University of Wollongong University of Wollongong
Research Online Research Online
Faculty of Engineering - Papers (Archive)
Faculty of Engineering and Information
Sciences
December 2001
Discovery of Ghost Cavities in the X-Ray Atmosphere of Abell 2597 Discovery of Ghost Cavities in the X-Ray Atmosphere of Abell 2597
B. R. McNamara
Ohio University, USA
M. W. Wise
Massachusetts Institute of Technology, USA
P. Nulsen
University of Wollongong
L. P. David
Harvard-Smithsonian Center for Astrophysics, Cambridge, USA
C. L. Carilli
National Radio Astronomy Observatory, Socorro, USA
See next page for additional authors
Follow this and additional works at: https://ro.uow.edu.au/engpapers
Part of the Engineering Commons
https://ro.uow.edu.au/engpapers/300
Recommended Citation Recommended Citation
McNamara, B. R.; Wise, M. W.; Nulsen, P.; David, L. P.; Carilli, C. L.; Sarazin, C. L.; O'Dea, C. P.; Houck, J.;
Donahue, M.; Baum, S.; Voit, M.; O'Connell, R. W.; and Koekemoer, A.: Discovery of Ghost Cavities in the X-
Ray Atmosphere of Abell 2597 2001.
https://ro.uow.edu.au/engpapers/300
Research Online is the open access institutional repository for the University of Wollongong. For further information
contact the UOW Library: research-pubs@uow.edu.au
L149
The Astrophysical Journal, 562:L149–L152, 2001 December 1
䉷 2001. The American Astronomical Society. All rights reserved. Printed in U.S.A.
DISCOVERY OF GHOST CAVITIES IN THE X-RAY ATMOSPHERE OF ABELL 2597
B. R. McNamara,
1,2
M. W. Wise,
3
P. E. J. Nulsen,
4
L. P. David,
2
C. L. Carilli,
5
C. L. Sarazin,
6
C. P. O’Dea,
7
J. Houck,
3
M. Donahue,
7
S. Baum,
7
M. Voit,
7
R. W. O’Connell,
6
and A. Koekemoer
7
Received 2001 August 31; accepted 2001 October 29; published 2001 November 14
ABSTRACT
A Chandra image of the central 100 kpc of the Abell 2597 cluster of galaxies shows bright irregular X-ray
emission within the central dominant cluster galaxy (CDG) and two low surface brightness cavities located
30 kpc from the nucleus of the CDG. Unlike the cavities commonly seen in other clusters, the “ghost” cavities
in Abell 2597 are not coincident with the bright central radio source. Instead, they appear to be associated with
faint extended radio emission seen in a deep Very Large Array radio map. We interpret the ghost cavities as
buoyantly rising relics of a radio outburst that occurred between 50 and 100 Myr ago. The demography of cavities
in the few clusters studied thus far shows that galactic radio sources experience recurrent outbursts on an
∼100 Myr timescale. Over the lifetime of a cluster, ghost cavities emerging from CDGs deposit ⲏ10
59
–10
61
ergs
of energy into the intracluster medium. If a significant fraction of this energy is deposited as magnetic field, it
would account for the high field strengths in the cooling flow regions of clusters. The similarity between the
central cooling time of the keV gas and the radio cycling timescale suggests that feedback between cooling gas
and the radio source may be retarding or quenching the cooling flow.
Subject headings: cooling flows — galaxies: clusters: general — intergalactic medium —
radio continuum: galaxies — X-rays: galaxies: clusters
1. INTRODUCTION
Early Chandra images of galaxy clusters have shown that
the X-ray–emitting gas in their centers is bright and irregularly
structured and that much of this structure is associated with
powerful radio sources. The radio sources in the Hydra A (Mc-
Namara et al. 2000b), Perseus (Fabian et al. 2000), and Abell
2052 (Blanton et al. 2001) clusters appear to have pushed aside
the keV gas, leaving low surface brightness cavities in the gas.
The cavities in Hydra A, Perseus, and Abell 2052 are filled
with bright radio emission and are confined by the pressure of
the surrounding keV gas. The cavities may be supported against
collapse by pressure from relativistic particles, magnetic fields,
and/or hot, thin thermal gas. Since the cavities have a lower
gas density than their surroundings, they should behave like
bubbles in water and rise buoyantly in the intracluster medium
(ICM; McNamara et al. 2000b; Churazov et al. 2001).
