Title: The MOSDEF survey: a stellar mass-SFR-metallicity relation exists at z∼2.3
Authors: Ryan L. Sanders, Alice E. Shapley, Mariska Kriek, William R. Freeman, Naveen A. Reddy, Brian Siana, Alison L. Coil, Bahram Mobasher, Romeel Davé, Irene Shivaei, Mojegan Azadi, Sedona H. Price, Gene Leung, Tara Fetherholf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro
First author’s institution: University of California, Davis
Journal: Published in APJ; open access on arxiv
Galaxy evolution is a complicated thing! Our current theory is that gas comes in, stars get made & explode, the surrounding interstellar medium (ISM) heats up and gets enriched with metals, and then gas goes out. These processes are happening in different stages all across the galaxy and can make simulating and observing galaxy evolution very difficult. Thankfully, through years of observation of local galaxies, we know that some galactic properties are correlated! For example, Tremonti et al. (2004) showed that there is a relation between stellar mass (M∗) and gas-phase oxygen abundance (12 + log(O/H) or Z) in the local universe (redshift, z ~ 0). In 2008, Ellison et al. discovered that this M∗-Z relation also has a dependence on the star-formation rate (SFR), in the local universe. This relation was shown later to be more correlated than the M∗-Z relation on it’s own!
The questions that then arise are: is there evidence for a M∗-SFR-Z relation at high-z? And if so, does it agree with the one at z~0? Or does it evolve with redshift? Many have tried to answer these questions, but most of these studies were based on large samples with low signal-to-noise (S/N) or small samples with intermediate S/N and have relied on a single metallicity indicator. But not anymore!
What did they do?
Completed in May 2016, the MOSFIRE Deep Evolution Field (MOSDEF) survey was a 4-year program in which the MOSFIRE instrument on the 10 m Keck 1 telescope was used to get near-IR spectra of ~1500 galaxies spanning redshifts 1.4 < z < 3.8. The authors elected to use the ~700 galaxies observed in the 2.01-2.61 redshift range. After S/N cuts, the authors were left with a 260 galaxy sample with an average redshift of z ~ 2.3 (see Figure 1, left). To make a conclusion about the (possible) redshift evolution of the FMR, the authors used a comparison sample of 208,529 star-forming galaxies at z~0 from Andrews & Martini (2013).
From their near-IR spectra, the authors measured SFRs from the H-alpha luminosities (which were then reddening corrected). This correction is needed because dust grains between us and the galaxies scatter blue light and cause the galaxies to appear redder than they really are. Using broad- and mid-band photometric fits, the authors were able to robustly determine galaxy masses. Finally, using six emission-line ratios (N2, O3N2, N2O2, O3, R32, & O32), they were able to get several metallicity estimates for each galaxy. Here, metallicity refers to the gas-phase oxygen abundance, 12 + log(O/H).
What did they Find?
The authors detected a M∗-SFR-Z relation at z ~2.3! This is best shown in Figure 2. This relation was found using the metallicity estimates from O3N2, N2, and N2O2. The ratios for R32 and O3 are double-valued with metallicity (think of these like a parabola) and can’t be used empirically to discover a relation like this. They can, however, be used to support a finding; in this case, the results from R32 and O3 are consistent with those found from the other three. Results from O32 were inconsistent with their findings and the authors concluded that this was likely due to biases in the reddening correction. Another main goal of this project was to determine if the M∗-SFR-Z relation evolved with redshift, which the authors found! At a given mass and SFR, the metallicity of the z~2.3 sample is 0.1 dex less than the z~0 sample, also shown in Figure 2. The authors speculate that this evolution may be caused by an increase in the mass-loading factor from z~0 to z~2 and by a decrease in the metallicity of infalling gas at z~2.
What’s Left to Discover?
The authors established that a M∗-SFR-Z relation exists at z~2.3 and that this relation evolves with redshift. The existence of this relation implies that our understanding of galaxy evolution is right… at least up to z~2.3. The next step is to investigate this relation at higher redshifts, but that is no trivial task. As shown through the use of five emission-line ratios, measuring metallicities at high redshift can be difficult and will take great care. Uncertainties and inconsistencies with reddening corrections and other calibrations can cause large uncertainties in the results, like the case with O32. Thankfully, the introduction of large telescopes (like JWST and TMT) will allow us to lessen these uncertainties through their increased sensitivities.