Dynamical Versus Stellar Masses in Compact Early-Type Galaxies: Further Evidence for Systematic Variation in the Stellar Initial Mass Function

TL;DR

This study provides strong evidence that the stellar initial mass function (IMF) in compact early-type galaxies varies systematically with galaxy velocity dispersion, becoming more bottom-heavy at higher dispersions, based on spectral analysis and dynamical modeling.

Contribution

It demonstrates consistent IMF variation evidence from both spectral features and dynamical models in a specific galaxy sample, supporting non-universal IMF hypotheses.

Findings

IMF becomes bottom-heavy at high sigma

Dynamical and spectral methods agree on IMF variation

IMF variation correlates with galaxy velocity dispersion

Abstract

Several independent lines of evidence suggest that the stellar initial mass function (IMF) in early-type galaxies becomes increasingly `bottom-heavy' with increasing galaxy mass and/or velocity dispersion, sigma. Here we consider evidence for IMF variation in a sample of relatively compact early-type galaxies drawn from the Sloan Digital Sky Survey. These galaxies are of sufficiently high stellar density that a dark halo likely makes a minor contribution to the total dynamical mass, Mdyn, within one effective radius. We fit our detailed stellar population synthesis models to the stacked absorption line spectra of these galaxies in bins of sigma and find evidence from IMF-sensitive spectral features for a bottom-heavy IMF at high sigma. We also apply simple `mass-follows-light' dynamical models to the same data and find that Mdyn is significantly higher than what would be expected if these galaxies were stellar dominated and had a universal Milky Way IMF. Adopting Mdyn M_* therefore implies that the IMF is `heavier' at high sigma. Most importantly, the quantitative amount of inferred IMF variation is very similar between the two techniques, agreeing to within 0.1 dex in mass. The agreement between two independent techniques, when applied to the same data, provides compelling evidence for systematic variation in the IMF as a function of early-type galaxy velocity dispersion. Any alternative explanations must reproduce both the results from dynamical and stellar population-based techniques.