Using simulations of supersonic jets expanding into the
ICM, Clarke, Harris, & Carilli (1997) and Heinz, Reynolds, &
Begelman (1998) argued that the cavities seen in ROSAT im-
ages of the ICM surrounding the Perseus and Cygnus A radio
sources (Heinz, Reynolds, & Begelman 1992; Carilli, Perley,
& Harris 1994) were caused by strong shocks. In this instance,
the X-ray emission from the rims surrounding the cavities
should be spectrally hard, and gas in the rim should have higher
entropy than the surrounding gas. The initial Chandra results
for the Hydra A (McNamara et al. 2000b; Nulsen et al. 2001),
1
Department of Physics and Astronomy, Ohio University, Athens, OH
45701.
2
Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cam-
bridge, MA 02138.
3
Center for Space Research, Massachusetts Institute of Technology, 70 Vas-
sar Street, Cambridge, MA 02139.
4
Engineering Physics, University of Wollongong, Wollongong NSW 2522,
Australia.
5
National Radio Astronomy Observatory, P.O. Box O, Socorro, NM 87801.
6
Department of Astronomy, University of Virginia, P.O. Box 3818, Char-
lottesville, VA 22903.
7
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD
21218.
Perseus (Fabian et al. 2000), and Abell 2052 (Blanton et al.
2001) clusters were surprising, as the emission from the rims
of the cavities was among the softest in the clusters. This im-
plies that the radio lobes expanded gently into the ICM at
roughly the sound speed in the keV gas (Reynolds, Heinz, &
Begelman 2001; David et al. 2000; Nulsen et al. 2001). The
rapidly growing number of cavities found in giant elliptical
galaxies (Finoguenov & Jones 2000), groups (e.g., Vrtilek et
al. 2000), and central dominant cluster galaxies (CDGs; e.g.,
Schindler et al. 2001) indicates that they are persistent features
of these systems.
An intriguing and potentially significant Chandra discovery
is the existence of cavities in the keV gas that do not have
bright radio counterparts. If such “ghost” cavities, as are seen
in Perseus (Fabian et al. 2000) and in Abell 2597 (discussed
here and in McNamara et al. 2000a), are generated by radio
sources, their properties would have significant consequences
for our understanding of the life cycles of radio galaxies and
the origin and dispersal of magnetic fields in clusters and gal-
axies. Here we discuss the remarkable properties of the X-ray
core of Abell 2597 and its ghost cavities. Throughout this
Letter, we assume , ,
⫺1 ⫺1
H p 70 km s Mpc Q p 0.3 Q p
0 M L
, a luminosity distance of 374 Mpc, and kpc.
0.7 1 p 1.67
2. OBSERVATIONS AND DATA REDUCTION
Abell 2597 is an ergs s
⫺1
(2–10 keV;
44
L p 6.45 # 10
X
David et al. 1993), richness class 0 cluster that lies at redshift
. The cluster possesses a bright cusp of X-ray emis-z p 0.083
sion associated with a cooling flow and the powerful radio
source PKS 2322⫺122, both centered on the CDG (Sarazin et
al. 1995).
A40ksChandra exposure was taken of Abell 2597 on 2000
July 28. The nucleus of the CDG was centered on node 0 of
the ACIS-S3 back-illuminated device. The pointing was chosen
to maximize the spatial resolution and soft energy response
available with Chandra while avoiding placing the interesting
central region on a node boundary. The observations were made
in faint, full-frame timed exposure mode, with the focal plane
L150 CAVITIES IN ABELL 2597 Vol. 562
Fig. 1.—Broadband-smoothed X-ray image of Abell 2597. The surface
brightness is irregular in the central 40⬙. The surface brightness depressions
associated with the cavities are seen 18⬙ to the southwest and northeast of
center. North is at the top; east is to the left.
Fig. 2.—Expanded view of the central region of Abell 2597 after subtracting
a smooth background cluster model. The 8.44 GHz radio contours are super-
posed. The cavities are seen as indentations in the bright emission. North is
at the top; east is to the left.
at a temperature of ⫺120⬚C. Grades 1, 5, and 7 were rejected
in our analysis, as were data below 0.3 keV and above 8.0 keV.
