Research

Up until ~1980 or so, AGN (Active Galactic Nuclei, which will now include quasars) were divided into many puzzling phenomenological subtypes based on correlated suites of observed traits. Much of the confusion was cleared up in the following decade: it turns out that while we see dramatically varied behavior, most of it depends only on the inclination of these roughly axisymmetric sources to the line of sight to Earth! This isn't so surprising in retrospect: people have many systematic differences in appearance, depending on whether they are seen from the front or the side or the back. Of course in AGN, the symmetry is approximately axial rather than bilateral. Perhaps astronomers didn't find this intuitive at first, because orientation has relatively little effect on the observed properties of stars, but that is only because they are round.

In the "Unified Model," a brilliant accreting SMBH (quasar) is surrounded by a toroidal distribution of opaque dust clouds. One can see the quasar directly when the torus happens to present a polar view to the Earth, and one can often detect the quasar and study it in detail even when the orientation of the torus is equatorial by using light scattered (polarized) by gas along the torus axis. This is exactly like a periscope, and allows us effectively an "overhead" view of the hidden AGN. This is how Joe Miller and I robustly unified quasars with active objects not showing quasar properties directly.

Another great example of orientation-dependence is the superluminal (apparent faster-than-light) motions in the jets of many radio-loud galaxies and quasars. The superluminal speeds are now attributed to motions actually at nearly light speed, traveling roughly (but not exactly) towards us along our line of sight. This is a classical effect due to sequentially lower light travel time for the little cannonball components shooting out of the radio cores as they move closer to us, producing the appearance of faster than light motions perpendicular to the line of sight. These sources are picked up preferentially because of the beaming (headlight) effect that comes from Special Relativity. In fact, this beaming effect nicely explains the apparently one-sided jets: we are quite sure now that there do exist "counter-jets" so that both lobes are being energized, but the jet on the far side beams its radiation away from the line of sight!

A spectacular corollary of the beaming idea is that the much larger number of equivalent radio sources which must be present but not pointing in our direction (according to the Copernican Principle) must be none other than the normal giant double sources (mostly radio galaxies), which do not show superluminal motion as seen from Earth. Credit goes mostly to P Scheuer, R Blandford, Martin Rees, and Mitchell Begelman for the theoretical insight.

The early work is described in detail in Annu. Rev. Astron. Astrophys. 31: 473-521 (1993). There's a brief email interview about this paper at In-Cites.

The spectropolarimetric method seems to work in most cases, but it needn't work in every case, because the gas acting as the periscope mirror may be absent, or the scattered light may be absorbed on the way to Earth. And it is mostly qualitative, providing no accurate information on the luminosity of hidden thermal AGN.

There is a complementary method of detecting hidden AGN, and also quantifying their luminosity. The idea is that the hidden light source radiates onto the dusty torus, which is on much larger scales than the energetically dominant optical/UV radiation described above, is much cooler, and radiates in the mid-infrared part of the spectrum. Our group and others have detected this reprocessed infrared light, confirming the "Unified Model." Some of this work was carried out in collaboration with graduate student D Whysong (e.g. Whysong, D, and Antonucci, R 2004, Astrophys J 602, 116) and researcher Patrick Ogle (e.g. Ogle et al 2006, 2010) of the Jet Propulsion Laboratory.

Almost all radio-quiet AGN follow this paradigm to first order. The same can be said for all of the most luminous radio-loud objects. But many radio galaxies of more modest luminosity lack this type of "central engine." We know this because they lack the reprocessed infrared radiation expected from the torus.

In retrospect there were intimations of the two types of radio galaxy - hidden quasar or no hidden quasar - going back for decades. And there is now very strong supporting evidence for the dichotomy all over the electromagnetic spectrum (Antonucci 2012 Astron Astrophys Trans 27, 557). The radio galaxies that lack the optical/UV emission from an accretion flow are only detected by their synchrotron radiation, and are called nonthermal sources. Thus it is now established that "radio jets" with impressive kinetic luminosity (power of the bulk acceleration of radio jets, as well as production of relativistic electrons and magnetic field producing the observed synchrotron radiation) - with little or no accretion to provide power from the gravity of the black hole! Historically it had been posited that some radio galaxies could be powered not directly by gravity, but by an electromagnetic process that taps the rotational energy of a spinning black hole, likely some form of the "Blandford-Znajek" mechanism. That seems likely for the lower luminosity radio galaxies.

Optical/near-IR Spectropolarimetry of Quasars: Separating the "central engine" light from that of surrounding matter.

