A&A 381, 841-847 (2002)
DOI: 10.1051/0004-6361:20011536
Yasuhiro Hashimoto1,2 - Günther Hasinger1,2 - Monique Arnaud3 - Piero Rosati4 - Takamitsu Miyaji5
1 - Astrophysikalisches Institut Potsdam, An der Sternwarte 16,
14482 Potsdam, Germany
2 -
Max-Planck-Institut für extraterrestrische Physik,
Giessenbachstrasse
85748 Garching, Germany
3 -
Service d'Astrophysique
CEA Saclay, 91191 Gif-sur-Yvette Cedex, France
4 -
European Southern Observatory, 85748 Garching, Germany
5 -
Department of Physics, Carnegie Mellon University,
5000 Forbes Ave., Pittsburgh, PA 15213, USA
Received 20 April 2001 / Accepted 29 October 2001
Abstract
We report on the XMM-Newton observation of RX J1053.7+5735,
one of the most distant
(z = 1.26) X-ray selected clusters of galaxies, which also shows an unusual double-lobed X-ray morphology, indicative of possible cluster-cluster interaction.
The cluster was discovered during our ROSAT deep pointings in the direction of the Lockman Hole.
The XMM-Newton observations were performed with the European Photon
Imaging Camera (EPIC) during the performance verification phase.
Total effective exposure time was 100 ksec.
The best fit temperature based on a simultaneous fit of spectra from
all EPIC cameras (pn+MOS) is
4.9+1.5-0.9 keV.
Metallicity is poorly constrained even using the joint fit of
all spectra, with an upper limit on the iron abundance of 0.62 solar.
Using the best fit model parameters, we derived an unabsorbed (0.2-10) keV
flux of
erg cm-2 s-1,
corresponding to a bolometric luminosity of
ergs-1.
Despite the fact that it was observed at fairly large off-axis angle,
the temperature errors are much smaller compared with
those of typical measurements based on ASCA or Beppo-Sax observations of
high-z (z > 0.6) clusters,
demonstrating the power of the XMM for
determining the X-ray temperature for high-z clusters.
The measured temperature and luminosity show that
one can easily reach the intrinsically X-ray
faint and cool cluster
regime comparable with those of
clusters observed by
past satellites.
The new cluster temperature and
we have measured for RX J1053.7+5735
is consistent with a weak/no evolution of the
-
relation out to
,
which lends support to a
low
universe,
although more data-points of z > 1 clusters are required for a more definitive statement.
The caution has to be also exercised
in interpreting the result, because of the uncertainty associated with
the dynamical status of this cluster.
Key words: galaxies: clusters: general - X-rays: galaxies - galaxies: evolution
Clusters of galaxies can be identified to high redshift and
can be used as tracers of the evolution of structure.
Since the evolution of structure depends on parameters such as
,
,
& P(k) in hierarchical cosmologies,
the study of the
evolution of galaxy clustering provides us with an important constraint on
the cosmology (e.g. Press & Schechter 1974; Peebles et al. 1989; Eke et al.
1996).
X-ray observations provide a powerful and unique means of selecting the clusters and
characterizing their properties, as
X-ray flux is
proportional to the square
of the electron density, and therefore
less affected
than optical data
by the superposed structures.
Measurements of the gas temperatures in high redshift clusters of
galaxies strongly
constrain cosmological models because cluster temperatures are closely
related to cluster masses,
and the evolution of the cluster mass function
with redshift is quite sensitive
to cosmological parameters.
X-ray cluster surveys based on ROSAT-PSPC
data (e.g. Rosati et al. 1995, 1998; Scharf et al. 1997; Burke et al. 1997;
Vikhlinin et al. 1998)
detect sizable samples of distant clusters (z > 0.5).
Although the number of high-z clusters has
significantly increased,
intracluster medium (ICM) properties
(such as X-ray temperature: )
at high redshift
are still largely unexplored. This is due to the limited
effective area and spatial resolution of the past X-ray missions;
only long ASCA/Beppo-SAX observations,
with a broad energy response, have
allowed studies of the high-z cluster ICM properties.
