Ultraluminous X-ray Source 1E 0953.8+6918 (M81 X-9): An Intermediate Mass Black Hole Candidate and its Environs
Extracted source text ยท arXiv astro-ph/0201230
%\documentclass[preprint2]{/home/wqd/text/aastex502/aastex}
%\documentclass[]{/home/wqd/text/aastex502/aastex}
\documentclass[]{aastex}
\usepackage{psfig}
\def\gsim{\lower 2pt \hbox{$\, \buildrel {\scriptstyle >}\over
{\scriptstyle \sim}\,$}}
\def\lsim{\lower 2pt \hbox{$\, \buildrel {\scriptstyle <}\over
{\scriptstyle \sim}\,$}}
\def\rosat{{\sl ROSAT}}
\def\ROSAT{{\sl ROSAT}}
\def\asca{{\sl ASCA}}
\def\ASCA{{\sl ASCA}}
\def\ha{H$\alpha$}
\def\HI{H{\small I}\ }
\def\hii{H{\small II}\ }
\newcommand{\as}{$^{\prime\prime}~$}
\newcommand{\am}{$^{\prime}~$}
\def\xs{{\it RX J0957.9+6903}}
\shortauthors{Wang}
\shorttitle{X-9 and its environs}
\begin{document}
\title{Ultraluminous X-ray Source 1E 0953.8+6918 (M81 X-9):\\
An Intermediate Mass Black Hole Candidate and its Environs
}
\author{Q. Daniel Wang}
\affil{Astronomy Department, University of Massachusetts, Amherst, MA 01003, USA}
\affil{Email: wqd@astro.umass.edu}
\begin{abstract}
We present a {\sl ROSAT} and {\sl ASCA} study of the {\sl Einstein} source X-9
and its relation to a shock-heated shell-like optical nebula in a
tidal arm of the M81 group of interacting galaxies. Our {\sl ASCA} observation
of the source shows a flat and featureless X-ray spectrum well
described by a multi-color disk blackbody model. The source most likely
represents an optically thick accretion disk around an intermediate mass
black hole ($M \sim 10^2 {\rm M}_{\odot}$) in its high/soft state, similar to
other variable ultraluminous X-ray sources observed in nearby disk galaxies.
Using constraints derived from both the innermost stable orbit around a
black hole and the Eddington luminosity, we
find that the black hole is fast-rotating and that its mass is between
$\sim 20/({\rm cos}\ i)\ {\rm M}_{\odot} - 110/({\rm cos}\ i)^{1/2}
{\rm M}_{\odot}$, where $i$ is the inclination angle of the disk. The inferred
bolometric luminosity of the accretion disk is $\sim (8 \times 10^{39}
{\rm~ergs~s^{-1}})/({\rm cos}\ i)^{1/2}$. Furthermore, we
find that the optical nebula is very energetic and may contain
large amounts of hot gas, accounting for a soft X-ray component as indicated
by archival \rosat\ PSPC data. The nebula is apparently associated with
X-9; the latter may be powering the
former and/or they could be formed in the same event (e.g., a hypernova).
Such a connection,
if confirmed, could have strong implications for understanding both the
birth of intermediate mass black holes and the formation of energetic
interstellar structures.
\end{abstract}
\keywords{
X-rays: galaxies --- ISM: binaries and bubbles --- galaxies: individual (M81)}
\section{Introduction}
One of the most enigmatic X-ray-emitting objects is the source X-9
(1E 0953.8+6918), discovered with the {\it Einstein
Observatory} in the field close to the galaxy M81 (Fig.\ 1; Fabbiano 1988).
The source is located about 12\farcm5 from the nucleus of the
galaxy and 2$^\prime$ from the galaxy's dwarf companion Ho IX
($1^\prime$ corresponds to a projected separation of 1 kpc at the distance
$D = 3.6$~Mpc; Freedman et al. 1994).
The 0.2--4~keV flux of the source is only about a factor of $\sim 2$ lower
than the flux of the nucleus (a LINER). X-9 was also detected in
subsequent \rosat\ and \asca\ observations at
comparable flux levels. Clearly the source is not a transient.
X-9 does, however, exhibit significant sporadic timing
variability and therefore must primarily be a compact
source (Immler \& Wang 2001; Ezoe et al. 2001).
