The Massive Ancient Galaxies at z>3 NEar-infrared (MAGAZ3NE) Survey is a multi-semester survey using the MOSFIRE instrument on the Keck telescopes in Hawaii. The project has spectroscopically confirmed the largest sample to date of ultramassive galaxies (stellar masses greater than 100 billion suns) from when the Universe was less than 2 billion years old.
Selection of candidate galaxies was done via analysis of deep multi-wavelength photometric catalogs in the VIDEO-XMM and COSMOS-UltraVISTA fields.
Spectral features including [OII] and [OIII] emission lines, Balmer series absorption, and the D4000 break were critical to putting caps on the ongoing star formation and AGN activity, as well as constraining the average age of stars in a galaxy.
Detailed modeling of the star-formation histories for the MAGAZ3NE sample indicates that these galaxies formed the majority of their stars in bursts forming hundreds to thousands of solar masses per year for several hundred million years.
Roughly half of the sample subsequently had their star formation abruptly truncated, putting constraints on the processes responsible for quenching at these early epochs. These galaxies may have run out of pristine gas with which to form new stars, or activity from a supermassive black hole may have snuffed out further star formation. Either way, the early epoch at which such an event must have occurred challenges current models of early galaxy star formation and quenching.
More details are available in our two publications thus far. The most extreme of our sample received significant press coverage on sites such as Scientific American, USA Today, CBC, Fox News, and included an interview for the BBC Newsroom.
Models shown in red. Note the clear Balmer absorption features and lack of emission lines.
The star-forming main sequence of 3<z<4 galaxies in COSMOS UltraVISTA DR3 (black), along with the confirmed UMGs (red-blue).
Left: The redshift at which the UMGs formed ~50% of their stellar mass. Right: The redshift at which the UMGs quenched their star formation.
The Multi-Object Spectroscopic Emission Line (MOSEL) Survey used the MOSFIRE instrument on the Keck telescope to follow-up a sample of galaxies with strong rest-frame optical emission features at 2.5<z<4 first published in Forrest, et al., 2017.
Spectroscopic confirmation of nearly 50 of these objects revealed low-mass galaxies with large gas fractions and [OIII]5007 equivalent widths. This indicates that the high star formation rates of these object could continue for some time, potentially doubling their stellar masses in several tens of Myr (Tran, et al., 2020).
Kinematic analysis reveals that the more massive of these objects have lower velocity dispersions than similar mass galaxies at lower redshifts, implying that ex situ stellar mass fractions increase with cosmic time (Gupta, et al., 2020).
The FourStar Galaxy Evolution Survey (ZFOURGE) was an international collaboration, with members in Australia, the U.S., and Europe.
With 5 near-infrared medium bandpasses and a deep K-band detection image, ZFOURGE provided precise photometric redshifts for galaxies out to z~4. This enabled a wealth of new scientific discoveries in the field of galaxy evolution.
A list of publications, as well as catalogs, images, and an entertaining data explorer are available on the survey website. The survey is described in detail in Straatman, et al., 2016.
Archival photometric bandpasses in the ZFOURGE fields are shown in gray, while the ZFOURGE medium bands are shown in orange.
With the increase in large surveys imaging the sky in multiple bandpasses, the number of galaxies with multi-wavelength imaging has increased dramatically over the last decade.
This information enables us to group galaxies together by the shapes of their observed spectral energy distributions (SEDs). When the galaxies in question span a range of redshifts, imaging probes different rest-frame wavelengths. By de-redshifting and combining the multiband imaging of galaxies across a redshift range, we fill in the ‘gaps’ in the photometry.
These composite SEDs allow analysis of features such as emission lines which are normally only available with spectroscopy. The details of our methodology are contained in Forrest, et al. (2018). This is based on the composite SED grouping method introduced in Kriek, et al. (2011).
Composite SEDs can be downloaded here. Users are highly advised to familiarize themselves with the details of construction before using.
Galaxies are statistically grouped together based upon their photometry, then deredshifted and scaled to increase the effective wavelength resolution of the composite SED.
The most massive, shortest lived stars emit large amounts of ultraviolet (UV) radiation. Due to their short lives, measuring the amount of UV light is therefore a reliable proxy for recent star formation.
However, many galaxies also have significant dust attenuation, which reduces the amount of this UV light observed- the energy is instead reemitted by the heated dust in the infrared (IR). As a result, measures of galaxy star formation use both UV and IR light when available.
The efficiency of this transformation depends on a variety of parameters, including how the dust is distributed relative to stars and what species of dust grain are present. By comparing the ratio of IR to UV light (the infrared excess, IRX) to the slope of the galaxy spectrum in the UV (β), one can infer the dust attenuation law.
In Forrest, et al. (2016), we built a set of composite SEDs from galaxies in the ZFOURGE survey at 1<z<3. Through the addition of far-IR data from the Candels-Herschel group, we plotted these composite SEDs on the IRX-β plane, finding results similar to a starburst attenuation curve.
In 2009, Cardamone, et al. found a population of galaxies in SDSS with impressively strong [OIII] 5007 emission. These ‘green pea’ galaxies have since been followed up with great interest and generally have low metallicities and high (specific) star formation rates.
Similar galaxies at higher redshifts, termed extreme emission line galaxies (EELGs) have now been found and studied extensively. Through the construction of composite SEDs, we have found such galaxies at 1<z<4 (Forrest, et al., 2017; Forrest, et al., 2018). EELGs appear more commonly at higher redshifts, and may have played a significant role in cosmic reionization.
While galaxies with very large [OIII]/[OII] line ratios have been theorized to possibly by Lyman continuum leakers, a stack of deep rest-frame UV imaging for a subsample of these galaxies yielded no detection (Naidu, et al., 2018).
A sample of these EELGs has also been spectroscopically observed as part of the MOSEL survey.
While many galaxies in the universe are actively forming new stars, there is a significant fraction that are not. While these ‘quiescent’ galaxies have become much more common over the lifetime of the cosmos, the mechanisms that cause galaxies to cease forming stars are not well understood.
Broadly speaking, if a galaxy runs out of gas to turn into stars, or if any remaining gas is prevented from collapsing gravitationally (via heating, ejection, or stabilization), star formation will shut down.
The key to unlocking the processes responsible is finding galaxies in the act of turning off their star formation. In Forrest, et al., 2018, we find a population of such galaxies, as identified by their composite SED D4000, emission lines, and dust content.
Interestingly, these galaxies have smaller sizes and larger Sérsic indices than star forming galaxies of similar mass, but larger sizes and smaller Sérsic indices than quiescent galaxies. These galaxies also become more prevalent as the universe ages, and their number densities imply quenching timescales on the order of 1-2 Gyr.