The optical/UV continuum is the energetically dominant result of real-time SMBH accretion. This emission is thought to be thermal radiation from heated accreting matter, but the theory remains in a primitive state. Existing models have shown virtually no predictive power, and in fact little "post-dictive" power. Calculations are very difficult (magnetohydrodynamics in the Kerr metric), and made more so by the fact that much of the observed spectrum from the deep gravitational potential well of the black hole ("where the money is") is heavily contaminated by emission from surrounding dust and gas. In that sense it has been almost impossible for theory to succeed: there is no real uncontaminated spectrum of the energy-producing black hole region. We have made a serious dent in this problem.

Our group has often used polarimetry as an astrophysical tool. As noted, scattered light can be studied via the "polarized light" spectrum, and this allows us to see things literally from other points of view, an unusual situation in astronomy.

Seeing astronomical objects from other points of view, and from other positions in space, can be profoundly valuable. In this section I describe the isolation of scattered light not from parsec scales as above, but on scales a millionth as large, perhaps from only hundreds of times the event horizon radius of the SMBH.

We have found several cases in which a bit of scattered light originates entirely from the "central engine". We can separate the wheat from the chaff quite accurately by studying the polarization as well as the wavelength of incoming photons in thermal objects. We have identified the first intrinsic spectral feature from quasars since their discovery in 1963, namely absorption in the Balmer continuum. This proves that the near-UV light originates from from an opaque thermal source with dissipation at large optical depth, as widely anticipated. The Balmer edge feature, and a strong continuum slope chance in the near-IR that we discovered subsequently (Kishimoto et al 2004, 2008), can seriously constraint theoretical models. For completeness I add that another spectral feature has been discovered in quasar spectra without the need for polarimetry. The key area around 1000 Å is pretty clear of contamination, and it's been found that many quasars have a change in spectral slope there, perhaps indirectly associated with Lyman continuum absorption.

A fourth project can be approached from the soft x-ray point of view or the optical point of view. There is major precedent in the former approach, pioneered by S Komossa, and recently, substantial precedent for the latter, especially by Arcavi et al. We take the optical approach, and our work (led by graduate student Eli Quetin) has a particular niche.

There is little direct evidence that black holes actually have gravitational potential wells which are as deep and steep as predicted by general relativity. (There are claims to the contrary from X-ray astronomers but I am deeply skeptical of them.) In these regions, relativistic effects be spectacularly manifest. Our project will not test the nature of spacetime in any detailed quantitative way. But it can test whether or not there is a sufficient and sufficiently concentrated gravitational force field to split a star. No other known astrophysical objects other than SMBHs can do this.

We seek to discover tidal disruptions of stars that pass very close to the putative supermassive black holes. According to theory, a star should be gravitationally scattered into a quiet black hole every ~10,000 years. This should produce an enormous (up to 1038 watts) temporary (months?) thermal electromagnetic display as the star is chewed up and digested by the black hole.

I emphasize that several astronomers, most notably S Komossa, have already provided good arguments for several individual tidal disruption events using X-ray observations. And recently some very good candidate events have been discovered and crucially followed in time with spectroscopy, especially by I Arcavi et al, based in the UCSB affiliated Las Cumbres Observatory.

Our plan is simple but arduous: we are interrogating the entire Hubble Space Telescope imaging archives to look for historical events of this nature. We expect to discover > 100 optical events, which would indicate the event rate and other demographic parameters with reasonable accuracy, and hopefully provide observational signatures which will help astronomers to discover such events in real time in the future.

Aside from this semi-quantitative though seemingly robust test of the existence of supermassive black holes, there is another purpose which is of great astronomical interest. People modeling quasars are dealing with a very messy process for which the "initial conditions" are unknown. The ideal experiment for theoretical understanding is to throw a single star into an isolated starved black hole, and watch the result. I call this getting the quasar "Green's Function." So far we have no results :--)

There are two ongoing technical breakthroughs in ground-based infrared astronomy infrared called adaptive optics and interferometry. Before these techniques (which have some limitations!), the best angular resolution attainable in the optical/UV range from the ground was ~1 arcsec (1/3600 of a degree). The main reason is that light from a point source no longer looks like a point source after passing through Earth's atmosphere. The light striking the various parts of a telescope mirror follows slightly different paths through the atmosphere. That means they suffer very different phase delays because of rapidly fluctuating changes in the air density (refractive indices). The light from a point source arrives not in the form of plane waves but as a crinkled wavefront, each part of which redirects incoming light at different angles on its way to the detector.