Interestingly, Miller (1995) discovered that the source was projected inside
a very unusual \ha-emitting nebula (Fig.\ 1), which also emits
strong [S{\small II}] and [O{\small I}] lines. He concluded that this
nebula was shock-heated and might be a very energetic supernova remnant (SNR)
or a superbubble. Furthermore, the nebula is apparently located
within a nearly ``quiescent'' massive atomic and molecular
gas concentration in a tidal arm (Concentration I, Yun et al. 1994;
Fig.\ 1a). The lack of a significant stellar population in this region has
further led to the suggestion of the
concentration being a protogalaxy (Henkel et al. 1993). The nature of
this combination of stellar and interstellar features remains
unknown.
\begin{figure}
\centerline{\hfil\hfil
\psfig{figure=f1a.ps,height=3.4truein,angle=90.0,clip=}
\hfil\hfil}
\centerline{\hfil\hfil
\psfig{figure=f1b.ps,height=3.2truein,angle=0.0
,bbllx=152bp,bblly=395bp,bburx=474bp,bbury=715bp,clip=}
\hfil\hfil}
\caption{\protect\footnotesize
The field including M81 in near-UV
($\sim 2490 $\.A; Hill et al. 1992), with overlaid HI contours
at 5, 10, 15, 20, and 25 $\times 10^{20} {\rm~cm^{-2}}$ (upper panel).
A close-up of the region outlined by a box in the
field of the protogalaxy is shown in the lower panel. The
H$\alpha$ image is provided by Miller (1995) and
the \rosat\ HRI intensity contours are plotted at 0.24, 0.34, 0.44, and
0.54 $\times {\rm~counts~s^{-1}~arcmin^{-2}}$. The structure of the X-ray
emission is uncertain because of the complicated PSF of the observations at
the large off-axis angle (Immler \& Wang 2001).
}
\end{figure}
In this paper we present results from the analysis of a dedicated
\ASCA\ and archival \ROSAT\ observations of the intriguing X-ray source X-9.
We concentrate on the nature of the source and its relation to the environs, not on the detailed processes involved in the emission and evolution of the source.
\section{X-ray Observations}
Our study is based primarily on our \asca\ observation of X-9,
taken in April of 1999. Located at ${\rm R.A.,
Dec.}$ $({\rm J2000})$ $=9^{\rm h}57^{\rm m}54^{\rm s}$, $69^\circ 03^\prime 50^{\prime\prime}$
(Source \# H44/P66; Immler \& Wang 2001), the source was
observed with the two pairs of
the SIS (exposure 41 ks) and GIS detectors (35 ks).
The SIS, however, experienced a loss
of sensitivity to low-energy ($\lesssim 2$~keV) photons
since approximately late 1994. This
loss increased with time and reached as much as 40\% at the time
of our observation. No method has been proposed to
quantify this loss. Nevertheless, the data from the SIS, with
a better spectral resolution than the GIS, show no
sign of any emission line in the spectrum of X-9. The present
study therefore includes only the data from the GIS, which
had a spectral resolution of $\delta E/E \approx 0.08$$ (6 {\rm~keV}/E)^{0.5}$
over the 0.5--10 keV band and a 50\% half power diameter of $\sim 3^\prime$ on axis.
We followed the standard
data processing procedure as detailed by Day et al. (1995), using the HEAsoft
software package. We extracted on-source counts from an aperture of 4\farcm5
radius and estimated the background contribution in a circle of $15^\prime$
radius around the GIS image center, excluding regions of $6^\prime$ radii
around the M81 nucleus and X-9. The data from the two GIS detectors
were first combined and then loaded into the software XSPEC for
spectral analysis.
We also incorporated archival data from \rosat\ PSPC and
\asca\ GIS observations in our analysis. All these observations were
pointed at either SN1993J or M81 and covered a period of about 9 years;
X-9 was included in the field of view at large off-axis angles
($\gtrsim 10^\prime$) in each of the observations.
The archival data were partly analyzed by Ezoe et al. (2001) and
by Immler \& Wang (2001), concentrating on timing variability. We extracted
the archival GIS data in a similar fashion as described
above for our observation. We combined PSPC spectra extracted
from an on-source aperture of 1\farcm3 radius and a background annulus
of inner and outer radii, 1\farcm8 and $5^\prime$, respectively. All the
spectra are corrected for off-axis PSF and effective area effects.
