Research

MAIN RESEARCH INTERESTS:

• infrared galaxies
• obscured quasars
• radio properties of infrared galaxies and AGN
• SED modelling and decomoposition
• mid-IR spectroscopy

SOME BACKGROUND (for students)

In 1996, the COBE satellite discovered the long predicted Cosmic Infrared Background (CIB) which is the cumulative light of all dusty sources from the first galaxies up until today. On average, local galaxies emit far more in the optical than in the infrared (due to most of the larger ones being essentially done with their star-formation). The infrared background was found to be comparable in strength to the optical background, implying that by contrast, sources in the past were on average much more infrared luminous or in other words were much more active star-formers. The sources that dominate this background are believed to be largely luminous and ultraluminous infrared galaxies at high redshift (z>~1). Classic hierarchical structure formation suggests that galaxies form by the mergers of smaller progenitors -- i.e. bottom-up. In this picture the most massive galaxies should be the ones that have formed their stars the latest. However, observations have long shown that the massive elliptical galaxies of today are dominated by fairly old stellar populations while the smaller galaxies are relatively younger. Different studies but most notably the analysis of a the optical spectra of a large sample of SDSS galaxies showed that indeed the formation of galaxies (in the sense of stellar mass build-up) is earliest in the most massive galaxies (whose stars are mostly in place by z~2), through intermediate mass galaxies such as most spirals (whose peak formation appears to have been at z~0.5-1.5 and was likely the driver behind the extremely strong evolution in the star-formation rate density with redshift between z~0 and z~1 (the Lilly-Madau plot), until today the most active galaxies (relatively speaking) are the dwarfs.

Meanwhile, it had become increasingly clear that not only all (large) galaxies contain supermassive nuclear black holes, but that the growth of those SMBHs is likely linked with the growth of the galaxy itself as evidenced by the correlation between black hole mass and stellar velocity dispersion (the Magorrian relation). This is also supported by the finding that the peak in the number of luminous quasars with redshift roughly coincides with the peak in the star-formation rate density. Although it is likely that the story is more complicated than that, one possible evolutionary path is: two gas rich progenitors merge triggering infall of gas in the central regions which leads to increased star-formation rates (up to ~1000Msun/yr) and as the central gas conccentration increases, the black hole begins to accrete more and more gas and thus grow. The accretion disk that forms produces prodigious amoiunts of energetic photons -- which is what we see as an active galactic nucleus (AGN) or in the case of the highest luminosities AGN -- a quasar. This activity is relatively short lived (maybe about 10 million years) as fairly quickly feedback processes kick in to quench it. Feedback takes the form of winds either driven by exploding supernovae or potentially the AGN itself which drive material outwards and thus quench further star-formation or black hole growth. Many quasars, especially at high redshift, are observed to be obscured behind large column densities of gas and dust. The classic picture here is that all AGN just outside their accretion disk have a structure known as a dust torus which means that depending on the orientation one either sees the quasar directly (unobscured) or through all that dust (obscured). The above evolutionary scenario however implies that at least some of the obscured quasars (and here the fraction of obscured quasars observed and its change iwth redshift and luminosity is critical) are in fact in this just past a merger stage when their host galaxy is also just forming and there is a great deal of both gas and dust obscuring the central regions more or less from all sides. Once feedback has pushed some of that material out of the way, the naked quasar would be revealed if only briefly (while the remainig fuel is exhausted). Many of the sources I have been studying lately are extremely dusty and luminous z~2 galaxies which host both starbursts and highly obscured quasars.

Below is a brief outline of the sort of things I've been working on lately. You'll have to ask me in person for more details.

SEDs

The attached figure shows the UV-to-IR SED of one of the z~2 obscured quasars I have been studying. The blue line shows the average SED of unobscured quasars from the SDSS survey (Richards et al. 2006). I have applied strong obscuration to this template (Av=7 based on Milky Way like dust) in order to get better agreement with the infrared data. The reproduce the UV-optical points however requires the addition of a host galaxy SED (the dot-dashed line).

One striking observervation was the finding of Polycyclic Aromatic Hydrocarbons (PAH) emission features in the spectra of z~2 sources. These features are fairly ubiquitous in the spectra of local star-forming galaxies but these objects are an order of magnitude more luminous than the most luminous local galaxies that are dominated by star-formation.

The above suggests that the dust properties in galaxies at these high redshifts are not too different from those observed today. However, the relative strength of the starbursts witnessed are unprecedented locally (the general idea is that greater gas densities, or available to be accreted, in the progenitors of these galaxies might be responsible). However the assumption that the dust properties have not varied too much from the past to day is yet unproven though crucial in our interpretation of these results. To address that, we need to observe features arising from different dust species. One of my current projects involves just that: it is deep mid-IR spectra of sample of 10 z~2 sources selected on the basis of deep 9.7 micron absorption features (arising from Si-O stretch in silicate grains). The new data allow to observe some other common molecular absorption features that have however never been seen before at those redshifts. Those include a 3.4um absorption feature due to the C-H stretch in hydrocarbons and the 3 and 6um water ice absorption bands arising from the bending and stretching modes respectively.

RADIO PROPERTIES:

One of the most interesting discoveries from our studies was the finding of significant radio activity including radio jets in some of our sources (see Figure). This combination of extreme obscuration and prominent radio jets has never been observed in the local Universe. This is an indication that we might be witnessing the transition between the dust obscured ULIRGs and classical quasars.

OUTFLOWS:

The radio jets shown above are one form of outflows (i.e. where material is moving out of the central regions of the galaxy. We have observed near-IR spectra (rest-frame optical) for many of our sources. These allow us to determine the redshifts with greater accuracy and also to look for additional signatures of the buried AGN. The figure below shows a serendipitous discovery among these spectra of a source whose OIII emission line profile shows not only the typical (in obscured AGN) narrow emission lines but also a broad blueshifted component which is an indication of gas moving toward us or in other words -- an outflow of ~260km/s.

THE COSMOLOGICAL EVOLUTION OF OBSCURED STAR-FORMATION AND QUASAR ACTIVITY:

I am interested in applying what we have learned from the detailed multiwavelength studies of these extreme sources to better understanding their role in the stellar mass and black hole build-up and the potential co-evolution of the two processes. As a first step in that direction we have constructed the mid-IR (8um) luminosity function for obscured quasars at z~0.6 and z~2. The results suggests that indeed the obscured quasars outnumber the obscured ones (Yan et al. 2009, in prep.).

RESEARCH WITH STUDENTS

I have been working with Anna Pancoast, a Haverford senior, on a project to study the morphological distribution of the dust obscuration and star-formation in NGC2442 (see Figure), a nearby southern sky galaxy that has recently undergone a tidal interaction with a companion distorting its spiral shape and triggering star-formation activity. All of the activity discussed above (at z~2) is believed to be merger driven, yet the details of this process are not clear. The state-of-the-art simulations still have trouble going down to sufficiently small scales in addition to as yet poorly understood details of the star-formation process. Hence such observations would provide invaluable insights into how is star-formation triggered following merger/interactions and provide constrains for future simulations of the process.