Star formation in the CDFS: observations confront simulations

TL;DR

This study compares observed star formation rates in the CDFS with galaxy formation models, revealing that models underestimate the rapid increase in growth rates from redshift 0 to 1, indicating a need for model refinement.

Contribution

It provides the first detailed comparison of observed specific star formation rate evolution with semi-analytic models across a broad redshift range.

Findings

Observed sSFR increases with redshift across all masses.

Models predict similar sSFR growth but underestimate the rate increase from z=0 to z=1.

Discrepancies suggest models need adaptation to match observed steep growth.

Abstract

We investigate the star formation history of the universe using FIREWORKS, a multiwavelength survey of the CDFS. We study the evolution of the specific star formation rate (sSFR) with redshift in different mass bins from z = 0 to z ~ 3. We find that the sSFR increases with redshift for all masses. The logarithmic increase of the sSFR with redshift is nearly independent of mass, but this cannot yet be verified at the lowest-mass bins at z > 0.8, due to incompleteness. We convert the sSFRs to a dimensionless growth rate to facilitate a comparison with a semi-analytic galaxy formation model that was implemented on the Millennium Simulation. The model predicts that the growth rates and sSFRs increase similarly with redshift for all masses, consistent with the observations. However, we find that for all masses, the inferred observed growth rates increase more rapidly with redshift than the model predictions. We discuss several possible causes for this discrepancy, ranging from field-to-field variance, conversions to SFR, and shape of the IMF. We find that none of these can solve the discrepancy completely. We conclude that the models need to be adapted to produce the steep increase in growth rate between redshift z=0 and z=1.