Searching for observational evidence of QSO feedback

Our standard models of galaxy formation and evolution appear to be incomplete in detail. We predict only small galaxies to exist at early times in the history of the Universe, but we observe many massive galaxies. And we predict many more massive galaxies at the present day than we observe. "Feedback" from supermassive black holes (SMBH) has been suggested as one mechanism by which these problems, amongst others, may be solved. Simulations have shown great success (Figure 1), but observational evidence for SMBH driven outflows driving gas out of galaxies has proved ellusive. Scientists at Max Planck Institute for Astrophysics (MPA) have been studying absorption lines in the spectra of QSOs, bright objects powered by SMBHs, at redshifts of 1 to 2 to try to uncover observational evidence for ubiquitous loss of gas from the host galaxies.

Fig. 1: From Li, Hernquist, Robertson et al., 2007, ApJ 665:187. Simulations of QSO feedback, which removes all the gas from the host galaxy of the QSO preventing subsequent star formation and black hole growth, are successful in reproducing many observational results. But can we find direct observational evidence for such efficient and ubiquitous feedback? Credit: Pierre Kervella (Paris-Meudon Observatory) and European Southern Observatory

Fig. 2: The number of MgII absorption line systems observed around QSOs, as a function of distance from the QSO, normalised by the number of absorbers that would be expected if absorbers did not cluster around QSOs. We observe the same clustering signal seen in galaxy-galaxy clustering, with an amplitude consistent with absorbers originating in typical massive galaxies.

Fig. 3: The number of CIV absorption line systems observed, along the line-of-sight from a QSO (histogram). The central peak could be explained by galaxy clustering, but not the high velocity tail. Overplotted in black is our best-fit model which suggests that 40% of CIV absorbers observed to be within 170Mpc of the QSO are actually intrinsic to the QSO and its host galaxy.

Absorption lines in QSO spectra have been studied since QSOs were first observed with spectrographs. They appear as black regions against the background light of the QSO, and their observed wavelengths tell us they are caused by chemical elements situated between the QSO and the Earth. In general these elements lie in ordinary gas clouds: the denser and more metal rich clouds lie close to, or within, galaxies; and the more diffuse, metal poor clouds lie alone in "voids", regions with no galaxies. But it has always been a possibility that some of the absorption lines may be directly associated with the QSO itself - they may be the evidence we have been searching for that the QSO is expelling the gas from the galaxy in which it lives.

Indeed, we do observe an increase in the number of these absorbers close to the QSO. However, because absorbers are also caused by ordinary galaxies, and galaxies tend to live together in groups, this peak may simply be explained by ordinary galaxies in the QSOs neighbourhood and thus be of no relevance to the search for outflows. A simple test, performed by Vivienne Wild and her collaborators, is to measure directly the clustering of absorbers around QSOs, by looking for absorbers next to the QSO i.e. using background QSOs as probes of the environment around foreground QSOs. The result is shown in Figure 2 for singly ionised Magnesium (MgII): we detect the clustering of absorbers around QSOs and show that it follows exactly the same powerlaw as for ordinary galaxy-galaxy clustering. However, there is a surprise when we translate the MgII clustering signal into a prediction for the number of absorbers along the line-of-sight between the QSO to us. We massively overpredict the number of absorbers: we can conclude that the ionising radiation of the QSO affects the gas clouds in and/or around most nearby galaxies, ionising it beyond the levels normally experienced by such clouds.

Therefore, we must look at absorption lines with a higher ionisation state to search for outflowing gas. Performing the same experiment for triply ionised Carbon atoms (CIV), we see a very different picture (Figure 3): an excess of absorbers which can only be caused by gas flowing towards us fast from the QSO. We find that around 40% of absorbers close to the QSO are infact truly associated with the QSO, with a range of velocities from a few hundred to many thousands of kilometers per second.

Are these absorbers the evidence we have been searching for, that QSOs expel a large fraction of the gas from the galaxies in which they live? Unfortunately, it has become clear that the QSO has a considerable ionising effect on the gas clouds of its own galaxy. Combining this unknown ionisation state with the unknown distance of the clouds from the QSO, it is impossible to measure the mass of the clouds which are being expelled at such speeds. In the near future, we hope that further constraints will be placed on the true nature of these clouds by measuring the strengths and widths of other absorption features from other elements that are present in the QSO spectra.


Vivienne Wild, Guinevere Kauffmann, Simon White


Publications

Vivienne Wild, Guinevere Kauffmann, Simon White et al., "Narrow associated QSO absorbers: clustering, outflows and the line-of-sight proximity effect", 2008, submitted to MNRAS, linkPfeilExtern.gifarXiv:0802.4100