The main purpose of the Hubble Space Telescope was to get around this by observing from above Earth's atmosphere, and the results have been spectacular. The limiting factor on the Hubble's angular resolution is the unavoidable diffraction effect due to the Uncertainty Principle. The resolution in this case is given by the wavelength of observation divided by the telescope mirror diameter (2.4m for the Hubble); in that case, it comes out to be ~0.1 arcsec, a full 10x better than from the ground. If the big ground-based telescopes could avoid the fluctuating atmospheric diffraction (known to astronomers as seeing), they would see ~4x sharper than the Hubble at a given wavelength, because they are so much bigger (up to 10m).

Suppose we want to observe a galaxy from a big ground-based telescope. If there happens to be a bright star next to it on the sky, by measuring the exact shape of the wavefront from the star after it passes through the atmosphere (this must be done every ~30 millisec of time), we can correct the wavefront crinkling, because we measure in real time the wavefront from the star's light is intrinsically plane waves. The correction is then also made in real time using many mechanical actuators (pistons), constantly reshaping a "rubber mirror" in the optical path.

If the galaxy target is very close to a bright star in the sky, the effect of the atmosphere will be automatically corrected for the science target (galaxy) as well, and diffraction-limited resolution is achieved. Very few galaxies or other objects of interest happen to have bright stars very close by, but that problem has been solved: we use a sodium laser beam which can for example be sent through the telescope backwards. A bit of the laser light is resonantly scattered by sodium atoms in the upper atmosphere, effectively forming an artificial bright star anywhere it's desired!

I worked on some of the first papers using this method on AGN, with some of the largest telescopes in the world (the Kecks, 10m). These telescopes are jointly own by the University of California and Cal Tech, with NASA now contributing and getting some guaranteed observing time, mostly for exoplanet work. The major UC participation gives good access to these giants. This is a great attraction for students and researchers coming here!

Our early images and spectra of the Cygnus A radio galaxy revealed cones of light emitted by a hidden quasar, as well as a tiny companion galaxy close to the core, which does not show up on the blurrier Hubble picture (Canalizo et al 2003). My colleagues had the technical expertise, and deserve most of the credit. They recruited me because of my knowledge of AGN.

All this is by way of introducing the next great (another factor of 10!) leap forward in improved angular resolution. I'm now working with some colleagues (my past postdocs S Hoenig and M Kishimoto) using this spectacular gain. combining the light coherently from multiple giant telescopes spaced ~100m apart. This has been done routinely in the radio regime since the 1950s, but it has just become possible for the much more challenging near and mid infrared spectral ranges. (It can't yet be done in the optical region for faint objects like AGN.) The Hubble mirror diameter is 2.4m, and that of the Keck is 10m. But by using multiple telescopes, the angular resolution is given by the distance between the telescopes, 10x larger than the Keck mirror size, so another factor of 10 in resolutions is achieved! We use simultaneously both giant Keck telescopes in Hawaii, or three (and soon all) of the European Very Large Telescopes (VLT - there are four in total) in Chile. Each telescope pair acts as a double slit interferometer and measures a Fourier component of the spatial brightness distribution. Current images, while incredibly sharp, are very crude compared with the detailed images from single telescopes, but very recently we and others have made a major advance: we can now recover the phases of the interference patterns in real time, and not just their amplitudes. This was a first for faint objects like AGN, and it greatly increases the amount of information collected. (I personally have no expertise in this technology, and I'm just involved in the science end.)

It's a powerful indication of the value that the community places on this work that we are given the use of multiple giant telescopes simultaneously, when the competition for the use of single giant telescopes is already very stiff.

I'll just say a bit about what we've achieved and published so far - and it'll sound boastful. The reason this doesn't embarrass me is that I'm really flattering former postdocs M Kishimoto and S Hoenig, who contribute the lion's share of the very difficult observing, analysis, and of the interpretation.

The AGN obscuring torus has been described above, and it is universally accepted that it absorbs much of the primary optical/UV radiation from the thermal AGN. But our group and others are finding that the near-mid IR emission associated by us and others with the bulk of the thermal re-radiation comes instead, at least in some cases, mostly from along the AGN polar axis. We tentatively suggest that it's a broad spectral peak at 3 microns which comes directly from the torus. If confirmed, it is very likely that there is a dusty wind driven in the polar regions by radiation pressure, redirecting a large fraction of the matter flowing towards the black hole upward and outward. We would like to know how much of this gas escapes to infinity, and how much may perhaps be recycled back into the torus. Our group and others hope to understand this relatively new part of AGN physics in the coming years.