Multi-wavelength key areas

The next decade will see the completion and exploitation of a number of new facilities capable of mapping the distant Universe in a wide range of wavebands spanning the X-ray to radio. XMM-Newton, GALEX, SIRTF, Astro-F, Herschel, ALMA, SCUBA-2 and eVLA/eMERLIN all offer the capability to survey tens of square degrees and detect populations of sources at median redshifts in excess of one. The first three of these missions are undertaking or planning surveys of 10-100 sq. degs in complementary regions. To exploit fully these multi-wavelength surveys deep near-ir imaging data are required to maximise the identification of objects that are heavily obscured at optical wavelengths. These include EROs (i.e. a population defined by their optical to near-ir colours) as well as other objects identified at sub-mm/far-ir wavelengths or in hard X-ray surveys. Near-ir imaging is also essential for normal galaxies, as it defines the rest-frame optical SEDs at z>1. We have an opportunity to take the lead in the definition of the next generation of ''Selected Areas'' (or ''standard survey areas'') by knitting together the XMM, GALEX, and SIRTF missions with deep near-ir imaging over tens of square degrees.

These multiwavelength observations will provide the possibility of a comprehensive analysis over all wavelengths of the relative contributions to the cosmic energy budget of stars and AGN at z>1, for comparison against similar measurements in the local Universe. These two populations are considered individually below.

Obscured starbursts observed by SIRTF in the far-ir: the wavelength dependence of the cosmic energy budget due to star formation

The three target fields listed in Table 5.1 (section 5.5) will all be covered in the SIRTF passbands, from 3.6 to 160 micron, by the SWIRE consortium, and will also be observed in the ultraviolet (1350 to 3000Å) by GALEX. The DXS J and K imaging will be particularly useful for identifying highly reddened dusty starbursts which will be undetectable at optical wavelengths. In Figure 5.3 we present calculations by Xu (2001) showing the expect K magnitude distributions for populations selected in the different SIRTF passbands, for a range of redshift slices. The depth required to detect nearly all the SIRTF sources is established by this calculation to be K=21.

Figure 5.3. This plot shows a calculation of the expected K-band magnitude distribution of sources detected in the SWIRE survey, using SIRTF, based on the models of Xu et al. (2001). SIRTF has 7 bands in the mid to far-ir, from 3.6 to 160 micron. For clarity the calculations for only 4 bands are shown. For each band the plot shows the predicted cumulative K-band magnitude distributions for the sources detected in each of several redshift slices. This indicates that essentially all sources are detectable in a survey reaching K=21.

The multiwavelength coverage provided by GALEX, SIRTF, the DXS, and suitable optical data (Table 5.1) will allow a comprehensive study at z>1 of the contribution from star formation to the cosmic energy output over the electromagnetic spectrum. The wide area coverage allows the calculation of the correlation function of galaxies of different star formation histories. Taken together the goal can be thought of as mapping the star formation of galaxies in the Universe in 4D i.e. spatial and temporal.

Over the planned 35 sq. degs the anticipated numbers of sources detected by SWIRE are 5X10⁴ starbursts, 5X10⁵ ellipticals, 15,000 classical AGN, and 100,000 obscured AGN, with fractions ~25% at z>2. The near-ir data will be particularly useful for identifying reddened sources not detected at optical wavelengths. These will be predominantly starbursts. In addition the near-ir data anchors the SED of sources at these wavelengths, to establish photometric redshifts. Finally, for sources that are not strongly reddened, the K-band light provides a measure of the stellar mass. This is the integral of the star formation up to that redshift, and therefore provides an additional constraint in determining the history of star formation of the different populations.

Obscured AGN observed by XMM and by SIRTF in the mid-ir: the contribution of AGN to the cosmic energy budget

Recent results (eg Mushotzky et al. 2000) suggest that the majority of the X-ray Background (XRB) arises from active galaxies (AGN), obscured by various amounts of dust. Dust reddens the spectrum leading to detections only in the infrared, with no optical counterparts to faint limits (eg Newsam et al. 1997; Lehmann et al. 1999). To confirm this model, to quantify the amount of obscuring dust, and to determine the possible evolution of dust content with redshift (eg Bechtold et al. 1994), we require near IR observations of very deep XMM and Chandra X-ray fields.

The far-ir background is dominated by emission from dusty galaxies. Many such galaxies show signs of strong star formation, which may heat the dust, but the dust may alternatively be heated by an obscured AGN (eg Genzel et al. 1998). Obscured AGN models for the XRB (eg Gunn and Shanks 2001; Almaini et al. 1999) predict (conservatively) that 5-25% of the far-ir background may be produced by obscured AGN. To distinguish the relative contributions of starbursts and obscured AGN we require, together with far-ir observations, deep X-ray and near-ir observations to define the spectral energy distributions (SEDs) and so reveal the presence of an AGN, quantify the amount of dust obscuration and measure dust temperatures.

Proposed Observations. Our observations must target the deepest available X-ray fields in order to find the most absorbed AGN and so put the tightest constraints on the obscured AGN model of the XRB. However even the very deepest X-ray observations, from XMM (~200 ksec), will only contain ~200 X-ray sources. Thus one field does not contain sufficient AGN to study evolutionary changes, for which ~1,000 AGN are required. The XMM-LSS field (Table 5.1) is the primary target for this work. However as described in section 5.2 our eventual goal is complete coverage of all three fields.

X-ray background models based on obscured AGN and their evolution indicate that to detect 80% of the AGN in a 50ksec XMM observation requires K=21, and K=22 for a 200ksec observation. Going to K=23 detects only 1% more. Our strategy therefore is to observe selected key fields to K=21 as early as possible within the DXS programme, and to make the case for deeper observations later as required. The highest priority fields are those where 3.6-160 micron observations are scheduled with SIRTF, to complete the SEDs from X-ray to far-ir wavelengths. The SIRTF GTOs reach to L~21 and M~19.5, providing equivalent depth to WFCAM at K=22. The SIRTF Legacy observations will typically be about 2 magnitudes less deep.