Only a few high-z (z > 0.6) clusters have
directly measured X-ray temperatures
(e.g. Donahue et al. 1999; Della Ceca et al. 2000).
Moreover, these high-z samples are inevitably biased toward
intrinsically X-ray bright, and thus high temperature clusters,
which complicates the
investigation of ICM evolution.
Advent of Chandra, with its
bigger effective area and higher angular resolution than the past satellites,
made us possible to effectively study the ICM properties of high-z clusters.
Chandra's high angular resolution, in particular,
greatly helps us to avoid confusion
with point sources (e.g. Stanford et al. 2001; Borgani et al. 2001;
Jeltema et al. 2001; Cagnoni et al. 2001).
Unfortunately, for those with directly measured ,
the errors on
are
still relatively large.
Moreover, Chandra is not particularly sensitive to
low surface brightness emission, which is sometimes unfavorable
for the analysis of faint X-ray features often present in high-z clusters.
The XMM-Newton (XMM) pn and MOS CCD cameras
have 10 times larger
effective area than ASCA GIS/SIS & Beppo-SAX MECS
(in addition to the fact that it has significantly
sharper PSF than ASCA & Beppo-Sax).
Even compared with Chandra ACIS, the effective area
is more than 5 times larger
which helps to
increase the accuracy of the
determination per given
exposure time.
With its high throughput and moderate (compared with Chandra)
angular resolution, XMM is extremely sensitive to low surface brightness
X-ray emission.
These properties make
XMM pn/MOS very suitable instruments
to investigate the ICM properties of distant clusters.
In this paper, we report on the XMM observation of one of the most distant X-ray selected clusters (z = 1.26) of galaxies, which also shows an unusual double-lobed X-ray morphology. This cluster, RX J1053.7+5735, was discovered during our deep ROSAT HRI pointings in the direction of the "Lockman Hole", a line of sight with exceptionally low HI column density (Hasinger et al. 1998a). RX J1053.7+5735 is one of the most distant clusters ever selected by diffuse X-ray emission (for another most distant X-ray selected cluster at z=1.26, RX J0848.9+4452, see Rosati et al. 1999; Stanford et al. 2001). Together with the fact that its X-ray morphology is clearly double-lobed and highly anisotropic, which may well be a sign of a cluster in the making, the RX J1053.7+5735 may provide us with unique information to better understand the evolution of clusters and galaxies in the early universe. Throughout the paper, we use H0 = 100 h kms-1 Mpc-1and q0 = 0.5.
![]() |
Figure 1: The ROSAT HRI contours of the cluster RX J1053.7+5735 overlaid on the color image made from V, R, and I (color coded in blue, green, and red) band exposures. North is up and East is left. |
Open with DEXTER |
Light curves were visually inspected,
and time intervals with high background (0.5-10 keV count rate higher than 8 cts/s for the pn
and 3 cts/s for MOS) were excluded.
We used both single and double events.
The remaining effective exposure times for each of three detectors were
approximately 100 ksec.
The astrometries calculated from the WCS keywords were offset by
5
-25
from the known optical counterparts. Thus, a transversal
shift and
a rotation angle were fit for each dataset, leading to residual systematic
position errors of 1
-2
.
Figure 2 shows the exposure corrected XMM image of RX J1053.7+5735
in the 0.5-2 keV band, an energy range where the bulk of the redshifted
cluster emission
falls upon and therefore the contrast against the background is the strongest.
The image was created by combining all events from three
(pn, MOS1, and MOS2) cameras.
The raw data were smoothed with a Gaussian with
= 7
.
The lowest contour is 2
above the background, and
the contour interval is 0.5
.
A combined exposure map was calculated for the pn plus MOS cameras.
Figure 2 clearly shows that the cluster emission is extended and double-lobed, consistent with the ROSAT HRI image, and its shape is indicative of non-regularity of this cluster. Unfortunately, the pointing variation between the different revolutions was not large enough to clearly remove the effect of inter-CCD gaps, which happen to run through the cluster, particularly around the eastern lobe, and this somewhat distorted the level and shape of the eastern-lobe contours.
![]() |
Figure 2:
The exposure corrected XMM image of the cluster RX J1053.7+5735
in the 0.5-2 keV band.
