XMM–Newton observations of seven soft X-ray excess quasi-stellar objects
K. L. Page
1
N. Schartel
0
M. J. L. Turner
1
P. T. O'Brien
1
0
XMM-Newton Science Operations Centre, European Space Agency
, Villafranca del Castillo, Apartado 50727, E-28080 Madrid,
Spain
1
X-Ray and Observational Astronomy Group,
Department of Physics & Astronomy, University of Leicester
, Leicester LE1 7RH
A B S T R A C T XMM-Newton observations of seven quasi-stellar objects (QSOs) are presented and the EPIC spectra analysed. Five of the active galactic nuclei (AGN) show evidence for Fe K emission, with three being slightly better fitted by lines of finite width; at the 99 per cent level they are consistent with being intrinsically narrow, though. The broad-band spectra can be well modelled by a combination of different temperature blackbodies (BBs) with a power law, with temperatures between kT 100 and 300 eV. On the whole, these temperatures are too high to be direct thermal emission from the accretion disc, so a Comptonization model was used as a more physical parametrization. The Comptonizing electron population forms the soft excess emission, with an electron temperature of 120-680 eV. Power-law, thermal plasma and disc BB models were also fitted to the soft X-ray excess. Of the sample, four of the active galactic nuclei are radio quiet and three radio loud. The radio-quiet QSOs may have slightly stronger soft excesses, although the electron temperatures cover the same range for both groups.
1 I N T R O D U C T I O N
At energies below 2 keV, the spectra of most active galactic nuclei
(AGN) show an upturn, away from the extrapolation of the
highenergy (210 keV) power law. This so-called soft excess
emission is thought to be common in both Seyfert galaxies and QSOs
(quasi-stellar objects). The first such soft excess was identified in
Mrk 841 by Arnaud et al. (1985); Turner & Pounds (1989) then
found that 50 per cent of their EXOSAT sample showed steeper
spectral slopes at low energies. Likewise, Walter & Fink (1993) and
Schartel et al. (1996) found that the ROSAT Position Sensitive
Proportional Counter spectral index tends to be significantly steeper
than that measured above 2.4 keV (typically 1.9; e.g. Nandra &
Pounds 1994). More recently, Pounds & Reeves (2002) have
discussed the frequent presence of soft excesses in XMMNewton data.
Many papers have been published about the soft excess, covering
both observational results with ROSAT (e.g. Fiore et al. 1994; Piro,
Matt & Ricci 1997), Ginga, EXOSAT (e.g. Saxton et al. 1993) and
Einstein (e.g. Masnou et al. 1992; Zhou & Yu 1992) and theoretical
work (e.g. Czerny & Elvis 1987; Czerny & Zycki 1994; Xia & Zhang
2001). It is generally thought that the soft excess may be linked to
the hot tail-end of the Big Blue Bump (BBB), or is an extension of
the ultraviolet (UV) band. The BBB/UV excess is likely to be owing
to thermal emission from the accretion disc surrounding the black
hole (e.g. Shields 1978; Malkan & Sargent 1982). However, this
thermal emission is not hot enough to also account for the soft X-ray
flux; hence, Comptonization is often invoked to explain the resultant
emission. In this scenario, the direct thermal emission from the
accretion disc is observed as the optical and UV spectrum. Some of
the disc photons, however, undergo inverse Compton scattering with
a population of hot electrons, thus gaining energy and producing a
broader spectrum, which appears similar to a power law over a
limited energy range (assuming unsaturated Comptonization).
All seven objects in this paper, listed in Table 1, have been
previously observed by ROSAT (Schartel et al. 1996) and were each noted
to have steep photon indices over the 0.12.4z,keV ROSAT band.
The Galactic absorption in the direction of each low-redshift QSO
is small and it has been previously found that there is no significant
evidence for additional, intrinsic absorption in any of the objects.
These QSOs, therefore, represent a useful sample to investigate the
soft excess in both radio-quiet and radio-loud AGN.
2 X M M N E W T O N O B S E RVAT I O N S
The QSOs in this paper were observed by XMMNewton between
revolutions 105 and 315 (Table 2). SAS (Science Analysis Software)
v. 5.4 was used to produce the event lists for the MOS and PN EPIC
(European Photon Imaging Camera) instruments, which were then
filtered using XMMSELECT. For most of the objects, events covering
patterns 012 (single, double, triple and quadruple events) were
selected for the MOS exposures, while 04 (singles and doubles)
were used for the PN, after ensuring that there were no problems
with pile up in either instrument. PG 0804+761 did appear to show
pile-up effects (using the SAS task EPATPLOT), though, so pattern zero
spectra (i.e. single events only) were used for this object. Spectra
were extracted within a circular region of 40 arcsec centred on
V band
1.4-GHz
flux (mJy)a
Q 0056363
PG 0804+761
Mrk 1383
Mrk 876
B2 1028+31
B2 1128+31
B2 1721+34
aRadio fluxes from the NVSS.
Q 0056363
PG 0804+761
Mrk 1383
Mrk 876
B2 1028+31
B2 1128+31
B2 1721+34
Observation
date (rev.)
2000-07-05 (105)
2000-11-04 (166)
2000-07-28 (116)
2001-04-13 (246)
2001-08-29 (315)
2000-12-06 (182)
2000-11-22 (175)
2001-02-13 (216)
2001-02-26 (223)
Exposure time (ks)
MOS1 MOS2
aLW, large window; SW, small window; FF, full frame.
each object; background spectra were obtained from an adjacent
area of blank sky. Finally, the spectra were binned in such a way
that each individual bin contained a sufficient number of counts to
apply the 2 statistic (the minimum used was 25) and the energy
range of each bin is larger than one third of the FWHM to avoid
oversampling effects (i.e. to ensure that the bins were uncorrelated).
Version 11.1.0 of XSPEC was then used to analyse the data.
Table 1 lists whether the QSOs are radio loud or quiet, their
redshift and Galactic absorption. The 1.4-GHz radio fluxes were
taken from the National Radio Astronomy Observatories Very Large
Array Sky Survey (NVSS; Condon et al. 1998) catalogue browser.
The completeness limit of the survey is 2.5 mJy so, as no object
was catalogued at the coordinates of Q 0056363, 2.5 mJy is taken
as the upper limit to the radio flux. Table 2 gives information about
the set-up of XMMNewton for each observation, while Table 3 then
gives the right ascension and declination of the objects, together with
source count rates and net counts.
Although the objects will be referred to as QSOs throughout this
paper, from their definition in the VeronVeron catalogue, their
magnitudes are on the border between bright Seyfert galaxies and
QSOs (M B 23) and they are thus sometimes referred to as type 1
Seyfert galaxies in the literature. H0 has been taken as 70 km s1
Mpc1, and q 0 = 0.5; errors are given at the 1 level (e.g. 2 =
2.3 for two interesting parameters).
3 S P E C T R A L A N A LY S I S 3.1 212 keV As is conventional, each background-subtracted spectrum was first simply fitted with a single power law over the entire bandpass of
Gal. abs.
(1020 cm2)
(...truncated)