using a sample of equilibrium galaxies with stellar masses in the range $10^{10} \leq M_\mathrm{gal}/M_\odot \leq 10^{12}$. Our assessment involves several metrics, including the completeness and purity of stellar particle classification, scale lengths, mass density profiles, velocity dispersion, and rotational velocity profiles. While all methods approximate the properties of the original components, such as mass fractions and density or velocity profiles, JEHistogram demonstrates a better accuracy, particularly in the inner regions of galaxies where component overlap complicates separation. Additionally, we apply JEHistogram to a Milky Way-like galaxy from the IllustrisTNG cosmological simulations, showcasing its capability to derive properties like size, mass, velocity, color, and age of dynamically defined disk and spheroidal components. All dynamical decomposition methods analyzed are publicly accessible through the Python package GalaxyChop.