The image was created by combining all events from three
(pn, MOS1, and MOS2) cameras.
North is up and East is left.
The raw data were smoothed with a Gaussian with ![]() ![]() ![]() ![]() |
Open with DEXTER |
The spectra were extracted from an elliptical region centered at the
cluster image with a semi-major/minor axis 0
9 and 0
6, respectively
(corresponding to 0.23 and 0.16 Mpc/h at z= 1.263).
The backgrounds were estimated from a region of the detector
at the same off-axis angle
surrounding the cluster, after removing point sources.
We obtain
and
net counts (in the 0.2-10.0 keV band)
for pn and MOS(1+2), respectively.
We regrouped the counts in order to have a
in each bin after the
background subtraction.
We fitted the spectra with a
thermal plasma emission model from Raymond & Smith (1977) using the
XSPEC 11.0 with respective response matrices (Haberl 2001).
For the MOS cameras, two filters were used alternately, thus we summed
all the MOS spectra according to the used filter and fit two sets of spectra
jointly.
We fixed a characteristic column density of neutral
hydrogen
at a value of
cm-2 and redshift at z = 1.263.
For the case of a joint fit,
each spectrum was fitted with its own normalization, but
with common
temperature (and metallicity, if applicable).
![]() |
Figure 3: Rebinned spectra and best-fit models for cluster RX J1053.7+5735 with pn a) and MOS1+2 b) cameras. The crosses are the observed spectrum and the solid line denotes best-fit model. |
Open with DEXTER |
![]() |
Figure 4:
Two-dimensional ![]() ![]() |
Open with DEXTER |
![]() |
Figure 5:
a)
New XMM temperature and its 2![]() ![]() ![]() ![]() |
Open with DEXTER |
Detector | kTa | Abundance | ![]() |
(keV) | (![]() |
||
pn | 4.8 (3.7-6.3) | 0.3 (fixed) | 0.86 (43) |
MOS1+2 | 4.9 (3.6-7.9) | 0.3 (fixed) | 0.78 (29) |
ALL | 4.9 (4.0-6.4) | <0.62 | 0.83 (72) |
Lobe | kT | Lbol | Abundance | ![]() |
(keV) | (1044 erg/s) | (![]() |
||
East | 3.7 (2.7-4.9) | 1.5 | 0.3 (fixed) | 1.03 (30) |
West | 6.2 (4.4-10.4) | 1.9 | 0.3 (fixed) | 0.82 (43) |
E+W | 4.9 (4.0-6.4) | 3.4 | <0.62 | 0.83 (72) |
In Fig. 3, and Table 1, the results of the best-fits are shown.
Figure 4 shows the two-parameter
contours for the cluster metallicity
and
.
The best fit temperature based on a simultaneous fit of all (pn+MOS)
spectra is
4.9+1.5-0.9 keV.
All temperature uncertainties quoted are at the 90% confidence levels for a
one-dimensional fit (
).
No prominent iron K
complex is visible at
3 keV
(redshifted 6.7 keV). At
keV, and the cluster redshift, only
the K
is strong enough to constrain the abundance
with our spectrum.
As a result, metallicity is poorly constrained even using the joint fit of
all spectra,
with an upper limit on the single iron abundance of 0.62 solar
(with 68% confidence).
Despite the fact that the cluster was observed at fairly large off-axis angle,
the temperature errors are much smaller compared with
those of typical
measurements based on ASCA or Beppo-Sax observations of
high-z (z > 0.6) clusters (Fig. 5a),
and the errors
are comparable with typically "on-axis'' Chandra
observations of
clusters (Fig. 5b),
demonstrating the power of XMM for
determining the X-ray temperature for high-z clusters.
Moreover, the measured temperature and luminosity show that
one can easily reach the intrinsically X-ray
faint and cool cluster
regime comparable with those of
clusters observed by
past satellites, which enables the investigation of the evolution
of various cluster X-ray properties
without additional
evolutionary assumptions.
Vignetting correction is estimated by using the average vignetting model
between 0.5-7 keV, also averaged over the spectral extraction region.
