(U)LIRGs : Supernova Fireworks

(U)LIRGs must have high rates of supernova because they have high SFR. A number of repeated IR and radio observations on a few local (U)LIRGs have shown that they indeed have high occurrence of supernovae and/or supernova remnants. However, this is yet to be demonstrated beyond the local universe. The difficulty is two-fold. First, one would need to do this in the rest-frame IR or radio to circumvent the severe dust extinction. Second, one would need high resolution to discern the SNe. Both become exceedingly difficult beyond the local universe. In Yan et al. (2018, ApJ, 867, 21), we presented indirect evidence by using the data in the Spitzer IRAC Dark Field (IDF), whose 3.6/4.5 μm observations spanned 14+ years on a nearly bi-weekly basis. Herschel/SPIRE also observed this area (as it is the region that has the lowest IR background), so we have (U)LIRGs by SPIRE detections. IRAC did not have the spatial resolution to discern SNe individually; however, the outburst of SN would change the brightness of the host galaxy. Therefore, by searching for 3.6/4.5 μm variability of the (U)LIRG hosts, the high rate of SNe could be revealed. We indeed found two very promising cases, whose IRAC Ch1/2 light curves are shown below in the left panel (the middle two objects). Their "net" variations after subtracting the host light are shown in the right panel, which are consistent with the summation of multiple SNe overlapping in time.

A JWST monitoring program will be able to resolve such SNe in (U)LIRGs, which will open up a new window of studying high-z SNe. The IDF is the best field for this purpose because it is in the JWST's CVZ and thus is the easiest to schedule observations.

"Red" but not "Dead" : W4BCGs

Elliptical galaxies are thought to be dominated by old stellar populations ("red") and have long ceased their star formation ("dead"). But it's known that there are counter-examples. Some of the most interesting counter-examples are found among the brightest cluster galaxies (BCGs). The vast majority of the BCG family are indeed red-and-dead, but a very small minority are not: they reside in the so-called "cool-core" (CC) clusters, and they can have SFRs from a few to ~ 100 M_solar/yr fueled by "cooling flows".

While exploring the WISE data archive, we have found an interesting population, which we dubbed the "W4BCGs". Runge & Yan (2018, ApJ, 853, 47; which was James' thesis paper) presented the analysis on a sample of 389 such objects (the largest sample of its kind): they are brightest cluster galaxies (BCGs) selected from the SDSS data, so in theory they should be some of the best representatives of local elliptical galaxies (i.e., high-mass and and long being "dead"); but these are bright in the WISE W4 band (22 μm) and not like anything passively evolving. They are IR-luminous, with the median IR luminosity well in the LIRG regime; in fact, > 27% of them qualify as ULIRGs. While some of them are AGNs, the majority are not (as shown the BPT diagnostics) and they should owe their strong IR emissions to dust-enshrouded star formation (the latter is confirmed by a small sub-sample that has Herschel SPIRE data). Ten of the W4BCGs have Chandra X-ray data, and it seems that seven of them might be in CC clusters. However, we show that in most cases (5 out of these 7) the mass deposition rates from cooling flows cannot account for their observed SFRs. This casts doubt on the theory that cooling flows are the cause of the star formation in non-quiescent BCGs.

ULIRG Quasars : starburst and AGN have nothing to do with each other

"AGN feedback" is an often-used mechanism in theoretical works to turn off star formation in galaxies so that they don't grow too big to be matched with observations. On the other hand, there have also been suggestions that this "feedback" could be "positive" so that the AGN could somehow trigger/enhance star formation. Quasars detected in sub-mm/mm are used as examples for the latter. But really?

In Ma & Yan (2015, ApJ, 811, 58), we assembled a large sample of quasars that have strong far-IR emissions and studied the co-evolution of starburst and AGN. This sample has 354 quasars spanning 0.14 ≤ z ≤ 4.7, and was derived based on the SDSS quasar sample and the far-IR data from all the Herschel wide-field surveys. This paper, while being only part of Zhiyuan's final thesis, is 25 pages long and contains a lot of interesting analysis and implications. If I were to highlight, two points are the most important in terms of the "big picture": (1) the amplitude of the starburst in the quasar host galaxies (indicated by the IR luminosity) has nothing to do with the AGN strength (in terms of quasar absolute magnitude, black hole mass, or X-ray luminosity), and (2) the fraction of IR-luminous quasars peaks at around 1.5 ≲ z ≲ 2.5, which is different from the distribution of the parent SDSS quasar sample. See figures below. Both suggest that starburst has nothing to do with AGN if we see them co-exist in galaxies.

Dusty starbursting regions in ULIRGs must be of small sizes (r ≲ 2 kpc)

Inspired by one of the preliminary conclusions in Ma & Yan (2015, ApJ, 811, 58), we carried out a further analysis of the L_IR-T relation (which was first observed in SMGs). As here we only cared about L_IR, we used a much enlarged far-IR quasar sample. The details are given in Yan & Ma (2016, ApJ, 820, L16). Basically, an object following the L_IR-T relation is best understood using the equivalent of the Stefan-Boltzmann law for an MBB with an effective radius of R_eff. The very existence of the L_IR-T relation means that the size of the dusty starbursting regions in an ULIRG must be small (with R_eff ≲ 2 kpc). While strictly speaking R_eff is not exactly the physical size, it is a good proxy. Outside of the region outlined by the L_IR-T relation is the "zone of avoidance" in terms of minimum star formation rate density of the starbursting region ("minSFRD" as we call it).