Physics
Physics Research Publications
Purdue University Year
Chandra X-ray spectroscopic imaging of
Sagittarius A* and the central parsec of
the Galaxy
F. K. Baganoff, Y. Maeda, M. Morris, M. W. Bautz, W. N. Brandt, W. Cui, J.
P. Doty, E. D. Feigelson, G. P. Garmire, S. H. Pravdo, G. R. Ricker, and L. K.
Townsley
This paper is posted at Purdue e-Pubs.
http://docs.lib.purdue.edu/physics articles/368
CHANDRA X-RAY SPECTROSCOPIC IMAGING OF SAGITTARIUS A*
AND THE CENTRAL PARSEC OF THE GALAXY
F. K. Baganoff,
1
Y. Maeda,
2
M. Morris,
3
M. W. Bautz,
1
W. N. Brandt,
4
W. Cui,
5
J. P. Doty,
1
E. D. Feigelson,
4
G. P. Garmire,
4
S. H. Pravdo,
6
G. R. Ricker,
1
and L. K. Townsley
4
Received 2001 February 2; accepted 2003 February 28
ABSTRACT
We report the results of the first-epoch observation with the ACIS-I instrument on the Chandra X-Ray
Observatory of Sagittarius A* (Sgr A*), the compact radio source associated with the supermassive black
hole (SMBH) at the dynamical center of the Milky Way. This observation produced the first X-ray (0.5–
7 keV) spectroscopic image with arcsecond resol ution of the central 17
0
17
0
(40 pc 40 pc) of the Galaxy.
We report the discovery of an X-ray source, CXOGC J174540.0290027, coincident with Sgr A* within
0>27 0>18. The probability of a false match is estimated to be d0.5%. The spectrum is well fitted either by
an absorbed power law with photon index 2:7 or by an absorbed optically thin thermal plasma with
kT 1:9 keV and co lumn density N
H
1 10
23
cm
2
. The observed flux in the 2–10 keV band is
1:3 10
13
ergs cm
2
s
1
, and the absorption-corrected luminosity is 2:4 10
33
ergs s
1
. The X-ray
emission at the position of Sgr A* is extended, with an intrinsic size of 1>4 (FWHM), consistent with the
Bondi accretion radius for a 2:6 10
6
M
black hole. A compact component within the source flared by up
to a factor of 3 over a period of 1 hr at the star t of the observation. The search for K line emission from
iron was inconclusive, yielding an upper lim it on the equivalent width of 2.2 keV. Several potential stellar
origins for the X-ray emission at Sgr A* are considered, but we conclude that the various properties of the
source favor accretion onto the SMBH as the origin for the bulk of the emission. These data are inconsistent
with ‘‘ standard ’’ advection-dominated accretion flow (ADAF) models or Bondi models, unless the accretion
rate from stellar winds is much lower than anticipated. The central parsec of the Galaxy contains an 1.3
keV plasma with electron density n
e
26
1=2
f
cm
3
, where
f
is the filling factor. This plasma should supply
10
6
M
yr
1
of material to the accretion flow at the Bondi radius, whereas measurements of linear polar-
ization at 150 GHz and above limit the accretion rate near the event horizon to d10
8
M
yr
1
, assuming an
equipartition magnetic field. Taken together, the X-ray and radio results imply that outflows or convection
are playing a role in ADAF models and subequipar tition magnetic fields in Bondi models, or else the X-ray
emission must be generated predominantly via the synchrotron self-Compton (SSC) process. The measured
extent of the source and the detection of short timescale variability are evidence that the emission from Sgr
A* contains both thermal and nonthermal emission components at comparable levels. We also discuss the
complex structure of the X-ray emission from the Sgr A radio complex and along the Galactic plane. Mor-
phological evidence is presented that Sgr A* and the H ii region Sgr A West lie within the hot plasma in the
central cavity of Sgr A East, whi ch we interpret as a supernova remnant that may have passed through the
position of the SMBH, leading to a period of increased activity that ended within the past 300 yr. Similarly,
we have discovered bright clumps of X-ray emission located on opposite sides of the Galactic plane, along a
line passing through the central parsec of the Galaxy. The arrangement of these lobes suggests that Sgr A*
may have experienced an earlier period of increased activity lasting several thousand years during which it
expelled hot gas in a bipolar outflow oriented roughly perpendicular to the Galactic plane. Additionally, we
present an analysis of stellar emission within the central parsec of the Galaxy.