Unfortunately, the observations were compromised by strong
flares. As a consequence, only 18.54 ks of useful data were
gathered.
3. COMPARISON BETWEEN THE CENTRAL X-RAY AND
RADIO STRUCTURES
An adaptively smoothed X-ray image of the central
centered on the CDG is shown in Figure 1. The
150 # 150
passband of this image is from 0.3 to 8.0 keV. The image
immediately reveals the gross structure in the inner 1⬘ or so of
the cluster, first seen in an earlier ROSAT High Resolution
Imager observation (Sarazin et al. 1995). However, the fine
structure and particularly the surface brightness depressions
located 18⬙ to the southwest and 16⬙ to the northeast of the
center were not seen in the ROSAT image.
The central structure was isolated by modeling and subtracting
the smooth background cluster emission, leaving theexcessemis-
sion shown in Figure 2. This difference image was then adap-
tively smoothed, which revealed the structure shown above the
4 j significance level. In the bright regions, surface brightness
variations of 50%–60% are present. The large surface brightness
depressions to the northeast and southwest are 2–3 times fainter
than the surrounding regions, the depression to the southwest
being deeper. In order to examine the relationship between the
central radio source and the structure in the gas, we have super-
posed radio contours of Abell 2597 in Figure 2. The radio image
was obtained with the Very Large Array (VLA) A configuration,
tuned to a frequency of 8.44 GHz (Sarazin et al. 1995). The
radio source is relatively small and has a steep spectrum with a
spectral index of about equal to ⫺1.5 (O’Dea, Baum, & Gal-
limore 1994; Sarazin et al. 1995). Its full extent from north to
south is only 5⬙, or roughly 8 kpc. Although a great deal of
structure is seen surrounding the radio source, there is no evi-
dence for cavities there in spite of its powerful nature. However,
the central radio source is small compared to the bright X-ray
structure in the inner 40 kpc or so. Therefore, any cavities that
may exist would be difficult to detect owing to emission from
intervening material along the line of sight.
Unlike the cavities in Hydra A, cavities in Abell 2597 are
located at larger radii and are more than twice the size of the
central radio source. After subtracting the background cluster,
the cavities appear to be more extensive than the impression
given in Figure 1. The cavity to the northeast is roughly circular,
with a ⯝9⬙ diameter, corresponding to a linear diameter of
15 kpc. The cavity to the southwest is elliptically shaped with
major and minor axes ⯝14⬙ and ⯝9⬙, or 23 and 15 kpc linear
diameter, respectively. The cavities are surrounded by shells of
X-ray gas on most sides, but perhaps not at the outermost radii.
We divided the data into soft- and hard-band images with
0.5–1.5 and 1.5–3.5 keV passbands, respectively, and we ar-
rived at the following conclusions. To within the accuracy of
the data, (1) the emission immediately adjacent to each cavity
is generally no harder or softer than its surroundings, and
(2) deep surface brightness depressions are present in both
bands. Therefore, there is no compelling evidence for heating
by the agent that inflated the cavities, nor are the depressions
likely to be caused by absorption.
In order to determine whether faint radio emission is asso-
ciated with the cavities, radio observations of Abell 2597 were
made with the VLA at 1.4 GHz on 2001 June 21. The total
observing time was 12 hr, and the array configuration was
mixed between the 3 km and 10 km configurations, leading to
a synthesized beam of FWH with the major axis
M p 11 # 6
position angle of 90⬚. Standard amplitude and phase calibration
were applied, as well as self-calibration using sources in the
field. The dominant source in the field is the nucleus of Abell
2597 itself (Fig. 2), which has a peak surface brightness in our
1.4 GHz image of Jy beam
⫺1
. This bright source1.49 Ⳳ 0.03
limits the sensitivity of the final image to about 0.1 mJy beam
⫺1
No. 2, 2001 McNAMARA ET AL. L151
Fig. 3.—VLA 1.4 GHz image of Abell 2597 at resolution (oval
11 # 6
contours), with the major axis position angle of 90⬚. The first five contour
levels are a geometric progression in the square root of 2, with the first level
being 0.9 mJy beam
⫺1
. The higher contours are a geometric progression by
factors of 2. Boxes correspond roughly to the positions and sizes of the
X-ray cavities. The faint contours extending south of center are data artifacts.
