Astron. Astrophys. 358, 88-94 (2000)

4. Summary and discussion

Our original interest in the field of Q0151+048A was to identify the DLA galaxy in front of it. This identification was accomplished via imaging in Ly (Fynbo et al. 1999), but our broad band images left some questions open. The purpose of the deeper broad-band data presented in this paper was to clarify this situation. We shall here first summarise our findings, then briefly consider their implications.

4.1. Results summary

Our new data have unambiguously confirmed the presence of extended emission in the field in all three bands I, B and U. The different morphology seen in the three bands strongly suggest that we see three objects superimposed: The quasar, the DLA absorbing galaxy and the quasar host galaxy.

The superposition of three close objects of widely differing brightnesses causes considerable degeneracy for any attempt to determine the brightness of the faintest sources, and it is therefore impossible to find a unique solution for the flux of the faintest object (the DLA galaxy S4). Nevertheless, we find that S4 is clearly detected in the U image. The U-band magnitude of S4 determined via our minimum procedure is fully consistent (to within 1 ) with being caused by the known Ly flux at 3565Å alone. The data are therefore consistent with a zero contribution from any continuum source in the U-band.

It is difficult to determine the exact errors on the I and B magnitudes of S4, but for both images we found a very significant improvement in the reduced of the fit when we included S4. It is therefore likely that S4 is indeed a low surface brightness continuum source, but this question is going to be extremely hard to settle.

The existence of a separate extended continuum source centred on qA is, however, clearly demonstrated independently in all bands. This result was arrived at independently via image deconvolution, and via our iterative object fitting technique.

4.2. Discussion: Starburst galaxy or dust scattering

The distance modulus (for z=1.93) in the assumed cosmology with h=0.5 is 45.8. Assuming instead =0.3 and 0.7 the corresponding distance modulus becomes 46.6. Hence, the absolute AB magnitudes of the host galaxy HGa is -24.0(-24.8) in U (rest frame 1100-1300 Å), -24.5(-25.3) in B (rest frame 1300-1500 Å) and -24.0(-24.8) in I (rest frame 2300-3400 Å). Such extremely bright magnitudes are in the local universe only connected with brightest cluster galaxies (for comparison M87 and Centaurus A both have absolute magnitudes of roughly -23 in the V-band). Brightest cluster members can be as bright as -26 (Oemler 1976). Interestingly we find that the absolute magnitude of HGa is similar to those of the extended `fuzz' that have been detected around other high redshift QSOs by Lehnert et al. (1992), Carballo et al. (1998) and Aretxaga et al. (1998a,b).

The morphology of the host galaxy HGa is best fit by a de Vaucouleurs profile. The fit to an exponential-disc leads to a much poorer fit. A plausible interpretation of the data is therefore that we see the early stage of a massive elliptical galaxy in the process of forming the bulk of its stars. Assuming that all the light is coming from stars, and not e.g. scattered quasar light (see below), we can estimate the star formation rate (SFR) needed to explain the observed fluxes. In the case of continuous star formation we can adopt the relation between the SFR and the luminosity at 1500 Å SFR = erg s^-1 Å^{- 1}) commonly used for LBGs (Pettini et al. 1998) and we hence infer a star formation rate of order 100(200) yr^-1 for =1(0.3) and =0(0.7). For instantaneous bursts we can use the Starburst99 package (Leitherer et al. 1999) to infer the colours of models calculated with solar metallicity and ages 1, 10 and 100 million years. The colours for these three models are given in Table 4.

Table 4. The colours of instantaneous starbursts with four different ages. The colours of HGa are listed for comparison.

The colours of the host, 0u-B1.1, -1.1B-I0.1 from Deconvolution and 0.90.3, -0.70.3 from PSF-subtraction, are roughly consistent with instantaneous bursts with ages in the range 10-100 Myr. The number of stars formed in the burst would be in the range from 10⁸ to a few times 10⁹ stars depending on the age of the burst and on the assumed cosmology.

Another interpretation of the extended fuzz frequently seen around quasars, is light from the quasar itself scattered by dust. This mechanism is well known from radio galaxies at high redshifts where scattering off dust grains has revealed the existence of "hidden" quasars in the galaxy cores. It is likely that radio quiet QSOs have similar non-isotropic radiation fields (see e.g. the discussion in Moller & Kjærgaard 1992), and in that case our line of sight is such that we look straight down the emission cone inside of which the scattering is taking place. In this case we therefore expect to see the quasar emission cone "end on" via forward scattered quasar light. The scattering process is expected to be essentially grey and recent calculations predict that as much as 10% of the quasar light could be scattered in this way (Witt & Gordon 1999; Városi & Dwek 1999; Vernet et al. in prep.). If considering a clumpy medium, we would expect dust scattered light to be emitted from inside a very large volume in front of the quasar. When taking the cone geometry into account one would expect its total flux to be roughly a few% of the quasar flux at any given wavelength (Fosbury, private communication). From Table 3 we find that the flux from HGa is 3, 6 and 2% of the flux from Q0151+048A in U, B and I respectively. Similar, but less significant, results are found for HGb. It is not yet known if the light profile of scattered light from a cone will reproduce a de Vaucouleurs profile, but since this seems to be a universally preferred profile it is not unlikely. One thing worth noting in Fig. 3e are the negative residuals surrounding the position of qA at a distance of 2-3 arcsec after subtraction of the fitted de Vaucouleurs profile. This indicates that the true profile of HGa in reality falls off steeper than a de Vaucouleurs profile. If model calculations were to show such a steep profile for forward scattered light in a radiation cone, that would be a strong hint towards the nature of the quasar fuzz.

However, the colours of the extended emission as seen in Table 3 and Fig. 4 are significantly different from those of the two QSOs, which argues against the scattering hypothesis. Hence, we conclude that at least a significant fraction of the observed extended emission must be caused by a star burst.

Fig. 4. The B(AB)-I(AB) vs. u(AB)-B(AB) colours of qA and qB (marked by ), of the instantaneous star burst models of Table 4 (thick, full drawn line). The dotted, dashed and long dashed lines show the colours of reddened bursts assuming A()=0.5 and MW, LMC and SMC extinction curves respectively. The colours of the extended emission under qA is marked by the error-bars. The dashed error-bar represents the measurement from deconvolution and the full drawn error-bar the measurement from PSF-subtraction.

© European Southern Observatory (ESO) 2000

Online publication: June 26, 2000
helpdesk.link@springer.de