In the following each parameter is described in general together with the error (where applicable). A table then indicates the combinations of camera (ca = EP, PN, M1, M2) and bands (b = 1, 2, 3, 4, 5, 8, 9; see 4XMM UG Tab. 1). 'EP' stands for EPIC and indicates parameters combined from those from the available cameras. Most parameters are determined by the SAS task emldetect, which does a simultaneous maximum likelihood PSF fit to the source count distribution in each energy band and camera (for details see 2XMM UG Sec. 3.1.2 f))
XID values (band 9) are derived in a seperate run of emldetect on input images, exposure maps and background maps that are merged for the bands 2, 3, 4. The input source position were taken from the first run (which determined the bands 1 - 5 and band 8 values) and kept fixed. As a consequence XID values are single-band values, which ensures a better handling of the error values, cf. 2XMM UG Sec. 3.1.2 g).
Entries have a NULL when the detection was not observed with the respective camera.
The individual parameters are (ca = EP, PN, M1, M2; b = 1, 2, 3, 4, 5, 8, 9; n = 1, 2, 3, 4):
| Fluxes: | ca_b_FLUX and ca_b_FLUX_ERR |
| Count rates: | ca_b_RATE and ca_b_RATE_ERR |
| Source counts: | ca_b_CTS and ca_b_CTS_ERR |
| Maximum likelihoods: | ca_b_DET_ML |
| Extent parameters: | EP_EXTENT, EP_EXTENT_ERR, and EP_EXTENT_ML |
| Hardness ratios: | ca_HRn and ca_HRn_ERR |
| Exposure map: | ca_b_EXP |
| Background map: | ca_b_BG |
| Pileup: | ca_PILEUP |
| Vignetting: | ca_b_VIG |
| Ontime values: | ca_ONTIME |
| Offaxis angle: | ca_OFFAX |
| Maskfrac value: | ca_MASKFRAC |
| Distance to the nearest neighbour: | DIST_NN |
Fluxes are given in erg/cm2/s and are calculated by the SAS tasks emldetect and by srcmatch for the input bands (see Secs 3.1.2 f) and i) in the 2XMM UG). Note that they correspond to the flux in the entire PSF and do not need any further corrections for PSF losses.
For the individual cameras, individual-band fluxes (bands 1 - 5, 9) are calculated from the respective band count rate using the filter- and camera-dependent energy conversion factors given in Tab. 2 and corrected for the dead time due to the read-out phase. These can be 0.0 if the detection has no counts. The respective errors are calculated from the respective band count rate error using the respective energy conversion factors.
Combined band fluxes (band 8) for the individual cameras are the sum of the fluxes and errors from each band (1 - 5).
The EPIC flux in each band is the mean of the band-specific
detections in all cameras weighted by the errors, with the error on
the weighted mean given by
EP_b_FLUX_ERR = SQRT (
1.0 / SUM ( 1 / ca_b_FLUX_ERR2 )),
where ca = PN, M1, M2.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] |
| PN | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] |
| M1 | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] |
| M2 | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] | flux [err] |
Count rates and errors are given in counts/second and are derived by the SAS task emldetect (see 2XMM UG Sec. 3.1.2 f).
The individual-band count rate (bands 1 - 5, 9) is the band-dependent source counts (ca_b_CTS) divided by the exposure map, which combines the mirror vignetting, detector efficiency, bad pixels and CCD gaps, and an OOT-factor (Out Of Time), depending on the PN modes (PN_SUBMODE). The source counts and with it the count rates were implicitly background subtracted during the fitting process. They correspond to the count rate in the entire PSF and do not need any further corrections for PSF losses. Note that rates can be 0.0 (but not negative) if the source is too faint in the respective band to be detectable.
Combined band count rates (band 8) for each camera are calculated as the sum of the count rates in the individual bands 1 - 5.
The EPIC rates are the sum of the camera-specific count rates in the respective band.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | rate [err] | - | - | - | - | - | rate [err] |
| PN | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] |
| M1 | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] |
| M2 | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] | rate [err] |
Source counts and errors are given in counts and are derived by the SAS task emldetect (see 2XMM UG emldetect).
The individual-band source counts (not given in the catalogue) are derived under the total PSF (point spread function) and corrected for background. The PSF is fitted on subimages of radius 60 arcseconds in each band (CUTRAD), which means, that in most cases at least 90 % of the PSF (if covered by the detector) was effectively used in the fit.
