Angular correlation functions of X-ray point-like sources in the full exposure XMM-LSS field

<BR /> Aims: Our aim is to study the large-scale structure of different types of AGN using the medium-deep XMM-LSS survey. <BR /> Methods: We measure the two-point angular correlation function of 5700 and 2500 X-ray point-like sources over the 11 sq. deg. XMM-LSS field in the soft (0.5-2 keV) and hard (2-10 keV) bands. For the conversion from the angular to the spatial correlation function we used the Limber integral equation and the luminosity-dependent density evolution model of the AGN X-ray luminosity function. <BR /> Results: We have found significant angular correlations with the power-law parameters γ = 1.81 ± 0.02, θ[SUB]0[/SUB] = 1.3'' ± 0.2'' for the soft, and γ = 2.00 ± 0.04, θ[SUB]0[/SUB] = 7.3'' ± 1.0'' for the hard bands. The amplitude of the correlation function w(θ) is higher in the hard than in the soft band for f[SUB]x[/SUB] ≲ 10[SUP]-14[/SUP] erg s[SUP]-1[/SUP] cm[SUP]-2[/SUP] and lower above this flux limit. We confirm that the clustering strength θ[SUB]0[/SUB] grows with the flux limit of the sample, a trend which is also present in the amplitude of the spatial correlation function, but only for the soft band. In the hard band, it remains almost constant with r[SUB]0[/SUB] ≃ 10h[SUP]-1[/SUP] Mpc, irrespective of the flux limit. Our analysis of AGN subsamples with different hardness ratios shows that the sources with a hard-spectrum are more clustered than soft-spectrum ones. This result may be a hint that the two main types of AGN populate different environments. Finally, we find that our clustering results correspond to an X-ray selected AGN bias factor of 2.5 for the soft band sources (at a median <inline-formula>bar{z} ≃ 1.1) and 3.3 for the hard band sources (at a median <inline-formula>bar{z} ≃ 1), which translates into a host dark matter halo mass of 10[SUP]13[/SUP]h[SUP]-1[/SUP]M[SUB]&sun;[/SUB] and 10[SUP]13.7[/SUP]h[SUP]-1[/SUP]M[SUB]&sun;[/SUB] for the soft and hard bands, respectively. This paper is dedicated to the memory of Olivier Garcet who has initiated the present work just before his sudden death.