Introduction

Nature of the X-ray population

The X-ray observations of the SXDS comprise the widest contiguous area ever observed in hard X-rays to a depth of ~10^-15 erg cm^-2 s^-1. The XMM-Newton observations comprise a single 100-ksec pointing and six 50-ksec pointings which flank the central field. The 1.3 square degree area is 10 times larger than Chandra surveys with similar depth. Based on the data we have so far obtained, we predict there will be in excess of 1,000 X-ray sources. They therefore represent a key resource in the efforts to fully understand the origin and nature of the Cosmic X-ray Background. Currently, ~70% of the soft (0.5-2 keV) X-ray background and ~60% of the hard X-ray background is resolved. Based on the optical follow-up observations of X-ray sources, the Cosmic X-ray Background is thought to be the integrated emission from AGNs with various redshifts, luminosity and types: thus, resolving the origin of the Cosmic X-ray Background is equal to revealing the history of accretion on to the massive black holes in the Universe. Models of the X-ray background predict that a significant fraction of such accretion occurs in absorbed AGNs, and the X-ray spectroscopy such accretion occurs in absorbed AGNs, and the X-ray spectroscopy needed to quantitatively examine the absorption in such objects can only come from XMM-Newton. (It should be noted that XMM-Newton has 2-4 times the photon collecting power of Chandra.)

The number density of luminous QSOs in the very high-redshift Universe (z > 5) is also an important issue for deep X-ray surveys. The number of such QSOs in SXDS will provide information about the formation epoch and process of massive objects in the early Universe. Models of quasar evolution predict the existence of ~100 X-ray quasars with z >5 in the SXDS Field (see Figure, adapted from Haiman & Loeb 1999). These objects will be identified based on their unusual colors (invisible in B, red in R-I due to the Lyman forest, blue in I-z' due to Ly-alpha emission in the I-band) and are readily accessible to FOCAS spectroscopy. Although the Sloan Digital Sky Survey has already identified a number of z >5 quasars, they represent only the very brightest tip of the quasar luminosity function -- our observations will probe approximately two orders of magnitude fainter. Haiman & Loeb's model predicts about 10 quasars with z >7 which will have z' <26, and therefore be easily detectable in this band and our deep infrared observations. Unfortunately, Ly-alpha is redshifted beyond the reach of FOCAS, but infrared spectroscopy (e.g., with OHS) will reveal their characteristic strong emission lines and hence provide a redshift.

Since the point spread function of XMM-Newton is larger than that of Chandra, the positional uncertainty of each source could encompass a number of possible identifications. However, high-resolution multi-color data can provide significant help in determining which object within the error circle is the correct ID, and our preliminary FOCAS spectroscopic observations indicate that the percentage of misidentifications will be low.

From our spectroscopy, we can obtain an estimate of the ratio of Type 1 to Type 2 (i.e., broad-line to narrow-line) AGN to compare with models for the X-ray background (e.g., Comastri et al. 1995). While luminous Type 2 AGN should substantially outnumber Type 1 AGN in such models, their optical appearance as normal galaxies makes them virtually impossible to find except through X-ray observations. The spectral information provided by XMM-Newton will allow us to determine which sources contribute to the X-ray background at both soft and hard energies and study the properties of their optical spectra; for example, do the hard, absorbed sources display emission lines or could their entire emission-line regions be obscured, as suggested by Fabian (1999).