Suzaku Constraints on the Soft and Hard Excess Emissions from Abell 2199

Abstract

The nearby ($z=0.03015$) cluster of galaxies Abell~2199 was observed by Suzaku in X-rays, with five pointings for $\sim 20$ ks each. From the XIS data, the temperature and metal abundance profiles were derived out to $\sim 700$ kpc (0.4 times virial radius). Both these quantities decrease gradually from the center to peripheries by a factor of $\sim 2$, while the oxygen abundance tends to be flat. The temperature within 12' ($\sim 430$ kpc) is $\sim 4$ keV, and the 0.5--10 keV X-ray luminosity integrated up to 30' is $(2.9 \pm 0.1) \times 10^{44}$ erg s$^{-1}$, in agreement with previous XMM-Newton measurements. Above this thermal emission, no significant excess was found either in the XIS range below $\sim 1$ keV, or in the HXD-PIN range above $\sim 15$ keV. The 90%-confidence upper limit on the emission measure of an assumed 0.2 keV warm gas is (3.7--7.5) $\times 10^{62}$ cm$^{-3}$ arcmin$^{-2}$, which is 3.7--7.6 times tighter than the detection reported with XMM-Newton. The 90%-confidence upper limit on the 20--80 keV luminosity of any power law component is $1.8 \times 10^{43}$ erg s$^{-1}$, assuming a photon index of 2.0. Although this upper limit does not reject the possible 2.1$\sigma$ detection by the BeppoSAX PDS, it is a factor of 2.1 tighter than that of the PDS if both are considered upper limits. The non-detection of the hard excess can be reconciled with the upper limit on diffuse radio emission, without invoking the very low magnetic fields ($< 0.073 \mu$G) which were suggested previously. Comment: 28 pages, 10 figures, accepted for publication in PASJ

Full text

Page 1

arXiv:0911.5205v1 [astro-ph.CO] 27 Nov 2009
Suzaku Constraints on the Soft and Hard Excess
Emissions from Abell 2199
Madoka kawaharada,1Kazuo makishima,2,1Takao kitaguchi,2Sho okuyama,2
Kazuhiro nakazawa,2and Yasushi fukazawa3
1Cosmic Radiation Laboratory, RIKEN, 2-1 Hirosawa, Wako, Saitama 351–0198.
E-mail : kawahard@crab.riken.jp
2Department of Physics, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113–0033.
3Department of Physical Science, Hiroshima University, 1-3-1 Kagamiyama, Higashi-hiroshima
739-8526.
(Received 2009 July 9; accepted 2009 November 26)
Abstract
The nearby (z = 0.03015) cluster of galaxies Abell 2199 was observed by Suzaku
in X-rays, with five pointings for ∼ 20 ks each. From the XIS data, the temperature
and metal abundance profiles were derived out to ∼700 kpc (0.4 times virial radius).
Both these quantities decrease gradually from the center to peripheries by a factor
of ∼ 2, while the oxygen abundance tends to be flat. The temperature within 12′
(∼ 430 kpc) is ∼ 4 keV, and the 0.5–10 keV X-ray luminosity integrated up to 30′is
(2.9±0.1)×1044erg s−1, in agreement with previous XMM-Newton measurements.
Above this thermal emission, no significant excess was found either in the XIS range
below ∼1 keV, or in the HXD-PIN range above ∼15 keV. The 90%-confidence upper
limit on the emission measure of an assumed 0.2 keV warm gas is (3.7–7.5) ×1062
cm−3arcmin−2, which is 3.7–7.6 times tighter than the detection reported with XMM-
Newton. The 90%-confidence upper limit on the 20–80 keV luminosity of any power
law component is 1.8×1043erg s−1, assuming a photon index of 2.0. Although this
upper limit does not reject the possible 2.1σ detection by the BeppoSAX PDS, it
is a factor of 2.1 tighter than that of the PDS if both are considered upper limits.
The non-detection of the hard excess can be reconciled with the upper limit on diffuse
radio emission, without invoking the very low magnetic fields (<0.073µG) which were
suggested previously.
Key words: galaxies: clusters: individual (Abell 2199) — X-rays: galaxies: clus
ters
1

