Predictions for the Chandra Multi-wavelength Project

See the project web page for a description of the Chandra Multi-wavelength Project (ChaMP), ).  The primary contact person is Belinda Wilkes.

How the predictions were derived


So far, only ACIS fields have been considered.  They are much more numerous, generally longer, and cover a larger fraction of the sky.  The ACIS sensitivity is somewhat better than that of the HRC and also has energy resolution so that one may define a hard band and a soft band so that one may detect sources separately in the two bands.  In keeping with tradition, I have computed expectations separately for two different bandpasses: 0.5-3.5 keV (soft) and 2-10 keV (hard).

My objective is to estimate the numbers of extragalactic sources that may be found in the ChaMP survey.  There are several important steps:

  1. Compute the expected count rates of X-ray sources that may be found in the ChaMP survey.  This step is complicated by the fact that sources have a distribution of spectral shapes and that the spectra are not simple power laws.  As with previous studies, I divide the population into "absorbed" and "unabsorbed" AGN.  The spectral parameters for the two populations are given in table 1.    The count rates give the expected count rate for a source with the given spectral parameters in the specific portion of the bandpass (restricting the effective area file) for a flux in that bandpass of 1.e-15 erg/cm2/s.  Note also:
    2. Type Class Bandpass Gamma log NH Rate (FI) Rate (BI)
    Unabsorbed QSO/Sy1 0.5-3.5 keV 2.0 20.5 0.158 0.357
    Absorbed Sy 2 0.5-3.5 keV 1.5 22.0 0.105 0.136
    Unabsorbed QSO/Sy1 2-10 keV 1.5 20.5 0.036 0.040
    Absorbed Sy 2 2-10 keV 1.5 23.5 0.170 0.150
    Table 1.  Parameters of AGN spectra assumed for the number-flux predictions.
  2. Divide the ACIS-I and ACIS-S into regions (and chip type) based on the mirror beam size.
  3. Fit a simple relation to exact calculations of the detection threshold, T, required so that there are less than 10 false sources per year due to background fluctuations in the entire ChaMP survey.
  4. Using estimates of the expected detector background in each region defined in step 2 and the threshold computed from step 4, compute a detection threshold and the resultant source flux sensitivity.  The  actual range  (figure 4) of detection thresholds is only a factor of 10.  When using a region containing half the power, the observed counts represent half the flux, so the predicted count rates are actually twice the value used for detection.
    5. ACIS Chip Bandpass Internal BG Cosmic BG Total BG
    ACIS-I (FI) 0.5-3.5 keV 0.000199 0.00037 0.00057
    2-10 keV 0.00052 0.00016 0.00068
    ACIS-I (BI) 0.5-3.5 keV 0.00059 0.00129 0.00187
    2-10 keV 0.00152 0.00016 0.00169
    Table 2. Background values taken from the Proposer's Guide in count/sq arcsec/ks.
  5. Derive the coverage functions, which give the sky solid angle surveyed to any specified flux limit.  These curves depend on the flux-count conversion factor and the background, so curves are shown for each bandpass/source type combination.  Figure 5 shows the hard band coverage function and figure 6 shows the coverage functions for the soft band.
  6. Set parameters the number count distributions, N(>S).  The models are taken from the X-ray background synthesis work of Comastri et al. (1995, A&A, 296, 1) and are summarized in Table 3.  The general form is given by: N(>S) = K*{S/Sb}^a, breaking at flux Sb to a steeper slope.  In the soft band, the absorbed sources are dominated by AGN with column densities near 3e22 below the break and 3e21 above the break (see Comastri et al, figure 2).  In the hard band, AGN with column densities near 3e23 dominate, especially below the flux break (see Comastri et al, figure 5).  Figure 7 shows the models for the hard band and the limit given by the diffuse X-ray background.
    7. Source Type Bandpass K (src/deg^2) Sb (erg/s/cm^2) a_low a_high
    Unabsorbed 0.5-3.5 100 1.0e-14 0.92 1.55
    2-10 55 1.6e-14 0.94 1.60
    Absorbed 0.5-3.5 90 3.0e-15 0.85 1.70
    2-10 200 1.2e-14 1.00 1.75
    Table 3.  Parameters of the N(>S) models used for predicting the ChaMP source counts.
  7. Integrate the predicted number of sources in all ACIS regions as a function of sensitivity.  A total N(S) is computed for each band by combining the absorbed AGN with the unabsorbed quasar counts.

Results

About 2000 sources are expected in the soft band.  Figure 8 shows the expected distribution of source fluxes.  The absorbed sources (akin to Sy 2s) comprise a relatively small fraction of the total.


 


Figure 8.  Expected counts for the soft band.  The curves give the number of sources that will be found in the survey with a flux greater than S.   Figure 9 is a version of this figure with a linear vertical scale.


 


About 700 sources are expected in the hard band, about equally divided between absorbed and unabsorbed AGN.   Figure 10 shows the expected distribution of source fluxes.  The unabsorbed AGN will probably also be detected in the soft band while the majority of the absorbed AGN will not be detected there, due to the high expected column densities.


 


Figure 10.  Expected counts for the hard band, as in Figure 8.  In this case, the expected number of absorbed AGN is comparable to that of unabsorbed AGN.  Figure 11 is a version of this figure with a linear vertical scale.


 

PS versions of figures

Left to do

  1. Compute predictions for the HRC-I observations (there are 5 in the Cycle 1 ChaMP).  The exposure times are relatively short, however.
  2. Add area from other ACIS chips that are used besides those considered in step 2.  This will add several sq. deg at brighter flux limits.

Send comments and questions to me at hermanm@space.mit.edu

Author: Herman L. Marshall
Updated: 4/30/99