The prototypical starburst galaxy NGC253

NGC253 is one of the nearest starburst systems at a distance of only 3.5 Mpc. It is considered one of the prototypical starburst galaxies with a star formation rate surface density of ΣSFR~102 M yr-1 kpc-2 in the nuclear region and a molecular depletion time that is 5-25 times lower than what is found in local disks. A galactic wind emerges from the central 200 pc of NGC253 that has been characterized in H , X-ray, as well as neutral and molecular gas emission. Due to the close proximity, starburst and galactic winds can be studied in detail and individual structures can be resolved.


False color image of the velocity structure in the center of NGC253
The central ~800 pc of NGC253. The false color image shows CO(3-2) gas moving towards us (blue), at systemic velocity (green) and moving away from us (red). Note the structures perpendicular to the major axis of the central edge-on gas disk, e.g. towards the bottom center. Their mismatching color makes them stand out and indicates the different velocity structure. These so-called streamers are molecular outflows launched by feedback of the intense starburst.

The molecular outflow in NGC253 at a resolution of two parsecs

In Krieger et al. 2019a, we study the molecular outflow in NGC253 in very high resolution of up to 2.5 pc in three CO lines. Thanks to the hig reslution, we are able to kinematically decompose the observed emission into a disk and non–disk component. A significant amount of 7-16% of the CO luminosity can be associated with the non-disk component. Of the total molecular gas mass of ~3.6×108M in the center of NGC253, 0.5×108M (~15%) in the non-disk component. The high-resolution CO(3–2) observations allow us to identify the molecular outflow within the non-disk gas. In the derivation of the physical properties of the outflow (molecular mass outflow rate, kinetic energy and momentum), the unknown outflow geometry and launching sites of the outflow are the primary source of uncertainty. The observed gas distribution in position-position-velocity space is projected on the plane of the sky and needs to be deprojected to obtain physically meaningful quantities. This is a particularly difficult task but can be done under the assumption of different geometries that all can reproduce the observed emission.

Molecular mass outflow rate in NGC253
The molecular mass outflow rate in NGC253 as a function of distance from launching site. Only with the high spatial resolution of 2.5 pc, we are able to kinematically separate outflows from disk emission and other kinematic components.

The figure above shows the deprojected molecular mass outflow rate as a function of distance from the estimated launching site. Out to ~300 pc, the rate is approximately constant at ~39 M yr-1. Assuming an alternative outflow evolution, it can be as low as ~14 M yr-1. At high-redshift where we don't have such detailed observations, the geometry is even less uncertain which causes large systematic unvertainties that are often not accounted for. The majority of this outflow rate is contributed by distinct outflows oriented perpendicular to the disk, with a significant contribution by diffuse molecular gas. The mass loading factor η = Ṁout / ṀSFR is the range η ~ 8-20, already a high value when considering that the star formation rate in NGC253 is relatively low compared to some other local starbursts and especially high-redshift star forming galaxies. The kinetic energy of the outflow is 2.5-4.5×1054 erg which is only ~0.1% of the total or ~8% of the kinetic energy supplied by the starburst. The outflow momentum is 4.8-8.7×108 M  km s-1 or ~2.5-4% of the kinetic momentum released into the ISM by feedback.
This detailed case study highlights some important uncertainties for outflow analyses that are even more relevant for lower resolution data at low and high redshifts.

The molecular ISM in the super star clusters of the starburst NGC253

Leroy et al. (2018) detected 14 forming super star clusters (SSCs) in 350 GHz submillimeter continuum in the center of NGC253. One of them has been known before and is visible in IR imaging but the other 13 are still deeply embedded in their natal gas and dust clouds. Invisible in optical and IR wavelengths due to extinction, sub-/millimeter or radio observations are necessary to penetrate into the dense clouds surrounding them.
In Krieger et al. (2019b), we study the ISM of the molecular clouds around the SSCs to determine the conditions in the gas where the super star clusters form and outflows are probably launched from. The 2.5 pc resolution of our ALMA cycle 3 observations in band 7 (~350 GHz) approach the size of the SSCs.

SSC 14 spectrum LSB SSC 14 spectrum USB
High-resolution spectrum on a super star cluster in the center of the NGC253 starburst. The spectrum is obtained from a single pixel at ~2.5 km s-1 spectral resolution and 2.5 pc spatial resolution. A fit including the labeled species is plotted in red on top of the observed spectrum (in black). (Krieger et al. 2019b)

In the 14 sources, we detect up to 55 lines of 14 species in the spectral windows 342.0-345.8 GHz and 353.9-357.7 GHz. The SSCs differ significantly in chemical complexity between 5 and 15 detected species. Multiple spectral components are present in most SSCs in CO, HCN,HCO+ and CS, some of which are probably the result of (self-)absorption. We further detect HCN isotopologues and isomers (H13CN, HC15N, H15NC), abundant HC3N, SO and S18O, SO2 and H2CS.
The dense gas ratios CO/HCN, CO/HCO+ of ~1-10 are low which implies high dense gas fractions in the SSCs. We compare the observed line ratios to chemical models and find that all SSCs are consistent with photon-dominated regions. None of the SSCs near the galactic center show evidence for an X-ray dominated region due to a potential AGN. The existence of a low luminostiy AGN in NGC253 thus stays elusive.
The frequent detection of vibrationally excited HCN and HC3N species implies strong IR radiation fields that are potentially trapped by a greenhouse effect due to high continuum opacities. This is on contrast to a photon-dominated region as implied by the line ratios. Hence, the energy input in the SSCs is either dominated by mechanical heating with a contribution by UV heating or the HC3 emission originates from different regions in the clouds than other lines.
When fitting the observed spectra with XCLASS, we find high gas temperatures in most sources with an average ~130 K rotational temperature in SO2.