Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0280-1-2.ps.gz Äàòà èçìåíåíèÿ: Wed Mar 28 12:35:03 2012 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 23:04:29 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: http astrokuban.info astrokuban

XMM­Newton CCF Release Note
XMM­CCF­REL­280
2­D PSF parameterization
A. M. Read & R. Saxton
March 28, 2012
1 CCF components
Name of CCF VALDATE List of Blocks changed Change in CAL HB
XRT1 XPSF 0014.CCF 2000­01­01 ELLBETA PARAMS YES
XRT2 XPSF 0014.CCF 2000­01­01 ELLBETA PARAMS YES
XRT3 XPSF 0014.CCF 2000­01­01 ELLBETA PARAMS YES
2 Change
The elliptical beta, 2­D PSF model (ELLBETA) consists of an elliptical King function plus a central
narrow Gaussian to model the core plus spoke structures (see [1] and CAL­SRN­0263 for a fuller
description). This release includes two separate changes: a modification to the strength of the spokes
and a tuning of the slopes of the elliptical King function.
2.1 Spokes
A spatial analysis of fits of the 2­D PSF to real sources showed that small problems with the spoke
strength still exist [2]. To correct this, the spoke structure has been reworked in this release to vary
the intensity along the spokes such that
1. The overall spoke intensity has been modified to CHNGFRAC=0.42 from its previous value
of 0.375.
2. The spokes start at a radius, RD1=10''. For smaller radii the spoke strength is zero.
3. The spoke strength increases linearly until reaching a peak at RD2=110''.
1

XMM­Newton CCF Release XMM­CCF­REL­280 Page: 2
Table 1: Observations used in the determination of optimum King profile slopes
Source name Obsid
1H0414+009 0161160101
3C 273 0126700301
3C 273 0414190701
H1426+428 0165770201
IRAS0578+1626 0502090501
MCG­6­30­15 0029740101
MCG­6­30­15 0029740701
MCG­6­30­15 0029740801
PG1116+215 0201940101
4. It then decreases linearly until attaining zero intensity at RD3=180''. The spoke strength is
zero beyond this radius.
These values are stored in the following keywords in the header of the ELLBETA PARAMS
extension:
. CHNGFRAC ­ (maximum) fraction of image to be filtered (0[none]­1[all]) which defines the
strength of the spoke
. RD1 ­ minimum radius to start spoke filtering (arcsec)
. RD2 ­ radius at which the spoking peaks (arcsec)
. RD3 ­ radius at which the spoking tails to zero (arcsec)
Note that the keywords related to the spoke structure currently have the same values for the
three cameras.
2.2 King function slope
A detailed spectral investigation of the 2D­PSF fit to several bright on­axis sources has revealed
that at some energies the measured flux varies depending on the extraction region. This has been
resolved by tuning the slope of the elliptical King function individually for each camera and for 6
energy bands from 0.15--10 keV.
In detail, source spectra have been extracted from annuli of outer radius 40'' and inner radius
ranging from 0'' to 20'' from the set of observations of bright on­axis AGN listed in Table 1. E#ective
area files (ARFs) have been produced by the SAS task arfgen and the spectra have been fit using
an absorbed power­law model.

XMM­Newton CCF Release XMM­CCF­REL­280 Page: 3
Figure 1: The ratio of the model on­spoke to o#­spoke intensity plotted against radius (in 1 arcsecond
pixels).
3 Scientific impact of this update
3.1 Spoke structure
The spoke structure, as represented in XRTn XPSF 0013.CCF, had been tuned using data contam­
inated with a galactic scattering halo [2]. In this release, the radial structure of the spoke has been
reworked to account for this e#ect. The overall change is a slight increase in the strength of the
spokes plus a radial change in their intensity. The new implementation is a better fit to real data
and should lead to a reduction in the number of spurious detections which are found in the wings
of bright point sources during source detection runs.
3.2 Encircled energy
In the current PSF implementation it was found that at some energies the fitted source flux was
quite dependent on the extraction region. In this release the King profile slope has been adjusted for
each camera, and at each energy, to make the flux as region­independent as possible. In Fig. 2 we
show the relationship between extraction region, flux and power­law slope for the obsid=0126700301
observation of 3C 273 in the energy band 0.54--0.85 keV. The relationship can be seen to be flatter
for the new parameterisation. This gives higher confidence in the robustness of the solution and
means that fluxes will not be a#ected when the core of the PSF has to be excluded to counter pile­up
e#ects. Changes have been made to the slope parameterisation at 0 and 3 arcminutes o#­axis. The
parameters for the PSF at larger o#­axis angles have not been changed.