Subject headings: accretion, accretion disks — black hole physics — galaxies: active — Galaxy: center —
X-rays: ISM — X-rays: stars
1. INTRODUCTION
After decades of controversy, measurements of stellar
dynamics have confidently established that the nucleus of
the Milky Way harbors a supermassive black hole (SMBH)
with a mass M 2:6 10
6
M
(Genzel et al. 1997, 2000;
Ghez et al. 1998, 2000; Scho
¨
del et al. 2002). The SMBH
coincides with the compact nonthermal radio source
Sagittarius A* (Sgr A*), but no emission at other wave-
lengths has been convincingly associated with it (x 2.1). It is
also well known that the bolometric luminosity (L) and the
X-ray lumi nosity (L
X
) of Sgr A* are far lower than expected
from the standard thin accretion disk model used in the
study of X-ray binaries and quasars (Shakura & Sunyaev
1973; Watson et al. 1981; Bradt & McClintock 1983; Frank,
King, & Raine 1992; Morris & Serabyn 1996 and references
therein). The bolometric luminosity of a 2:6 10
6
M
black
hole radiating at the Eddington rate (L
E
)is3 10
44
ergs
s
1
, while the measured bolometric luminosity of Sgr A* is
d10
37
ergs s
1
(see Narayan et al. 1998a and references
1
Center for Space Research, Massachusetts Institute of Technology,
Cambridge, MA 02139-4307; fkb@space.mit.edu.
2
Institute of Space and Astronautical Science, 3-1-1 Yoshinodai,
Sagamihara 229-8501, Japan.
3
Department of Physics and Astronomy, University of California, Los
Angeles, CA 90095-1562.
4
Department of Astronomy and Astrophysics, Pennsylvania State
University, University Park, PA 16802-6305.
5
Department of Physics, Purdue University, West Lafayette, IN 47907.
6
Jet Propulsion Laboratory, California Institute of Technology,
Pasadena, CA 91109.
The Astrophysical Journal, 591:891–915, 2003 July 10
# 2003. The American Astronomical Society. All rights reserved. Printed in U.S.A.
891
therein). In the standard model, 10% of the luminosity is
in X-rays (Frank et al. 1992), so one would expect
L
X
3 10
43
ergs s
1
if Sgr A* were radiating at the
Eddington rate. The SMBH at Sgr A* has been undetected
in the 2–10 keV band with L
X
< 10
35
ergs s
1
(x 2.2), which
is 10
9
times fainter than the X-ray luminosity that would
be expected at the Eddington rate. Similarly, the 10
6
M
SMBHs at the cores of several nearby spiral galaxies emit
L
X
’ 10
37
–10
39
ergs s
1
, implying that they are 10
5
–10
7
times fainter in X-rays than woul d be expected at their
Eddington rates (Garcia et al. 2000; Terashima, Ho, & Ptak
2000; Ho et al. 2001).