Combined band source counts (band 8) for each camera are calculated as the sum of the source counts in the individual bands 1 - 5.
The EPIC band 8 counts are the sum of the (available) individual camera band 8 counts
The error is the statistical 1 σ error on the total source counts of the detection.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | cts [err] | - | - | - | - | - | - |
| PN | cts [err] | - | - | - | - | - | - |
| M1 | cts [err] | - | - | - | - | - | - |
| M2 | cts [err] | - | - | - | - | - | - |
Maximum likelihoods are derived by the SAS task emldetect (see 2XMM UG Sec. 3.1.2 f)).
The individual-band maximum likelihood values (bands 1 - 5, 9) stand for the detection likelihood of the source, L = - ln p, where p is the probability of the detection occurring by chance. While the detection likelihood of an extended source is computed in the same way, systematic effects such as deviations between the real background and the model, have a larger effect on extended sources and thus detection likelihoods of extended sources are more uncertain.
To calculate the maximum likelihood values for the combined band 8 and EPIC the sum of the individual likelihoods is being normalised to two degrees of freedom using the function
ML_corr = -ln(gammaq (ndof/2, ML)),
where
ndof = 2 (for
xpos,ypos) + N_images for point sources,
ndof = 3 (for xpos,ypos,extent) + N_images
for extended sources,
gammaq = - ln (Q(a,x)) = - ln (1 - P(a,x)),
and P is the incomplete gamma function.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | detml | - | - | - | - | - | detml |
| PN | detml | detml | detml | detml | detml | detml | detml |
| M1 | detml | detml | detml | detml | detml | detml | detml |
| M2 | detml | detml | detml | detml | detml | detml | detml |
The extent radius and error as well as the extent likelihood of a source detected as extended is determined by the SAS task emldetect (see 2XMM UG Sec. 3.1.2 f)). It is derived by convolving a beta-model profile with the source PSF and fitting the result to the source image. Anything below 6" is considered to be a point source and the extent is set to zero. To avoid non-converging fitting an upper limit of 80" is imposed.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | ext [err] | - | - | - | - | - | - |
| PN | - | - | - | - | - | - | - |
| M1 | - | - | - | - | - | - | - |
| M2 | - | - | - | - | - | - | - |
The extent likelihood is the likelihood of the detection being extended as given by EXTENT_ML = - ln (P) , where P is the probability of the extent occurring by chance.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | extent ml | - | - | - | - | - | - |
| PN | - | - | - | - | - | - | - |
| M1 | - | - | - | - | - | - | - |
| M2 | - | - | - | - | - | - | - |
The hardness ratios for each camera are derived by the SAS task emldetect (see 2XMM UG Sec. 3.1.2 f)). They are defined as the ratio between the bands A and B:
HR(A,B) = (band B - band A) / (band A + band B).
Note that in the case where the rate in one band is 0.0 (i.e., too faint to be detected in this band) the hardness ratio will be -1 or +1 which is only a lower or upper limit, respectively. In the case where the rate in both bands is zero, the hardness ratio is undefined (NULL).
There are four hardness ratios (n) using the following bands:
| HR1: | bands 1 & 2 |
| HR2: | bands 2 & 3 |
| HR3: | bands 3 & 4 |
| HR4: | bands 4 & 5 |
Errors are the 1 σ error on the hardness ratio 1 as derived by the SAS task emldetect.
EPIC hardness ratios are calculated by the SAS task srcmatch and are averaged over all three cameras (PN, M1, M2). Note that no energy conversion factor was used and that the EPIC hardness ratios are de facto not hardness ratios but an equivalent number helpful to characterise the hardness of a source.