Using the best fit model parameters, we derived an
unabsorbed (0.2-10) keV
flux of
erg cm -2 s-1,
corresponding to a
luminosity in the cluster rest frame of
ergs-1 and a bolometric luminosity of
ergs-1.
Due to the still preliminary status of the
current EPIC calibration, and some simplification about the vignetting
correction, we estimate systematic flux errors
on the order of 10%.
Temperature and luminosity are also estimated for each (eastern and western)
lobe of RX J1053.7+5735, using a circular (radius = 0
47) extracting region centered at each lobe.
Our fitting for each lobe using all data (pn+MOS) shows that
is
3.7 +1.2-1.0 keV and
6.2 +4.2-1.8 keV, while
is
ergs-1and
ergs-1, for eastern and western
lobes, respectively (Table 2).
For any other regions, such as the bridge between the two lobes,
the counts are less than 10% of total cluster counts and not sufficient
for any reasonable
investigations.
![]() |
Figure 6:
Newly measured X-ray properties for a high-z (z = 1.263)
cluster, RX J1053.7+5735
(the square symbol),
is compared
with other clusters at various redshifts.
The error on the temperature represents the 90% confidence range.
The error on the RX J1053.7+5735 luminosity is 10%.
The eastern and western lobes of RX J1053.7+5735 (denoted as "East'' and "West'')
are also shown.
The solid line is the
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Open with DEXTER |
Clusters exhibit a correlation between their X-ray
luminosity ()
and temperature (
). This correlation is often
used to compare gas and dark matter contents in clusters, as
is
connected with the gas while
is related
to the total gravitating mass of the cluster.
The cluster
-
relationship therefore,
can be used to examine the evolution of clusters and should be
able to tell the range of redshifts where the cluster formation
has taken place (e.g. Scharf & Mushotzky 1997).
At low redshift (z < 0.1), the -
relation is well measured
(e.g. Mushotzky 1984; Edge & Stewart 1991; David et al. 1993)
and expressed via power law:
,
with
-2.9 (Markevitch 1998; Allen & Fabian 1998;
Jones & Forman 1999; Arnaud & Evrard 1999).
At higher redshifts, however,
no significant difference from the low redshift relations are
reported, despite the fact that
a relatively large evolution of the -
relation is required in order to account,
within the frame work of the critical universe,
for the weak
evolution observed in the cluster X-ray luminosity function (XLF)
out to
(Rosati et al. 1998).
Mushotzky & Scharf (1997) (see also Henry 2000),
by comparing the clusters at z < 0.1 with those at z > 0.1,
found no convincing evidence for a significant evolution
of
-
relationship out to
-0.5,
which is consistent with recent result of
Wu et al. (1999) where they used the
largest sample of clusters to date from the literature, and estimated
.
For even higher redshifts,
Donahue et al. (1999),
using a complete sample of (five) high redshift (z > 0.5) EMSS clusters,
have detected no significant evolution up to
(see also Della Ceca et al. 2000; Borgani et al. 2001).
In Fig. 6, we compare our newly measured X-ray properties for
RX J1053.7+5735 with other clusters at
various redshifts from the literature
(Markevitch 1998; Mushotzky & Scharf 1997; Donahue et al. 1999; Della Ceca et al.
2000; Borgani et al. 2001).
The square denotes the new
and
for RX J1053.7+5735.
The error on the temperature represents the one dimensional
90% confidence range.
The error bar on the RX J1053.7+5735 luminosity is 10%.
The eastern and western lobes of RX J1053.7+5735 (denoted as "East'' and "West'')
are also separately shown.
The solid line is the
-
relation of
from Wu et al. (1999) together with
2
lines (dotted lines).
The dashed line represents the evolving
-
relation
in the form of
(1+z)2 at z = 1.263, which would be required to
make the observed weak-evolving XLF to be consistent with a
universe
(Borgani et al. 1999).
The circles, triangle, & stars denote low redshift (
0.1),
intermediate redshift (
), and high redshift
(z > 0.5) clusters, respectively.
Total cluster luminosity may be somewhat underestimated because of
the fundamental difficulty in measuring the
luminosity of an extended source using a finite-sized extraction region.