The absence of a strong, compact X-ray source associated
with the SMBH at the Galactic center has been one of the
profound mysteries of high-energy astrophysics and must
have at least one of three basic causes. First, the SMBH may
reside in an environment where the accretion rate
_
MM5
_
MM
E
¼ L
E
=ð
r
c
2
Þ’5:8 10
3
1
r
M
yr
1
, either
because the ambient gas has extremely low density, because
it is too hot or is moving too fast to accrete efficiently, or
because it is dynamically ejected prior to accretion. Here
r
¼ L=ð
_
MMc
2
Þis the radiative efficiency of the accretion flow
and c is the speed of light. Second, the mechanism of accre-
tion may be such that the radiative efficiency is extremely
low. The advection-dominated accretion flow (ADAF;
Narayan, Mahadevan, & Quataert 1998b) and related
models can achieve low values of
r
and hav e been
intensively applied to the Sgr A* problem. Third, the X-ray
emission from Sgr A* may be much higher than observed
because of anisotropy (e.g., a relativistic beam oriented per-
pendicular to the Galactic plane) and/or extremely high
absorption along the line of sight.
The Chandra X-Ray Observatory (CXO; Weisskopf et al.
1996), with its Advanced Charge-Coupled Device Imaging
Spectrometer (ACIS; G. P. Garmire, J. P. Nousek, & M. W.
Bautz, in preparation), provides a unique opportunity to
advance our knowledge of X-ray emission from Sgr A*. It
combines a mirror with subarcsecond resolution and an
imaging detector with high efficiency over a broad energy
band and moderate spectral resolution. The spatial resolu-
tion and accurate astrometry are essential to discriminate
the emission of Sgr A* from that produced in the surround-
ing compact cluster of massive stars and hot plasma in the
region. The sensitivity of Chandra/ACIS at energies above
2 keV is essential to penetrate the high interstellar absorbing
column along the line of sight to the Galactic center ( x 2.1).
The central black hole of our Galaxy sits amid a complex
of X-ray–emitting and absorbing entities, which compli-
cates the analysis of the emission ascribable to Sgr A* but
also provides a significant bonus in terms of the science that
can be extracted from a single pointing with ACIS. In the
immediate vicinity of Sgr A* lies a cluster of luminous
young stars, many of which are windy emission-line stars
that presumably supply the matter which the black hole is
presently accreting (Krabbe et al. 1995; Blum, Sellgren, &
DePoy 1996; Najarro et al. 1997; Coker & Meli a 1997;
Quataert, Narayan, & Reid 1999b; Paumard et al. 2001).
These stars and their colliding winds are themselves poten-
tial X-ray sources (Ozernoy, Genzel, & Usov 1997). This
cluster excites a parsec-scale H ii region, Sgr A West, that is
well studied at radio and infrared wavelengths. Surrounding
Sgr A West is the dense, predominantly neutral circumnu-
clear disk (Morris & Serabyn 1996 and references therein).
Beyond that, on a scale of 10 pc, the nonthermal radio
shell source Sgr A East surrounds the circumnuclear disk in
projection. In a separate study (Maeda et al. 2002), we have
interpreted Sgr A East—a strong X-ray source—as a super-
nova remnant. Figure 1 diagrams the relative placement of
these structures.
After a review of some relevant past studies (x 2), we
describe the observations, data analysis, source detection,
and astrometry (x 3). The resulting image of the inner
40 pc 40 pc (17
0
17
0
) of the Galaxy is presented in x 4,
and the properties of the innermost arcsecond associated
with Sgr A* are described in x 5. The integrated emission
from stellar sources (x 6) and the diffuse emission ( x 7)
within the central 10
00
of the Galaxy are discussed. Tentative
identifications of discrete X-r ay sources in the central 10
00
with bright IR sources and the effects of source confusion
on observations by other X-ray satellites are presented in
x 6. The origin of the X-ray emission at Sgr A* (i.e., SMBH
vs. stellar) is discussed (x 8) and implications for the astro-
physics of accretion onto the central SMBH of the Galaxy
are presented (x9). We summarize our findings, evaluate the
various models, and discuss the scientific goals of future
observations in x 10. This is one of several papers arising
from this Chandra observation. Future papers will present
our studies of the X-ray emission from the point sources
and the diffuse plasma distributed throughout the field.