XMM­Newton CCF Release XMM­CCF­REL­280 Page: 4
Figure 2: Spectral fit parameters for an observation of 3C 273 (obsid=0126700301) in the energy
range 0.54--0.85 keV. Spectra have been extracted from an annulus of outer radius 40'' and inner
radius ranging from 0 to 400 sky pixels (equivalent to 0 to 20''). Upper panels give the photon
index, normalisation and flux for the current 2D­PSF parameterisation (black=EPIC­pn, red=EPIC­
MOS1, green=EPIC­MOS2); lower panels give the same information for this new release of the
2D­PSF. This observation su#ers from pile­up in the inner 5'' (50 sky pixels).

XMM­Newton CCF Release XMM­CCF­REL­280 Page: 5
Figure 3: Spectral fit parameters for an observation of 1H 0414+009 (obsid=0161160101), in the
energy range 2.0--4.5 keV, using the current 2D­PSF parameterisation (upper panels) and this release
(lower panels) .

XMM­Newton CCF Release XMM­CCF­REL­280 Page: 6
4 Estimated scientific quality
The spoke model has been designed to fit the data in [2] to an accuracy of a few % in the three
EPIC cameras.
For the tested observations of 3C 273 and MCG­6­30­15 the changes introduced in this release
lead to fluxes which are consistent to within # 5% for annular extraction regions with outer radius
40'' and inner radius ranging from 5­20'', for all energies and for the three EPIC cameras (see Fig.
2 for an example at 0.54­0.85 keV). Note that data within the inner 5'' are likely a#ected by a small
amount of pile­up in these sources.
Some of the sources in the analysis had trends which went in opposite directions making it
impossible to resolve all the problems in all the sources with a single set of PSF parameters. For
example, the lower statistic sources, 1H 0414+009 and H 1426+428 show larger deviations in spectral
parameters with inner extraction radius for energies above 2.0 keV. This is perhaps related to low
statistics and background subtraction issues. An example is shown in Fig.3. Nevertheless, overall
the changes introduced here give an improvement in the flux consistency in most sources at most
energies. Figure 4 shows the standard deviation about the mean flux for the the old and new CCFs.
Points below the dotted line represent improved performance.
5 Test procedure and results
The new CCFs have been used within calview, psfgen and eregionanalyse to test that they
produce the correct PSF images and encircled energy values.
The modified SAS code needed to take advantage of the new CCF structure is not yet public.
To ensure that these changes do not adversely a#ect the current public SAS (v11.0) the same three
tasks, calview, psfgen and eregionanalyse were executed from the current SAS using the new
CCF elements. No problems were seen.
6 Future changes
It is likely that changes to the PSF parameters at o#­axis angles greater than 3 arcminutes will be
needed in the future.
7 References
[1] Read, A., Rosen, S., Saxton, R. & Ramirez, J. 2011, A&A, 534, 34
[2] Owen, R. & Ballet, J. 2011, SSC­CEA­TN­1001, v2.1. (http://xmm2.esac.esa.int/documents/CAL­
SSC­CEA­TN­1001­2.1.pdf)

XMM­Newton CCF Release XMM­CCF­REL­280 Page: 7
Figure 4: A plot of the consistency of the measured flux with changing extraction region. The
axes represent the standard deviation of the flux normalised by the mean flux for the current CCF
(version 12) and the new CCF (version 14). Di#erent cameras are represented by di#erent colours
while filled and open symbols represent 3C273 and 1H0414 respectively. Energies are represented
by the circle (0.54­0.85 keV), star (0.85­2 keV), triangle (2­4.5 keV) and square (4.5­10 keV) shapes.

XMM­Newton CCF Release XMM­CCF­REL­280 Page: 8