Page 2

1.Introduction
The intra-cluster medium (ICM) in clusters of galaxies, i.e, hot thermal plasmas in col-
lisional ionization equlibria confined within their gravitational potential, constitutes the most
dominant form of cosmic baryons that has ever been detected. Although the ICM radiates
predominantly optically-thin thermal X-ray emission, some clusters have been reported to ex-
hibit excess signals above the thermal emission at the lowest or highest energy ends of their
X-ray spectra. One interpretation of the soft excess is emission from Warm Hot Intergalactic
Medium (WHIM), which has been predicted from cosmological simulations (e.g. Cen & Ostriker
1999) and is expected to solve the so-called missing baryon problem (e.g Fukugita et al. 1998).
Another interpretation of the spectral soft and/or hard excess is non-thermal emission from
accelerated particles in galaxy clusters (e.g. Lieu et al. 1999; Fusco-Femiano et al. 1999): a
power-low shaped spectrum can exceed the dominant thermal emission at both sufficiently high
and low energies. This provides a possibility that galaxy clusters are giant accelerators in the
universe. Thus, the search for soft and hard excess signals in clusters forms a very important
research subject. However, the instrumental sensitivity in these photon energies has been in-
sufficient, and hence the existence of these excess components has remained controversial (e.g.
Bregman & Lloyd-Davies 2006; Nevalainen et al. 2007; Rossetti & Molendi 2004; Fusco-Femiano
et al. 2007).
Among those galaxy clusters from which the excess emission has been reported,
Abell 2199 is particularly interesting, because it has been suspected to exhibit both the soft and
hard excess components. In fact, Kaastra et al. (1999) reported the detection of both, based
on the BeppoSAX, Extreme Ultraviolet Explorer, and ROSAT observations. The two excess
components were simultaneously expressed by a single power law with a photon index ∼ 1.8
which is superposed on the thermal emission spectrum, and interpreted as inverse Compton
(IC) photons produced when the cosmic microwave background (CMB) photons are scattered
up by relativistic electrons in the cluster. However, as argued by Kempner & Sarazin (2000),
the magnetic field in Abell 2199 would have to be unusually weak (< 0.073 µG) in order for
the relativistic electrons not to produce synchrotron emission beyond measured upper limits
of diffuse radio emission. As an alternative interpretation, Kempner & Sarazin (2000) pro-
posed non-thermal bremsstrahlung by supra-thermal electrons which are being accelerated in
the cluster.
In the era of XMM-Newton, the possibility of thermal emission from a warm gas also
became another popular idea to explain the soft excess in several clusters. Kaastra et al. (2004)
reported a redshifted O VII emission line from a 0.2 keV warm gas in Abell 2199, along with
Coma, Abell 1795, Sersic 159-03, MKW 3s, Abell 2052, and Abell 3112, and hence ascribed
their soft excess to emission from WHIM. At the same time, Chandra has made progress on the
ICM physics of Abell 2199 especially in the central region, such as an asymmetric temperature
2