However,
the treatments (if any)
against the missing flux outside the extraction region
often require difficult extrapolation and are not uniform
for the low-z cluster
samples, or other high-z samples in the literature.
We therefore chose not to include the correction associated with the
missing flux.
Nevertheless, we crudely estimate the effect in our cluster
using a profile from Neumann & Arnaud (1999). We find that
the uncertainty in
is on the order of 10%, which is
similar with the size of error bar shown in Fig. 6.
The new cluster temperature and
we have measured for RX J1053.7+5735
is in agreement with other high-z
-
analyses (e.g. Donahue et al. 1999; Della Ceca et al. 2000; Borgani et al. 2001).
The dynamical state of high-z clusters provides us with valuable and direct information on the formation and evolution of clusters. In hierarchical models, structures form from the bottom-up and thus, for many clusters, "formation'' means an ongoing sequence of mergers and interactions with other clusters and groups. X-ray signatures of cluster interaction include surface brightness anomalies (e.g. non-symmetric, non-isotropic X-ray isophotes; substructures; elongate X-ray core; an X-ray peak offset from the peak of the galaxy distribution), and non-isothermal and asymmetric temperature distributions. Evidence for such cluster substructures and merger events in the local universe have been obtained with ROSAT and ASCA satellite (see Buote 2001 for a review). For higher redshifts (z > 0.5), the majority of clusters show clearly distorted X-ray morphology (e.g. Neumann & Böhringer 1997; Gioia et al. 1999; Della Ceca et al. 2000; Neumann & Arnaud 2000; Stanford et al. 2001; Jeltema et al. 2001), suggesting that they are in an unrelaxed state.
The fact that the X-ray morphology of RX J1053.7+5735 is
double-lobed suggests that we may be seeing a merger event at
,
although our optical/NIR data are currently insufficient for
a definitive statement about the dynamical state of this cluster.
As for the
distribution,
Fig. 7 shows the contour plot of an image in the 1-2 keV band overlaid on the
gray-scale image in the 0.5-1 keV band. The figure shows some hint of
a difference in the spatial distribution
of the 1-2 keV image contour with respect to the 0.5-1 keV map,
which may be interpreted as a
variation.
The western lobe also shows a hint of higher
,
however, it
is statistically inconclusive (Table 2).
Future XMM (or Chandra) data of this cluster are
needed, as well as more optical/NIR data,
to improve our knowledge of the dynamical state of this cluster.
![]() |
Figure 7:
The contour plot of 1-2 keV image overlaid on the
gray-scale 0.5-1 keV image.
Both images are smoothed with a Gaussian with
![]() |
Open with DEXTER |
The new cluster temperature and
we have measured for
RX J1053.7+5735 is consistent with a weak/no evolution of the
-
relation out to
,
whereas a strong evolution of the
-
relation is required to make the observed weak-evolving XLF to be
consistent with a
universe. Thus, our results could be
interpreted as a support for a low
universe, although
more data-points of z > 1 clusters are needed for a more definitive
statement.
The caution has to be also exercised in interpreting the result,
because of the uncertainty associated with the dynamical state of the
cluster. In fact, our results may emphasize the fundamental difficulty in
constraining the cosmology from the evolution of the XLF and the
-
relation, since the baseline of the comparison, the local
-
relation and the theoretical M-T relationships, are essentially valid
for dynamically relaxed clusters (see also Ricker & Sarazin 2001).
If the majority of high-z clusters are in the unrelaxed state,
the impact of dynamically unrelaxed cluster states has to be carefully assessed.
In particular, during
mergers, clusters are expected to follow a complex track in the
-
plane as shown in the numerical simulations (Ricker & Sarazin 2001),
and therefore,
the relatively low luminosity of
RX J1053.7+5735 in view of its temperature may be interpreted
as further evidence for the unrelaxed state of the cluster.
Acknowledgements
We thank Pat Henry for a careful reading of the paper and useful comments. We also thank S. Majerowicz for his help in analyzing the MOS spectra. We acknowledge the referee's comments which improved the manuscript. Part of the work was supported by the Deutsches Zentrum für Luft- und Raumfahrt DLR project numbers 50 OX 9801 and 50 OR 9908.