| ca \ n | 1 | 2 | 3 | 4 |
| EP | hr [err] | hr [err] | hr [err] | hr [err] |
| PN | hr [err] | hr [err] | hr [err] | hr [err] |
| M1 | hr [err] | hr [err] | hr [err] | hr [err] |
| M2 | hr [err] | hr [err] | hr [err] | hr [err] |
The exposure maps are made by the SAS task eexpmap (see 2XMM UG Sec. 3.1.2 a)); they combine the mirror vignetting, detector efficiency, bad pixels and CCD gaps. The exposure map values in the catalogue are given in seconds and are derived by the SAS task emldetect as the PSF weighted mean of the area of the subimages (radius 60 arcseconds) in the individual-band exposure maps.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | - | - | - | - | - | - | - |
| PN | - | exp | exp | exp | exp | exp | - |
| M1 | - | exp | exp | exp | exp | exp | - |
| M2 | - | exp | exp | exp | exp | exp | - |
The background maps are made by the SAS task esplinemap (see 2XMM UG Sec. 3.1.2 d)); they are made using a 12 x 12 nodes spline fit on the source-free individual-band images. The background map values in the catalogue are given in counts per pixel and are derived by the SAS task emldetect as the background map value at the given detection position. Note that the source fitting routine uses the background map itself rather than the single value given here. The value is zero if the detection position lies outside the FOV.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | - | - | - | - | - | - | - |
| PN | - | bg | bg | bg | bg | bg | - |
| M1 | - | bg | bg | bg | bg | bg | - |
| M2 | - | bg | bg | bg | bg | bg | - |
As of 4XMM we provide three new columns (PN_PILEUP, M1_PILEUP and M2_PILEUP) quantifying whether each detection may be affected by pile-up in any instrument. A value below 1 corresponds to negligible pile-up (less than a few % flux loss) while values larger than 10 denote heavy pile-up. Pile-up is dependent on time for variable sources. Our pile-up levels are not based on a fit of the full images using a pile-up model (Ballet 1999). For point sources, they are based on the measured count rates reported in the catalogue over the full energy band, transformed into counts per frame. The thresholds (at which the pile-up level is set to 1) are set to 1.3 cts/frame for MOS and 0.15 cts/frame for PN. For extended sources, the pile-up level is based on the measured count rate per CCD pixel at the source position, and therefore refers to the peak brightness, assuming this can be considered uniform at the pixel scale (4.1 arcsec for PN). The threshold is set for all instruments to 5 x 10-3 cts/frame/pixel, such that the flux loss is also a few % when the pile-up level is 1 (see the 4XMM User Guide Sec. 3.4))
The vignetting values in the catalogue are derived by the SAS task emldetect (see 2XMM UG Sec. 3.1.2 f)); they are a function of energy band and offaxis angle. (Vignetting values used in the source parametrization come from the vignetted exposure maps.)
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | - | - | - | - | - | - | - |
| PN | - | vig | vig | vig | vig | vig | - |
| M1 | - | vig | vig | vig | vig | vig | - |
| M2 | - | vig | vig | vig | vig | vig | - |
The ontime value, given in seconds, is the total good exposure time (after GTI filtering) of the CCD where the detection is positioned. Note that some source positions fall into CCD gaps or outside of the detector and can therefore, in some cases, have a NULL. The all EPIC (EP) value represents the largest ontime value of any of the individual cameras used.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | ontime | - | - | - | - | - | - |
| PN | ontime | - | - | - | - | - | - |
| M1 | ontime | - | - | - | - | - | - |
| M2 | ontime | - | - | - | - | - | - |
The offaxis angle, given in arcminutes, is the distance between the detection position and the onaxis position on the respective detector. Note that the offaxis angle for a camera can be larger than 15 arcminutes when the detection is located outside the FOV of that camera. The all EPIC (EP) value reflects the smallest of the individual camera values.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | offax | - | - | - | - | - | - |
| PN | offax | - | - | - | - | - | - |
| M1 | offax | - | - | - | - | - | - |
| M2 | offax | - | - | - | - | - | - |
The maskfrac value is the PSF weighted mean of the detector coverage of a detection as derived from the detection mask (see 2XMM UG Sec. 3.1.2 b) for details). It depends slightly on energy; only band 8 values are given here which are the minimum of the energy-dependent maskfrac values. Sources which have less than 0.15 of their PSF covered by the detector are considered as being not detected.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | - | - | - | - | - | - | - |
| PN | maskfrac | - | - | - | - | - | - |
| M1 | maskfrac | - | - | - | - | - | - |
| M2 | maskfrac | - | - | - | - | - | - |
The distance to the nearest neighbouring detection is given in arseconds; they are derived by the SAS task emldetect (see 2XMM UG Sec. 3.1.2 f)). Emldetect uses an internal threshold of 6 arseconds (before positional fitting) for splitting a source into two.
| ca \ b | 8 | 1 | 2 | 3 | 4 | 5 | 9 |
| EP | distnn | - | - | - | - | - | - |
| PN | - | - | - | - | - | - | - |
| M1 | - | - | - | - | - | - | - |
| M2 | - | - | - | - | - | - | - |