Page 3

distribution in the direction perpendicular to the jets within 30′′of the center (Kawano et al.
2003), and a weak isothermal shock associated with the central active galactic nucleus (Sanders
& Fabian 2006).
In order to examine the soft and hard excess phenomena of Abell 2199 with a better
sensitivity, we observed the object with Suzaku (Mitsuda et al. 2007). The excellent low-energy
capability of XIS-BI (back illuminated CCD, Koyama et al. 2007) and the low background of the
silicon PIN detectors in the HXD (Takahashi et al. 2007; Kokubun et al. 2007) have provided
the best data in the soft and hard X-ray bands, respectively. As a result, we have obtained
tight upper limits on both excess components, and strengthened the dominance of the thermal
emission from this cluster.
In the present paper, quantities quoted from the literature are rescaled to the Hubble
constant of H0= 71 km s−1Mpc−1, and errors are given at 90 % confidence range unless other-
wise stated.
2.Observation and Data Preparation
2.1.Observation
As shown in figure 1 and table 1, we observed Abell 2199 with Suzaku in the beginning of
October 2006. The observation was conducted in 5 separate pointings (“Center” and “Offset1”
through “Offset4”), for ∼ 20 ks each, with the XIS field of view (FOV) partially overlapping.
Throughout the 5 pointings, all the XIS sensors were in normal clock mode without window
or burst options. These observations became one of the first ones in which spaced-row charge
injection (SCI) was applied to the XIS detectors (Nakajima et al. 2008). In the present paper,
we use data of XIS0, XIS1, XIS3 and the HXD-PIN. Although XIS2 was still operational, we
do not use its data because a database of non X-ray background (NXB) for SCI-ON XIS2 is
too small, which is available only until the anomaly in XIS 2 on 2009 November 91.
Between the Offset2 and Offset4 observations, operations of changing the HXD-PIN
bias voltage were carried out in response to flare-like events in a PIN diode2(W10P0). As
summarized in table 2, data of all the 64 PINs were available throughout the Center and
Offset1 pointings, and for a part of Offset2 and Offset4. In contrast, only one fourth of the PIN
diodes were operational in Offset3. Therefore, we do not analyze the Offset3 HXD-PIN data in
the present paper. For the other observations, we exclude the data of the four PIN diodes in
W10 unit, and use the remaining 60 channels. Backgrounds and responses excluding the four
PIN diodes were prepared with a help of the HXD team.
1
http://www.astro.isas.ac.jp/suzaku/doc/suzakumemo-2007-08.pdf
2
http://www.astro.isas.jaxa.jp/suzaku/log/hxd/
3

Page 4

2.2.XIS data reduction
The present paper utilizes the XIS and HXD-PIN data, prepared through version 2.0
pipe-line data processing, and obtained from the Suzaku ftp site3.
screened XIS event files using Suzaku software version 11 in HEAsoft 6.6, together with
the calibration datebase (CALDB) released on 2009 January 9.
xisputpixelquality, xispi, and xistime in this order, and performed the standard
event screening.Bad pixels were rejected with cleansis, employing the option of chip-
col=SEGMENT. We selected events with GRADE 0,2,3,4, and 6.
(GTI) were detemined by criteria as "SAA_HXD==0 && T_SAA_HXD>436 && COR>6 && ELV>5
&& DYE_ELV>20 && AOCU_HK_CNT3_NML_P==1 && Sn_DTRATE<3 && ANG_DIST<1.5".
of the processing and screening are the same as those described in the Suzaku Data Reduction
Guide.4
The redistribution matrix files (RMFs) of the XIS were produced by xisrmfgen, and
auxillary responce files (ARFs) by xissimarfgen (Ishisaki et al. 2007). As an input to the
ARF generator, we prepared an X-ray surface brightness profile of Abell 2199 using a double-β
model (sum of two β models; King 1962) of which the parameters were determined through a
least chi-square fit to the background subtracted and vignetting corrected 0.5–10 keV XMM-
Newton MOS1 image of the Abell 2199 cluster. The employed parameters are (rc, β)=(10.3±0.5
kpc, 0.650±0.008) for the narrower component and (51.0±0.8 kpc, 0.531±0.002) for the wider
component, where rcis the core radius. Absorption below 2 keV, caused by a carbon-dominated
contamination material on the XIS optical blocking filters, is included in the ARFs. There,
differences of the contaminant thickness among the XIS sensors are taken into account, along
with its radial dependences and secular changes (Koyama et al. 2007).
Non X-ray background (NXB) of each XIS sensor was created using xisnxbgen, which
sorts spectra of night earth observations according to the geomagnetic cut-off-rigidity (COR)
and makes an averaged spectrum weighted by residence times for which Suzaku stayed in
each COR interval (Tawa et al. 2008). We assumed systematic errors (90% confidence range)
of 6.0% and 12.5% for the NXB from XIS-FI (front illuminated CCD, Koyama et al. 2007)
and XIS-BI (Tawa et al. 2008; table 5 (c)), respectively, and added them in quadrature to
the corrensponding statistical errors. The X-ray background was estimated as described in
subsection 3.1.
We re-processed un-
We applied xiscoord,
Good time intervals
Details
2.3. HXD-PIN data reduction
Unscreened event files of HXD-PIN were re-processed by hxdtime, hxdpi, and hxdgrade
in this order. Using hxdgtigen, we rejected the time intervals when “FIFO Full”, “BFSH”, or
“TLMRJCT” happened, and further narrowed the GTI by imposing criteria of "SAA_HXD==0 &&
3
ftp://ftp.darts.isas.jaxa.jp/pub/suzaku/ver2.0/
4
http://ftools.gsfc.nasa.gov/docs/suzaku/analysis/abc/
4