60"
2 pc
Sgr A*
of Sgr A West
Northern Arm
Circumnuclear
Disk
Sgr A West
Sgr A East
Galactic plane orientation
Fig. 1.—Schematic diagram of the principle constituents of the Sgr A
radio complex, showing the previously known features discussed in this
paper. The outer ellipse depicts the location of the radio shell of Sgr A East,
which is a filled-center structure in X-rays (the red region in Fig. 2, and the
yellow and green region on the left side of the central structure in Fig. 3).
The circumnuclear disk, which is not evident in the X-ray maps, may affect
the morphology of the X-ray–emitting region (see Fig. 4). The black hole
candidate, Sgr A*, lies at the center of a cavity surrounded by the circumnu-
clear disk. The ionized gas features of Sgr A West and the central cluster of
luminous, hot stars also lie within this cavity.
892 BAGANOFF ET AL. Vol. 591
2. PAST STUDIES
2.1. Radio/IR
Sgr A* is a compact, nonthermal radio source (Balick &
Brown 1974; Backer 1996). Radio proper motion studies
performed over the last decade place Sgr A* at the dynami-
cal center of the Galaxy, and set a lower limit on its mass of
2 10
4
M
(Backer & Sramek 1999; Reid et al. 1999). It has
an intrinsic radio brightness temperature e10
10
K (Backer
et al. 1993; Rogers et al. 1994), and is weakly variable on
timescales of less than about a month in the centimeter and
millimeter bands (Zhao 1989; Wright & Backer 1993;
Falcke 1999; Tsuboi, Miyazaki, & Tsutsumi 1999; Zhao,
Bower, & Goss 2001). These properties are reminiscent of
the compact nuclear radio sources present in radio-loud
quasars and active galactic nuclei (AGN) and suggest that
Sgr A* may derive its luminosity from matter accreting onto
the SMBH at the center of the Galaxy (Lynden-Bell & Rees
1971).
Polarimetric and spectropolarimetric observations made
with the Very Large Array (VLA) and the Berkeley-Illinois-
Maryland-Association (BIMA) radio interferometers show
that Sgr A* is linearly unpolarized at frequencies up to 112
GHz (Bower et al. 1999a, 1999c, 2001); the upper-limit on
linear polarization at 112 GHz is 1.8%. Aitken et al. (2000)
report the detection of linear polarization from Sgr A* at
750, 850, 1350, and 2000 lm with the SCUBA camera on
the 15 m James Clerk Maxwell Telescope (JCMT). After re-
moving the effects of strong free-free emission and polarized
dust from the single-dish JCMT beam (34
00
at 150 GHz),
they found that the fractional linear polarization at 2000 lm
(150 GHz) is 10
þ9
4
%, and that it increases with frequency.
This result has recently been confirmed by Bower et al.
(2003) using higher angular resol ution observations with
the BIMA array.
Circular polarization of Sgr A* has been detected at 1.4
to 14.9 GHz with the VLA (Bower, Falcke, & Backer
1999b) and the Australia Telescope Compact Array
(ATCA; Bower et al. 2002). The fractional circular
polarization at 4.8 GHz is 0:37% 0:04%. The circular
polarization at 4.8 GHz is confirmed independently by Sault
& Macquart (1999) with the ATCA.
The total radio luminosity of Sgr A* is estimated to be a
few hundr ed L
(Morris & Serabyn 1996). This raises the
possibility that the emission could result from accretion
onto a cluster of compact stellar-mass objects (Ozernoy
1989; Morris 1993). However, recent proper motion studies
of stars within 6
00
of the Galactic center constrain the mini-
mum mass density of the central gravitational potential to
be e10
12
M
pc
3
(Eckart & Genzel 1997; Ghez et al.
1998). The best-fit model from Ghez et al. requires a dark
central object of mass M ¼ð2:6 0:2Þ10
6
M
within
0.015 pc of Sgr A* (see also Genzel et al. 2000). These
results rule out a cluster of compact stellar-mass objects as
the energy source for Sgr A* (see Maoz 1998) but provide
no direct evidence that the central engine is a SMBH. Fur-
thermore, dynamical studies cannot provide the spectral
information needed to identify the unde rlying emission
mechanism or mechanisms.