\section{X-ray Data Analysis and Results}
Because X-9 is variable and lacks an apparent point-like optical
counterpart, we concentrate on spectral models that are appropriate for
a compact X-ray source. We first fit the spectrum
from our own \asca\ observation alone.
Table 1 summarizes results from the fits of several commonly used models.
The power law is characteristic of AGN spectra in the \asca\ band,
although they sometimes exhibit soft {\sl excesses} at energies
$\lesssim 1$ keV and/or hard X-ray {bumps} at $\gtrsim 10$ keV.
Previous analysis by Ezoe et al. (2001) showed that a power law
gave an acceptable fit to an average X-ray spectrum of X-9, extracted from
archival GIS observations. The quality of this spectrum is problematic,
however, chiefly because of the poor PSF at off-axis angles greater than
10$^\prime$. There might also be substantial uncertainties in background
subtraction and energy-dependent PSF-related correction as well as in the
calibration of the instrument energy response. Indeed, the power
law model fits our high quality GIS spectrum of X-9 poorly. The spectrum is
convex-shaped with flux {\sl deficits} at both low and
high energy ends and shows no sign of an Fe K line at $\sim 6.4 - 6.7$ keV,
which often appears in the spectra of bright AGNs. Therefore, X-9 is unlikely
to be an AGN.
Could X-9 be Galactic in origin?
We considered the black body, which is a reasonably good
description of X-ray emission from the surface of an isolated neutron star.
The model, however, is much too sharply peaked to
satisfactorily represent the spectrum of X-9. Raymond \& Smith
optically thin thermal plasma also gives an unacceptable fit
to the spectrum. Our \asca\
observation with a timing resolution up to $\sim 0.1$~ms
was designed partly to test the hypothesis that the source might be
a moving X-ray pulsar, which could also power the optical nebula.
Our FFT and period folding timing analyses of
X-9, however, show no significant periodic signal in the
period range from 1 ms to several hours.
\begin{deluxetable}{lcc}
\tablecaption{Model fits to the \asca\ GIS spectrum}
\tablewidth{0pt}
\tablehead{
\colhead{Model} & \colhead{Parameter\tablenotemark{a}}& \colhead{$\chi^2/n.d.f.$}
}
\startdata
Power-law & energy slope $=1.3(1.2-1.4)$& 240/140\\
& $N_{\rm H} = 5.0(4.5-5.6) \times 10^{21} {\rm~cm^{-2}}$\\
Blackbody & Temperature $=0.74$~keV & 511/140\\
& $N_{\rm H} = 2.3(0.9-3.2) \times 10^{21} {\rm~cm^{-2}}$\\
MCD & $T_{\rm in}=1.34(1.31-1.37)$~keV & 93/130 \\
& $K_{\rm MCD} = 0.20(0.18-0.22)$\\
& $L_{\rm x} = (1 \times 10^{40}/{\rm cos}\ i) {\rm~ergs~s^{-1}}$\\
& $N_{\rm H} = 3.0(0.0-6.1) \times 10^{20} {\rm~cm^{-2}}$\\
\enddata
\tablenotetext{a}{$T_{\rm in}$ is the characteristic temperature of the disk
at its inner radius. The normalization of the MCD model
is defined as $K_{\rm MCD} = (R_{\rm in}/D)^2 {\rm~cos}\ i$, where $R_{\rm in}$ is the
inner disk radius in units of km, the distance to the source $D$ is
in units of 10 kpc, and $i$ is the disk inclination angle.
The disk luminosity $L_{\rm x}$ is calculated in the 0.5--10~keV
band.
Uncertainties in the parameter estimates (enclosed in parentheses)
are at the 90\% confidence limits.}
\end{deluxetable}
We find that an accreting black hole model is an attractive
option for the X-9 spectrum (Fig.\ 2). The source,
if at the distance of M81 and radiating isotropically, has an X-ray
luminosity of $\sim 1 \times 10^{40} {\rm~ergs~s^{-1}}$. This luminosity
is a factor of $\sim 10^2$ greater than the Eddington luminosity
for a $1.4~{\rm M}_{\odot}$ accreting neutron star, but is within the range
of the so-called ultra-luminous X-ray sources (ULXSs), which are detected
typically in disk galaxies and considered as
accreting BHs in the high/soft state (Colbert, \& Mushotzky 1999;
Makishima et al. 2000).
