Research
AGN Feedback
Powerful jets and winds from AGN have a strong impact on the interstellar medium of the host radio galaxy. The deposition of jet mechanical energy and transfer of momentum affects the formation of stars in the host galaxy and the accretion of matter down the gravitational potential onto the central supermassive black hole. AGN jets and winds may, therefore, be responsible for establishing the observed correlation between the mass of the black hole with the mass and the velocity dispersion of the host galaxy's bulge. We test these propositions with high resolution 3D hydrodynamical simulations. This work is done in collaboration with Geoff Bicknell, Masayuki Umemura, Ralph Sutherland, and Joseph Silk.
Galaxy formation and AGN
Over cosmic time, galaxies grow by accreting gas, forming stars, and merging together to form bigger galaxies. Meanwhile, at the centers of all galaxies reside supermassive black holes (SMBH) that also grow by accreting gas and merging (after a galaxy merger).
When black holes accrete rapidly they efficiently turn the gravitational energy of accreted material into radiation, jets, and winds. The total power released exceeds that emitted by all the stars in the galaxy, and the galaxy is said to harbor an "active galactic nucleus" (AGN). Two types of galaxies harboring AGN are quasars, in which the radiation from the AGN dominates, and radio galaxies, in which the jet dominates. AGN disc winds have never been directly observed, but are inferred from absorption of the AGN radiation by fast outflows.
The re-deposition of energy and momentum into the interstellar medium of a galaxy through the outflows and radiation from the AGN is called "AGN feedback".
Problems in galaxy formation theory
Because bigger galaxies and bigger black holes are expected to form later in the history of the unverse, this paradigm is referred to "hierarchical galaxy formation". While this simple picture fits nicely into the framework of the currently accepted cosmological model, a closer look at galaxy populations reveals trends that require more complicated physics is involved.
One famous observation, as yet eluding unambiguous explanation, is the correlation of the mass of the central SMBH and the mass of the bulge of the host galaxy. Silk & Rees (1998) first proposed and showed analytically that an AGN wind can naturally produce this correlation. One aim of my work was to test this with realistic (multiphase) 3D hydrodynamic simulations.
Another puzzle is that when one looks at the ages of stellar populations in galaxies, it appears that more massive galaxies, which are meant to form later, had their stars in place much earlier than less massive galaxies. The same goes for black-holes - more massive black holes were in place in massive galaxies earlier in time than less massive black holes. It is believed that AGN feedback may play in regulating the star-formation history of galaxies resulting in "anti-hierarchical" galaxy formation.
Simulations of AGN jet feedback
In order to investigate the effects of powerful AGN jets on the halo of galaxies, we ran almost 30 high-resolution 3D relativistic-hydrodynamic simulations of AGN jets interacting with a two-phase interstellar medium. We were particularly interested in the effects of the AGN jet on the cold phase, the dense clouds in which stars were formed. The simulations were performed with the hydrodynamic code FLASH.
One of the first results to emerge from the simulations is that the presence of dense clumpy clouds in the halo significantly changes the evolution of the radio sources (Wagner & Bicknell 2011, pdf | NASA ADS). Compared to the freely expanding jet, the jet propagating through a clumpy medium is strongly hampered by the dense clouds and floods and channels in an almost diffusive process through the interstellar medium until the main jet stream breaks out of the clumpy region. Although the jet is very powerful, it is incredibly light (8 orders of magnitude lighter than the clouds) so it is easily deflected. An almost spherical bubble filled with high-pressured and fast-streaming jet plasma engulfs the entire galaxy.
The strong interaction of the jet with the interstellar medium leads to a rapid dispersal of clouds, reducing the amount of material available for star-formation. The outflows of neutral and ionized gas generated can explain outflows in high-redshift (Wagner & Bicknell 2011, Mahony et al 2013, Nesvadba et al 2008) and low-redshift (Lehnert et al 2011, Dasyra & Combes 2012) radio-galaxies. The cloud cores, are, however compressed, and the enhanced density in the cloud cores may lead to increased star-formation within. The competition of these two effects determines whether feedback is negative (suppression of star-formation) or positive (induced star-formation).
With our 30 odd simulations, we determined that the sign and strength of the feedback strongly depends on the internal structure and spatial distribution of clouds (Wagner, Bickenll, & Umemura 2012, pdf | NASA ADS). In general, it is difficult to estimate accurately which of the two effects dominates because hydrodynamic ablation and conditions for star-formation are difficult to capture in simulations.
Encouragingly, however, the clouds get accelerated to velocites comparable to those observed in high and low-redshift radio galaxies. Moreover, the dependence of the velocities to which the clouds are dispersed on the power of the AGN jet (which relates directly to the mass of the SMBH), is in agreement with the M-sigma relation, and with the predictions by Silk & Rees (1998).
Simulations of feedback by AGN winds
If AGN winds are powerful enough, they may have a similar feedback effect as AGN jets do. We performed several simulations of the interactions of an AGN wind and an inhomogeneous two-phase interstellar medium to verify this (Wagner, Umemura, & Bicknell 2013, pdf | NASA ADS).
Some of the most powerful winds known to exist are Ultrafast Outflows (UFO, Tombesi et al 2010), which may be as common as in 40% of all AGN. Compared to jets, winds are heavier (carry more momentum) and slower, and have a wider opening angle, and are in general less powerful. Even UFOs are on average an order of magnitude less powerful than jets.
For this study, we show three simulations which highlight the importance of the two-phase gas distribution. They differ in the distribution of dense gas in the halo. In one simulation, there is no dense gas. In the second, there dense clouds are distributed spherically (in the shape of a bulge of a galaxy). In the third case, the clouds are confined to a disc. The simulations were performed with the magnetohydrodynamic code PLUTO.
Naturally, the evolution of the simulations, and the strength and sign of feedback depend strongly on the distribution of the clouds. The most informative comparison is that between the spherical and disc-like distribution of gas. In the former, negative feedback is very effective (although weaker than in the case of a jet with similar power). In the latter, very little dense gas is blown out or dispersed, and instead the energy bubble developing in the halo of the galaxy engulfs and compresses the entire disc, inducing star-formation within.
These two cases represent extremes in the efficiency of negative or positive feedback. The spherical distribution of dense gas maximises the efficiency for negative feedback, and the disc-like distribution of gas is optimal for positive feedback.
For both AGN-jet and AGN-wind driven feedback, it will be informative to make closer comparisons of with individual galaxies, and galaxy samples, especially those observed with ALMA, as this informs us of the dense gas distribution in galaxies, on which the feedback efficiency depends so sensitively.