Numerous models have been proposed that can produce
centimeter- through millimeter-band spectra that are at
least roughly consistent with the observations, but this spec-
tral range is too narrow to identify uniquely the nature of
the central engine. What is needed is a detection or strict
upper limit on the flux of Sgr A* at higher frequencies to fix
the overall spectrum on both ends.
Several claims have been made in the literature for the
detection of Sgr A* in the mid- and near-IR (e.g., Stolovy,
Hayward, & Herter 1996; Genzel et al. 1997). However, the
search for an infrared (IR) counterpart to Sgr A* is ham-
pered by source confusion and the strong IR background in
the Galactic center. Precise astrometric alignment of IR
images with radio maps using OH/IR stars that are also
masers indicates that none of the confirmed IR sources seen
so far can be associated definitively with the position of
Sgr A* (Menten et al. 1997). Furthermore, none of the near-
IR sources yet stands out spectroscopically as a possibly
nonstellar object. Consequently, claims of detection of IR
emission from Sgr A* are widely viewed as upper limits at
this time.
The Galactic center is heavily obscured by gas and dust in
the optical and ultraviolet wavebands (A
V
30 mag)
(Becklin et al. 1978; Rieke, Rieke, & Paul 1989). Thirty mag-
nitudes of visual extinction corresponds to a column density
N
H
6 10
22
cm
2
(Predehl & Schmitt 1995), so the
obscuring medium becomes partially transparent to X-rays
from the Galactic center at energies e2 keV. X-ray observa-
tions thus provide our best opportunity to constrain the
high-frequency end of the spectral energy distribution of
Sgr A*. Since strong, hard X-ray emission is a characteristic
property of AGNs, Sgr A* is expected to be an X-ray source
if it derives its energy from accretion onto a supermassive
black hole. However, no definitive detection of X-ray emis-
sion from Sgr A* had been made prior to the launch of
Chandra in 1999 July.
2.2. X-Ray
The earliest X-ray observations of the regions surround-
ing Sgr A* were carried out with rocket- and balloon-borne
instruments (see review by Skinner 1989), but detailed
observations started with Einstein, the first satellite
equipped with grazing-incidence X-ray optics (Watson et al.
1981). Einstein observed the Galactic center twice, 6 months
apart, with the IPC (0.5–4.0 keV) for a total of 18.3 ks and
once with the HRI (0.5–4.5 keV) for 9.1 ks. The IPC images
had an angular resolution of 1
0
(FWHM) and revealed 12
discrete sources within the central 1
1
of the Galaxy.
The error box for the strong est of these sources, 1E
1742.52859, was centered only 20
00
from the position of
Sgr A*. Assuming an absorbed thermal bremsstrahlung
model with kT ¼ 5keVandN
H
¼ 6 10
22
cm
2
, Watson
et al. estimated the absorption-corrected 0.5–4.5 keV lumi-
nosity of this source to be 9:6 10
34
ergs s
1
.
7
The Einstein
images showed that the discrete sources were embedded in a
bright, 25
0
15
0
elliptically shaped region of apparently dif-
fuse emission lying along the Galactic plane, which
accounted for 85% of all the emission from that region. No
variability was detected in the point sources over the 6
month baseline. The HRI image was essentially blank
because of the high absorbi ng column and the low detection
efficiency of that instrument.
Hard X-ray observations were made wi th 3
0
–5
0
resolution
in the late 1980s and early 1990s using Spacelab-2/XRT
7
Throughout this paper we adopt 8.0 kpc for the distance from Earth to
the center of our Galaxy (Reid 1993). All luminosities have been adjusted
to this distance, except where specified otherwise.