Page 5

T_SAA_HXD>500 && TN_SAA_HXD>180 && COR>6 && ELV>5 && AOCU_HK_CNT3_NML_P==1 &&
HXD_DTRATE<3 && ANG_DIST<1.5 && HXD_HV_Wn_CAL>700 && HXD_HV_Tn_CAL>700".
For the Center, Offset1, Offset2, and Offset4 pointings, we made special HXD-PIN RMFs
which exclude contributions of the four PINs in W10 unit. This was constructed by summing
RMFs from the remaining 60 PIN diodes. Since hxdarfgen did not support extended sources
yet, we created PIN ARFs for each 1′×1′pixel by hxdarfgen, assuming the same X-ray surface
brightness as used for XIS ARFs (subsection 2.2) . The ARFs were then averaged by weighting
with counts contained in the corresponding pixels.
As the HXD-PIN NXB, the “tuned” PIN NXB files published by the HXD team5were
used (Fukazawa et al. 2009). In the present paper, we used 2.3% as systematic error (90%
confidence range) of the PIN NXB. Details of its derivation is described in Appendix 1. The
PIN CXB was estimated from an observation of a nearby region in the same manner as the
XIS CXB (subsection 3.2).
3.Estimation of X-ray Backgrounds
3.1.XIS X-ray background
To analyze the XIS data of an extended object in general, we must subtract the Cosmic
X-ray background (CXB) and Galactic foreground emission (GFE). Since the emission from
Abell 2199 itself hampers their direct estimation using our own data, we used 73.6 ks Suzaku
observation of a nearby blank sky region named “High Latitude Diffuse A” (Hereafter HLD;
Observation ID = 500027010), which is 4◦.03 away from Abell 2199. The difference of their
Galactic latitudes is only 0◦.7, and hence the hydrogen column densities (Dickey & Lockman
1990, weighted average over 1◦cone radius) toward Abell 2199 (0.86×1020cm−2) and HLD
(1.02×1020cm−2) are similar.
After applying the standard processing described in subsection 2.2 to the unscreened
XIS data of HLD, we removed circular regions (1′in radius) around six point sources that are
visible in the 0.5–10.0 keV XIS images, and extracted spectra of the remaining region from the
four XIS detectors. The corresponding detection limit for the point sources is ∼ 6×10−14erg
s−1cm−2in 2–10 keV. Then, the NXB-subtracted XIS-BI and XIS-FI (averaged over the three
XIS-FI detectors) spectra in energy ranges of 0.3–5.5 keV and 0.5–5.5 keV, respectively, were
fitted simultaneously with a CXB+GFE model (see below) using XSPEC12 version 12.4.0ad,
incorporating the XIS response for a source of uniform brightness. In the analyses described
below, metal abundances refer to Anders & Grevesse (1989), and the photoelectric absorption
cross-sections to Balucinska-Church & McCammon (1992) with a new He cross section (Yan et
al. 1998).
As the CXB model, we used a power law with a fixed photon index of 1.41 (Kushino et
5
http://www.astro.isas.jaxa.jp/suzaku/analysis/hxd/pinnxb/tuned/
5