Fig. 4.—Radial variation of (a) surface brightness, (b) electron density,
(c) temperature, and (d) pressure.
rms, implying a dynamic range of 15,000. The radio emission
from Abell 2597 is marginally resolved with these observa-
tions, with a total flux density of 1.82 Jy.
In the context of studying the X-ray cavities, the most in-
teresting results from the radio observations are the extensions
of the radio source in the vicinity of the cavities, as can be
seen in Figure 3. These extensions are robust in the imaging
process and are highly significant with respect to the noise on
the image. Unfortunately, the resolution of the image is in-
adequate to make any firm conclusions concerning the mor-
phology of the extended emission in the vicinity of the cavities,
only that such emission exists. This detection of radio emission
and a future confirmation with higher resolution are keys to
understanding the nature of ghost cavities (§ 5).
4. THE PHYSICAL STATE OF keV GAS
The radial distribution of surface brightness, temperature,elec-
tron density, and pressure in the central region of Abell 2597 is
shown in Figure 4. The profiles were extracted from annular
apertures centered on the weak nuclear point source coincident
with the radio core (R.A
h
25
m
19
s
.7, decl ⬚07⬘27⬙. p 23 . p 12
[J2000.0]). The aperture sizes were chosen to include roughly
1000 and 5000 counts for surface brightness and temperature
profiles shown in Figures 4a and 4c, respectively. The density
profile, Figure 4b, was constructed by deprojecting the surface
brightness profile assuming an emission measure appropriate for
a 3 keV gas. The temperature in each aperture was determined
by fitting an absorbed MEKAL single-temperature model in
XSPEC with abundances fixed at 0.4 solar and a Galactic fore-
ground column of . Only data from the
20 ⫺2
N p 2.48 # 10 cm
H
ACIS-S3 device were considered.
The temperature drops from 3.4 keV at 100 kpc to 1.3 keV
in the inner several kiloparsecs. Over this same region, the
gas density and pressure rise dramatically, reaching values of
0.07–0.08 cm
⫺3
and ergs cm
⫺3
, respectively, in the
⫺10
2.5 # 10
inner few kiloparsecs. The radiative cooling time of the keV gas
in the central 10 kpc region surrounding the radio source is only
≈ 10
8
yr. Similar properties are found in other cooling flow3 #
clusters, such as Hydra A (McNamara et al. 2000a, 2000b; David
et al. 2000), Abell 2052 (Blanton et al. 2001), and Perseus (Fa-
bian et al. 2000).
5. THE ORIGIN AND ENERGY CONTENT OF THE GHOST CAVITIES
Without internal pressure support, the cavities would collapse
on the sound crossing timescale of ∼10
7
yr. Yet the existence of
ghost cavities in Abell 2597 and in NGC 1275 (Fabian et al.
2000) beyond their radio lobes shows that they almost certainly
persist much longer than 10
7
yr and therefore must have pressure
support. The gas density within the ghost cavities is much less
than the ambient density. Therefore, they must be buoyant and
have risen outward from the nucleus to their current projected
radii of ⯝30 kpc. The time required for the cavities to rise to
this radius is roughly 10
7.7
–10
8
yr (McNamara et al. 2000b; Chur-
azov et al. 2001). This is much larger than the minimum age of
the central radio source of ∼ yr (Sarazin et al. 1995),
6
5 # 10
so the ghost cavities are unlikely to be related directly to the
current nuclear radio episode.
Based on the properties of radio-bright cavities, it is rea-
sonable to suppose that the ghost cavities were produced in a
radio episode that predated the current one shown in Figure 2.
Their radio emission has faded presumably because they are
no longer being supplied with relativistic particles from the
nucleus. This hypothesis is supported by our detection of ex-
tended radio emission toward the cavities and their locations
along nearly the same position angle as the central jets (Sarazin
et al. 1995). A rough estimate of the minimum energy pressure
in the extended, 1.4 GHz emission is considerably less than
the ambient pressure, as is found in other clusters. This would
suggest a departure from the minimum energy condition or a
dominant pressure contribution from low-energy electrons or
protons.
Assuming the cavities formed through the action of a radio
source, the lower limit to the energy expended during their
formation is given by the PdVwork done on the surrounding
