The standard model for the origin of the extreme luminosity of quasars considers that a supermassive black hole shines as a quasar when intense mass inflow takes place, possibly as a consequence of tidal forces in dissipative events (e.g., Di Matteo et al., 2005). In this scenario, the circum-galactic medium of quasar host galaxies is expected to be populated by streams, cool gas clouds and tidal debris, as commonly observed in interacting galaxies (e.g., Sulentic et al., 2001; Cortese et al., 2006). Moreover the gas of the circum-galactic medium could be metal enriched by supernova-driven winds triggered by starbursts events associated to the mergers or by quasar-driven outflows of gas (e.g., Steidel et al., 2010; Shen and Ménard, 2012).

One of the effective ways to study the circum-galactic medium of galaxies at high redshift is to investigate the absorption features that they imprint in the spectra of quasars. In particular, projected quasar pairs (Figure 1 left) are ideal observational tools for this purpose, since the light of the very bright source in the background (z ≡ z_B) goes through the extended halo of the foreground (z ≡ z_F < z_B) object (e.g., Hennawi et al., 2006; Farina et al., 2013). This can be evidenced by absorption lines at the foreground redshift: an example is reported in Figure 1 right.

In our previous works we studied 49 quasar pairs (Farina et al., 2013, 2014; Landoni et al., 2016). We used the optical spectroscopy of close pairs (projected distance ≤ 200 kpc) obtained at the ESO-VLT and Gran Telescopio Canarias (GTC) to investigate the distribution of the absorbing gas at 100-200 kpc projected distance from the quasar studying the presence of Mg II or C IV absorption lines at the redshift of foreground quasar. In order to characterize the structure of circum-galactic medium of the foreground quasar host galaxy we estimated the covering fraction of Mg II or C IV as a function of the projected distance. We assumed a threshold in equivalent width of 0.3, then we subdivided the projected distance in bins. For each bin we computed the covering fraction as the ratio between the number of systems with Mg II or C IV absorption lines greater than the threshold and the total number of observed systems. Our previous results (Farina et al., 2013, 2014; Landoni et al., 2016) indicate that 22 spectra exhibit absorption lines of foreground quasar in the background quasar: 15 Mg II and 7 C IV. We found that the detected Mg II absorption systems of EW >0.3 Å associated with the foreground quasars are revealed up to ~200 kpc from the centre of the host galaxy. The structure of absorbing gas is patchy with covering fraction of the gas that quickly decreases beyond 100 kpc. This is illustrated in Figure 2 left. A similar behavior is present in absorption systems with C IV doublet (Figure 2 right).

In this paper we analyze optical and NIR images of foreground quasars in order to investigated their closed environments and their host galaxies.

The selection procedure of our quasar pair projected sample is reported in detail in Farina et al. (2013) and Farina et al. (2014). Here we summarize briefly the main criteria of our choice:

We acquired optical images at FORS2@VLT with the I BESS filter for the object with z ≤ 1 and the Z GUNN filter otherwise, and near-IR images at HAWK-I@VLT with H filter for a total of 24 quasar projected pairs (4 targets have only optical images, 8 ones have only NIR images and 12 objects have images in both bands.). With this configuration we explore galaxies at the redshift of the target in the B band in the rest frame. For a subsample of object reported in Table 1 we present here 8 deep high quality I-band images (seeing ~0.5 arcsec).

We performed the analysis of optical images in order to characterize the close environment of foreground quasar. We used the software SEXTRACTOR (Bertin and Arnouts, 1996) to identify all objects in the frame over a given magnitude limit and to distinguish galaxy-like objects from star-like ones (galaxies have CLASS_STAR < 0.5 and stars have CLASS_STAR >0.5). Then we evaluated the overdensity of galaxies around the foreground quasar calculating the ratio of number of galaxies per arcminute square to background estimated at distances greater than 500 kpc.

The near-IR images have been analyzed using the software package AIDA (Astronomical Image Decomposition Analysis, Uslenghi and Falomo, 2008). From this analysis of the near-IR images we are able to characterize the properties of the foreground quasars host galaxy via 2-d model fitting, assuming that they are the result of the superposition of two components: the nucleus, described by the local PSF, and the host galaxy, modeled by a Sersic function convolved with the proper PSF.

Based on our deep optical images of quasars we are able to characterize the galaxy environment up to I ~ 23.5 which is more than 2 magnitudes deeper than SDSS images (see Figure 3 left). This allows to investigate the galaxy environment down to about 3 magnitude fainter than M^*. We find that for 4 cases there is a clear galaxy overdensity around the foreground quasar while in the another 4 cases there is no evidence that quasars live in a group of galaxies (see Figure 3 and Table 1). In 3 quasars that exhibit overdensity there is also a detection of Mg II absorption systems at the same redshift of foreground quasar in the circum-galactic medium. In the cases of no galaxy overdensity Mg II absorption lines are detected in 2 objects. The small sample, investigated till now, does not permit us to draw firm conclusions on the relationship between galaxy environments and presence of cold gas in the intergalactic medium. We are completing the analysis of the full sample and extending it with other targets from ongoing observations at GTC.

Till now we analyzed 11 out of the 20 foreground quasars with NIR images. For 9 (~80%) the host galaxy is well resolved. In one case the detection of the host galaxy is marginal and only for one image no evidence of the host galaxy is found. The rest frame absolute magnitude in I band of the resolved host galaxies ranges from −22.5 to −24.9, with a median value of −23.7. These results are comparable to those reference to other quasars at similar redshift (Sanghvi et al., 2014).

RO provides the analysis of the environment. MU provides the analysis of NIR images. All other authors provide discussion of results and write the paper.