No. 2, 2003 CHANDRA IMAGING OF SGR A* AND THE GALACTIC CENTER 893
(Skinner et al. 1987), Spartan-1 (Kawai et al. 1988), and
Granat/ART-P (Sunyaev, Markevitch, & Pavlinsky 1993;
Pavlinsky, Grebenev, & Sunyaev 1994). The line of sight to
the Galactic center becomes optically thin to X-rays with
energies above a few keV; hence the fluxes measured by
these missions were nearly free from the effects of absorp-
tion. These observations suggested the presence of a long-
term variable source near the position of Sgr A* with an
average 4–20 keV lumin osity of 10
36
ergs s
1
.
Prior to Chandra, the highest angular resolution observa-
tions were made with the PSPC and the HRI instruments on
ROSAT (Predehl & Tru
¨
mper 1994; Predehl & Zinnecker
1996). The PSPC observed the Galactic center for 50 ks in
1992 March and detected 14 sources within the central
30
0
30
0
of the Galaxy. With the relatively high spatial reso-
lution of 10
00
–20
00
, it resolved 1E 1742.52859 into three
sources, of which RX J1745.62900 was coincident within
10
00
with the radio posit ion of Sgr A*. The high ab sorbing
column combined with the soft energy band (0.1–2.5 keV)
and modest spectral resolution of the PSPC limited its abil-
ity to constrain the parameters of the spectral fit. Following
Watson et al., Predehl & Tru
¨
mper adopted a thermal
bremsstrahlung model with kT ¼ 5 keV, but with
N
H
¼ 1:5 10
23
cm
2
to obtain broadband agreement with
the hard X-ray data described above, and derived an unab-
sorbed 0.8–2.5 keV luminosity of 6:6 10
35
ergs s
1
for the
source.
8
The HRI, with 5
00
resolution and lower sensitivity
than the PSPC, did not detect a source at the position of
Sgr A* in a 27 ks exposure.
The first X-ray imaging of the Galactic center with
charge-coupled devices (CCDs) was made in 1993 with
ASCA (Koyama et al. 1996). The angular resolution of
the ASCA mirrors was 1
0
. ASCA detected diffuse ther-
mal emission (kT 10 keV) with helium-like and hydro-
gen-like K emission lines of various elements covering
the central square degree of the Galaxy. A 2
0
3
0
ellipti-
cal region filling the Sgr A East shell showed bright dif-
fuse emission at a level 5 times that of the more extended
emission. After correction for a measured absorption of
N
H
7 10
22
cm
2
, the unabsorbed 2–10 keV luminos-
ity of this gas was found to be 10
36
ergs s
1
. No sub-
traction was performed for the spatially variable local
background, and consequently ASCA could only place
an upper limit of 10
36
ergs s
1
on the X-ray luminosity
of Sgr A*. ASCA detected fluorescent line emission from
cold iron in the molecular cloud Sgr B2 but could not
find a bright X-ray irradiator nearby. This lead Koyama
et al. to propose that the X-ray luminosity of Sgr A*
some 300 yr ago may have been 3 10
39
ergs s
1
.
Koyama et al. (1996) found a hard X-ray source located
1<3 away from Sgr A*. During their second observation
made in 1994, Maeda et al. (1996) discove red an X-ray burst
and eclipses with a period of 8.4 hr from the hard source,
establishing that it was an eclipsing low-mass X-ray binary
(LMXB). Only one cataloged transient source, A 1742289
(Eyles et al. 1975), which appeared in 1975, positionally
coincides within the error region. However, Kennea & Skin-
ner (1996) reanalyzed Ariel V data taken in 1975 and found
no eclipses from A 1742289. Hence, the hard source was
identified as a newly discovered LMXB and given the name
AX J1745.62901. Maeda et al. reported that the absorbed
flux from this source varied from 1 10
11
to 4 10
11
ergs
cm
2
s
1
, which was similar to the variations reported previ-
ously by the lower resolution hard X-ray instruments.