The X-ray spectra of these sources are well described by the multi-color disk
blackbody (MCD) model (XSPEC model {\sl diskbb}). Indeed, the model fits
the spectrum of X-9 well (Table 1).
\begin{figure}
\centerline{ {\hfil\hfil
\psfig{figure=f2.ps,height=2.5in,angle=270, clip=}
\hfil\hfil}}
\caption{Joint fit to the combined \asca\ GIS and \rosat\ PSPC spectra
of X-9. The histograms represent the best fit of the MCD model.
\label{fig2}}
\end{figure}
Next, we jointly fit the GIS and PSPC spectra with the
MCD model (Fig.\ 2). The PSPC data are particularly
sensitive to soft X-ray emission and
absorption. Because the flux of X-9 varies with time
(amplitude of variability exceeding a factor of $\sim2.5$; cf. Fig.\ 3 in
Immler \& Wang 2001) and because the spectrum could also
change accordingly, we allow both $T_{\rm in}$ and $K_{\rm MCD}$ to be
different for the GIS and PSPC spectra and joint-fit the X-ray absorbing
gas column density.
The MCD model gives a good fit to the spectra ($\chi^2/n.d.f. = 143/166$).
The fitted column density is $1.0 (0.9-1.1) \times 10^{21} {\rm~cm^{-2}}$
(90\% confidence limits), compared to the expected Galactic contribution
of $\sim 5 \times 10^{20} {\rm~cm^{-2}}$ in the direction. This
extra absorption is consistent with the hypothesis that X-9 is located
inside Concentration I. The MCD spectral parameters for the GIS data
are nearly identical to those from the above single spectrum fit.
But the spectral parameters for the PSPC data
are rather different: $T_{\rm in} = 0.59(0.55-0.63)$ and $K_{\rm MCD} =
1.7(1.3-2.2)$.
\section{Discussion}
The above spectral characteristics of X-9, together with its
strong aperiodic variation (Immler \& Wang 2001) and its lack of
periodicity, argues against a scenario of the source as
a Galactic compact object.
Indeed, at the source's Galactic position of $l, b = 143^\circ, 18^\circ$,
there is little room for such a scenario.
Ezoe et al. (2001) have made additional arguments against the Galactic origin
of X-9, based primarily on its high X-ray to optical emission ratio.
In contrast, the model of X-9 as an accreting BH inside Concentration I
at the M81 distance naturally accounts for the sporadic timing behavior,
the convex-shaped spectrum, the excess absorption, and the large X-ray flux
of the source.
But the apparent discrepancy between
the GIS and PSPC spectral characteristics demands an explanation.
The low $T_{\rm in}$ as inferred from the PSPC spectrum is very abnormal,
compared with similar ULXSs observed with \asca. X-9 also shows
strong flux variation: e.g., $2.5 \times 10^{-12}
{\rm~ergs~s^{-1}~cm^{-2}}$ for the PSPC spectrum vs.
$5.1 \times 10^{-12} {\rm~ergs~s^{-1}~cm^{-2}}$ for the GIS spectrum
over the overlapping 0.5--2 keV band. But this variation is
typical for ULXSs and is well within the range inferred from the
archival GIS spectra of X-9. Indeed, the MCD model
provides satisfactory fits to all individual GIS spectra, and
$T_{\rm in}$ is always greater than 1.2 keV. Clearly, the MCD model with the
abnormal low $T_{\rm in}$ cannot be a fair
characterization of the GIS spectra. It is important to note that
the MCD model has been tested primarily on X-ray spectra of heavily
absorbed Galactic black hole candidates. It is possible
that the model does not adequately describe the soft X-ray emission
from the accretion disk of X-9.
Alternatively, the PSPC spectrum of X-9 may contain
a significant soft X-ray contribution from diffuse hot gas. This may be
expected in the region enclosed by the shock-heated optical nebula, or
from relativistic particles accelerated in an expanding shock
related to the nebula. The narrow bandwidth and very limited spectral
resolution of the PSPC data, however, do not allow for a useful spectral
decomposition of such a soft component from the uncertain
disk emission of the source at the time of the observations.
A spatial analysis of the archival \rosat\ PSPC and HRI data also shows
evidence for the extended X-ray emission around the centroid of X-9 on
scales comparable to the size of the nebula. But
various systematic uncertainties (e.g., the source centroid shift
caused by the off-axis instrumental point spread function)
prevent us from a definitive measurement of the extended emission.
