4 Summary and conclusions

We have obtained near infrared imaging for a sample of fifteen QSOs, most of them selected as having extended ionized nebulosities, in order to analyze the possible connection of this gas with the underlying stellar population. We detect infrared extensions in at least eleven of them. The ionized gas extensions obtained from spectroscopic studies are within the extensions that we determine from NIR imaging. When images of the ionized gas are available, their extensions are compatible with the gas belonging to the host galaxies. The kinematical behaviour of this gas should be analyzed in order to be able to put some constraints on the possible mechanisms for the gas origin and fuelling.

We have determined the relative contribution of the host galaxies by subtracting the scaled PSF to the quasar profile. The corresponding magnitudes are given in Table 3. For the whole set of 15 objects, the resulting average contribution is $50 \,\pm\, 20\%$ in J ($48 \,\pm\, 22\%$for the twelve radio loud quasars and $58 \,\pm\, 7\%$ for the three radio quiet). We could extract the host galaxy for only four quasars in K', for which the host galaxy contributed $67 \,\pm\, 8\%$ in average to the total luminosity. For the four hosts extracted in both J and K', we find that their contributions are similar in both bands for IRAS 09149-6206 (radio quiet) and larger in K' than in J for the remaining three (radio loud), in agreement with the expectation that K' describes the stellar population better than J. The average absolute magnitudes of the host galaxies are $M_{J} = -25.5 \,\pm \,1.1$ ($-25.7 \,\pm\, 1.2$ for the radio loud and $-24.9 \,\pm\,0.5$ for the radio quiet) and $M_{ K'} = -27.0 \,\pm\, 0.6$.

This is in agrement with previous results on radio loud quasars in other bands (Véron-Cetty & Woltjer 1990; Taylor et al. 1996; Bahcall et al. 1997; Hooper et al. 1997, assuming typical values V-H=3.0, R-H=2.5, H-K=0.2 and J-H=0.8). J and K' absolute magnitudes are also compatible with those of Bright Cluster Member galaxies by Thuan & Puschell (1989) and Aragón-Salamanca et al. (1998).

Given our limited spatial resolution, we cannot always discriminate whether asymmetric extensions are intrinsic or are the consequence of the presence of objects in projection close to the line of sight to the quasar. We have attempted to answer this question first by estimating the shape of the PSF from starlike objects in the field, deconvolving the images and then subtracting the PSF of the deconvolved image to the quasar infrared profile, and second by using high resolution optical images retrieved from the HST archives, which were available for eight of the observed QSOs.

It was possible to resolve the extensions into close objects for PKS 0812+020, PKS 0837-120 and 3C 281, though for the second quasar the elongated structure seems also to be present in the HST images, suggesting that the underlying galaxy is indeed detected. In all the other cases the infrared extensions are symmetric, as expected for a normal underlying host galaxy; however, the corresponding average profiles are notably different from the PSF profiles in only four QSOs: IRAS 09149-6206, PKS 1302-102, Mrk 877 and MC 1745+163 (see Fig. 24), where the harboring galaxies are well fit by the r^1/4 profile describing elliptical galaxies. Out of these four quasars, note that PKS 1302-102 and MC 1745+163 are radio loud, while IRAS 09149-6206 and Mrk 877 are radio quiet.

Note that although HST images were needed to resolve close objects, they have proven to be sometimes less efficient to detect low surface brightness features, as for PKS 1302-102 (Bahcall et al. 1995, see Sect. 3.8). For 3C 215, the HST image does not allow to trace the elongation due to the host galaxy, whereas our J image does. The present study confirms that infrared imaging is well adapted to detect underlying galaxies in quasars, after objects located close to the quasar are identified with high spatial resolution imaging (HST or adaptive optics ground based data).

One of the possibilities to explain the presence of extended gaseous envelopes around QSOs is to claim their belonging to rich environments, with the presence of close companions producing the physical mechanisms (tidal forces or nonaxisymmetrical perturbations in accretion or merger events) that can account for gas fuelling. A number of redshifts have been measured for the objects in the vicinity of five of the QSOs we have observed: PKS 0812+020, PKS 0837-120, 3C 215, 3C 275.1 and 4C 11.50. Note that out of these five quasars four are known to be rich in ionized gas, and the only one for which ionized gas has not been reported is PKS 0837-120, which belongs to a richness class 1 cluster. For all five quasars, between one and five galaxies were found to have redshifts similar to that of the quasar. The presence of such a large number of companion objects is therefore likely to imply the existence of environmental effects which could account for the existence of extended ionized nebulosities in most of these objects.

Unfortunately, we do not have redshift information for field objects around the ten other QSOs in our sample. However, our infrared imaging shows that at least in projection most of these quasars have a number of galaxies in their environments. 3C 334 has at least one close galaxy, and some of the objects close to PKS 1011-282 and 4C 20.33 seem to be galaxies as well, while the fields around IRAS 09149-6206 and MC 1745+163 are crowded with mostly star like objets.

In order to improve such a study of the physical parameters of the underlying galaxies of quasars, and to describe their environments, high spatial resolution deeper infrared images of larger fields and with a higher sensitivity are obviously required (in particular, our limiting magnitudes together with the small frame sizes we are dealing with - less than 1' - result in small object numbers that prevent a reliable statistical analysis of colors and number counts of companions, such as that performed by Hall & Green 1998). After the neighbouring objects are detected and galaxies are separated from stars, spectroscopy is necessary to determine which galaxies really belong to the quasar environment.

Acknowledgements

I. Márquez acknowledges financial support from the Spanish Ministerio de Educación y Ciencia. I.M. acknowledges technical support from the IRAC team at ESO, specially Luis Ramirez. We are very grateful to Ron Probst, who made available the SQIID package for the reduction of infrared images within IRAF available to us. Finally, we thank the anonymous referee for useful comments.