Hence, the hard X-ray fluxes attributed to Sgr A* may have
been contaminated significantly by AX J1745.62901 and
A 1742289 (see also Beckert et al. 1996).
A BeppoSAX/MECS observation, with on-axis angular
resolution of 1<3 and an energy range similar to the
ASCA/SIS, was performed in 1997 (Sidoli et al. 1999b).
BeppoSAX detected the diffuse emission near Sgr A*, mea-
sured the absorption column to be N
H
8 10
22
cm
2
,
and set a tighter upper limit on the 2–10 keV luminosity of
Sgr A* of d10
35
ergs s
1
.
In the hard X-ray/soft gamma-ray band, observations
with the SIGMA telescope on Granat yielded an upper limit
of 6 10
35
ergs s
1
(35–150 keV; Goldwurm et al. 1994). The
EGRET instrument on the Compton Gamma-Ray Observa-
tory (CGRO) detected a strong excess of emission from the
Galactic center in the gamma-ray band (>30 MeV; Mayer-
Hasselwander et al. 1998). The error-circle radius of 0=2
included the position of Sgr A*. The source luminosity above
100 MeV was 2:2 10
37
ergs s
1
.Mayer-Hasselwanderetal.
reported that the angular extent of the excess was only
marginally consistent with that of a point source.
As we will see, most of the previously reported X-ray
fluxes can be attributed to a combination of diffuse emission
and numerous stellar sources lying within the instrumental
resolution elements containing Sgr A*. The extremely low
luminosity of Sgr A* and the detection of the multitude of
other nearby sources reported here, indicate that this is in
all likelihood the first true detection of X-ray emission from
Sgr A*. No evidence exists at present that there have been
any long-term variations of Sgr A* that would have allowed
its detection with previous instruments (see x 5.4.2).
3. OBSERVATIONS AND ANALYSIS
3.1. Data Acquisition and Reduction
We observed the center of the Milky Way with Chandra
for 51.1 ks on 1999 September 21. Detectors I0–I3 in the
2 2 element imaging array (ACIS-I) and detectors S2–S3
in the center of the 1 6 element spectroscop y array (ACIS-
S) were read out. The photosensitive region of each CCD is
comprised of 1024 1024 pixels, with each square pixel
subtending 0>492 on a side; hence, each CCD subtended
8<3 8<3 on the sky. Detector S3 is a backside-illuminated
CCD; the other five are frontside-illuminated. The ACIS
CCDs were clocked in timed-exposure (TE) mode using the
standard integration tim e of 3.2 s per frame. The focal plane
temperature was 110
C. To prevent telemetry saturation,
events with energies e15 keV and events with ACIS flight
8
Using the same data as Predehl & Tru
¨
mper (1994), Predehl &
Zinnecker (1996) reported the 0.8–2.5 keV luminosity of RX J1745.62900
as 1
2ðÞ10
34
ergs s
1
, assuming an absorbed power-law model with
¼ 1:6 and N
H
¼ 2 10
23
cm
2
. However, they did not specify whether
this luminosity was corrected for absorption. We used the spectral model of
Predehl & Zinnecker with the response matrix pspcb_gain2_256.rsp to com-
pute the predicted PSPC count rate. Normalizing the model to the count
rate observed by Predehl & Tru
¨
mper (8 10
4
counts s
1
), we found that
the luminosity reported by Predehl & Zinnecker had not been corrected for
absorption. Several papers in the literature have used the luminosity
reported by Predehl & Zinnecker under the assumption that it was unab-
sorbed (e.g., Melia & Falcke 2001). Consequently, the accretion models in
these papers, which were based in part on fits to the luminosity reported by
Predehl & Zinnecker, underestimated the upper limits on the accretion rate
and the X-ray luminosity of Sgr A* in 1992 by 1–2 orders of magnitude.
894 BAGANOFF ET AL. Vol. 591
