arXiv: Aravena, Spilker et al. 2016
A while ago, I posted about some low-resolution observations and my own high-resolution observations of the molecular gas in two of the SPT lensed sources. As with most things in our group’s follow-up observations of these sources, however, we’re trying to have a uniform set of information for every source, which makes it easier to interpret the population as a whole. In this paper, Manuel Aravena and I look at the low-lying transitions (1-0 and 2-1) of the CO molecule in a large sample of the SPT sources, which is probably the most robust way to understand the molecular fuel from which stars form in galaxies.
We start out by pointing out that you can’t actually infer the magnification factor of lensed galaxies using only the width of the CO line. This was an idea that’s been floating around in the literature for several years, but it’s been fairly hard to test because most samples of lensed galaxies didn’t have both CO observations and an estimate of their magnification. We compiled a number of objects from the literature and combined them with the magnification of our own sources which I derived. This method is especially inaccurate at low magnifications, but doesn’t do better than about a factor of 2 even at high magnifications. We think this might be either because some of the lens models are too challenging (they have multiple lens galaxies or weird lens geometries), but it may also just be a selection effect, because it’s hard to observe unlensed objects which are faint in CO emission.
Next, we estimated the size of the CO-emitting regions. Even though our observations didn’t spatially resolve the sources in this case, you can still get a handle on the emitting region if you assume the CO emission is optically thick (almost always the case) and has the same excitation temperature as the dust emission (which we have good constraints on). It looks like the sizes of the CO-emitting regions are, on average, larger than the size of the dust emitting regions. If this sounds familiar, it’s because it’s also what we found in my previous work and seems to be pretty common in galaxies as a whole. There were a handful of objects where the inferred CO size was actually smaller than the dust size, so it will be interesting to see if those sizes pan out with higher-resolution CO observations.
Of course, the real strength of observing low-J CO lines is that they’re good tracers of the total mass of molecular material in galaxies – you don’t have to convert from a higher-excitation line, and we have enough long-wavelength photometry to know what the dust (and dust-to-gas ratio) is doing. Manuel closes out by looking at how the gas fraction (ie, gas mass / total mass) evolves in highly star-forming galaxies over the history of the universe. Since this is directly what drives the ability of galaxies to form stars, it’s really important to be able to link the gas fuel with its product, stars. We find that the gas fraction rises quickly out to about z ~ 2, after which it flattens out at around 40%. This is consistent with previous measurements, and extends them even further into the past history of the universe.
All in all, I’m pretty happy with how this study turned out (as you might guess from this longer-than-usual writeup). We took a lot of data to make this paper possible, and it’s currently the most complete set of low-J CO follow-up observations of any sample. Many, many thanks to the folks at ATCA for their long history of support for this program.