Matus Rybak
Will a newborn star feel a difference between being born in Orion today, and in a massive, dusty galaxy 10 billion years ago? Forming stars at rates 100x higher than any present-day galaxy, the dust-enshrouded, sub-mm bright galaxies (SMGs) present a challenge to current models of star-formation. To uncover what processes drive these extreme star factories and how they impact the newborn stars, we need to study the physical conditions of their molecular gas.
However, the studies of physical conditions in SMGs have been long restricted to unresolved observations. These provide only a galaxy-averaged view of the gas and dust thermodynamics, mixing the actual star-forming sites with cold, quiescent gas.
To push beyond this resolution limit, I combine the superb angular resolution of the ALMA telescope with the extra magnification provided by strong gravitational lensing. Using an image reconstruction tailored to radio-interferometers such as ALMA and extensive multi-tracer (dust, CO, C+) observations of dust and gas, I have mapped - for the first time - the physical conditions in a dusty galaxy just 2 billion years after the Big Band at an unprecedented, sub-kiloparsec resolution. These results provide the first view of the physical conditions in SMGs on scales comparable to nearby galaxies, how they vary across the galaxy, and reveal the extreme nature of its star-forming regions, with higher densities and stronger radiation fields than in the very heart of Orion.
Will a newborn star feel a difference between being born in the nearby Orion cloud today, and in a massive, dusty galaxy 10 billion years ago? At redshift z=2-5 (10-13 Gyr ago), up to half of all the stars lived in the so-called dusty, star-forming galaxies. With star-formation rates exceeding 1000 solar masses per year, DSFGs are the most intensely star-forming galaxies in the entire history of the Universe. The unparalleled star production in DSFGs presents a challenge to our understanding of galaxy evolution and star formation. To understand the physical processes of in DSFGs and how they impact the newborn stars, we need to study the gas thermodynamics (density distribution, temperature, radiation fields...) of the molecular gas from which the stars are born. These can be inferred by comparing different emission lines to predictions from radiative transfer models.
However, because of their small size and faintness, studies of physical conditions in DSFGs have been mostly limited to unresolved observations. These provide only a galaxy-averaged view of the gas and dust thermodynamics and can be severely biased if different emission lines arise from different parts of the source, as is seen in some resolved observations.
The solution is to resolve DSFGs on sub-kiloparsec (kpc) scales. In this work, we combined the superb angular resolution of ALMA (Atacama Large Sub-Millimeter Array) with extra magnification provided by strong gravitational lensing. Our target is SDP.81: a redshift z=3.042 strongly lensed galaxy, with high-resolution imaging from ALMA which covers the CO(5-4), (8-7), and (10-9) lines. We complemented this data by new observations of the [CII] emission (from the ionized carbon) and the low-excitation CO(3-2) line. All data has a resolution of 0.2 arcseconds.
We reconstruct the lensed source using a visibility-fitting code that provides better fidelity than working on post-processed images.
Thanks to the lensing, we achieve a mean resolution of ~0.025 arcsec!
The reconstructed images reveal the complex structure of SDP.81. The far-IR continuum (tracing the dust-obscured starburst) is very compact; the CO and [CII] emission (coming from the molecular gas) are much more extended. The [CII] emission stretches over more than 10 kiloparsecs in a spectacular tidal tail! Crucially, the very different distributions of the far-IR continuum, CO, and [CII] show that the physical conditions must vary rapidly on sub-kpc scales.
We then use the PDRToolbox database of radiative transfer models to infer the gas density, far-UV irradiation, etc. for each 200x200 pc pixel. For each pixel, we fit the observed line and continuum data (or upper limits where applicable). We took great care when assessing our error budget: we include extra uncertainties due to flux calibration and lens modelling systematics. This is the first time that we could map the physical conditions of star-forming gas in a high-redshift galaxy.
Somewhat surprisingly, the observed data is well-fit by rather simple models (1-D clouds, no shocks). This is partially due to our conservative uncertainties. We have recently obtained new, high-sensitivity observations of the CO(3-2) line and will use them to improve our radiative transfer models.
Our 200-pc maps of physical conditions in SDP.81 show a significant variation in far-UV radiation and gas density across the source. Comparing our maps to nearby star-forming regions, the conditions in SDP.81 turn out to be similar to those in the Trapezium cluster in the Orion cloud. The Trapezium cluster is already extreme in itself, but each of our 200x200 pc pixels in SDP.81 probably contains few star-forming regions with "void space" in between. So the actual star-forming sites in SDP.81 must be true monsters!