\subsection{X-9 as an Accretting IMBH}
Assuming that the MCD model is correct, we can estimate the mass
of the BH: $M_{\rm BH} = (41 {\rm M}_{\odot}) \alpha^{-1} (K_{\rm MCD}
/{\rm cos}~i)^{1/2},$ where
$\alpha$ is the ratio of the inner disk radius to the last stable
orbit radius of a non-rotating (Schwarzschild) BH, and
both $K_{\rm MCD}$ and $i$ are defined in the note to Table 1. Adopting $K_{\rm MCD}$
in Table 1 gives
$M_{\rm BH} = 18/(\alpha^2 {\rm cos}~i)^{1/2} {\rm M}_{\odot}$.
For a general rotating Kerr BH, $\alpha \gtrsim 1/6$
(e.g., Zhang et al. 1997) and thus $M_{\rm BH} \lesssim 108
{\rm M}_{\odot}/({\rm cos}~i)^{1/2}$. The model, however, does not include various
general relativistic effects (e.g., light bending). Ongoing modeling
of such effects shows that the cosine law is
approximately preserved for Schwarzschil BHs and that for
extreme Kerr BHs the effective ${\rm cos}\ i = 0.17-0.4$ for $i =
5^\circ-85^\circ$ (Zhang et al. 2001). So the range of variation can be
considerably smaller than the cosine law.
Following Makishima et al. (2000), we can also estimate the bolometric
luminosity of the disk as
$L_{bol} = (8 \times 10^{39} {\rm~ergs~s^{-1}})/({\rm cos}~i)^{1/2}.$
Because $L_{bol}$ should typically be smaller than, or at most comparable
to, the Eddington luminosity $1.3 \times 10^{38}
(M_{\rm BH}/{\rm M}_{\odot})$,
we obtain $\alpha \lesssim 0.76 ({\rm cos}~i)^{1/2}$,
indicating that the BH is fast-rotating ($\alpha < 1$) and that
$M_{\rm BH} \gtrsim 24 {\rm M}_{\odot}/{\rm cos}\ i$. The 90\% statistical
uncertainties in these mass limits are about 10\%.
\subsection{Optical Nebula}
The presence of the unusual shell-like optical nebula
is an important part of the mystery about X-9 (Fig.\ 1b).
Optical spectroscopy of the nebula shows that its mean heliocentric
velocity ($\sim 47-52 {\rm~km~s^{-1}}$) agrees with the velocity
($\sim 45-65 {\rm~km~s^{-1}}$) of the \HI concentration
(Adler \& Westpfahl 1996; Miller 1995). The \HI velocity field further
shows that the concentration is part of the M81 group (Miller 1995).
In contrast, the dwarf galaxy Ho IX has an optical heliocentric
velocity of $119 \pm 60$ (de Vaucouleurs et al. 1991) and is
offset spatially from the centroid of the \HI\ concentration.
Therefore, Ho IX may not be related to either the \HI\ concentration or
the nebula (Fig.\ 1a).
What might be the origin of the nebula around X-9?
At the distance to M81, the nebula has an \ha\ luminosity
$L_{{\rm H}\alpha} \sim 1 \times 10^{38} {\rm~ergs~s^{-1}}$
(Miller \& Hodge 1994), which is about three times more luminous than the
bright SNR N49 in the LMC (Vancura et al. 1992).
But the most distinct difference between the two
is their sizes: $\sim 250$ pc $\times$ 475 pc for the
X-9 nebula (Miller 1995) vs. 8 pc radius for N49.
If the nebula is heated primarily by a shock (Miller 1995),
a significant fraction (parameterized here as
$\xi$) of $L_{{\rm H}\alpha}$ may then be produced by the
excitation as pre-shock hydrogen atoms drift into the post-shock region
(e.g., Cox \& Raymond 1985). Optical
spectroscopy so far, however, places only an upper limit of $300
{\rm~km~s^{-1}}$ on the
shock velocity of the X-9 nebula (Miller 1995).
Assuming a shock velocity $v_2 \gtrsim 2$
(in units of $10^2 {\rm~km~s^{-1}}$ and
the standard case B (that is, optically thick for all Lyman recombination
lines), we can estimate from
$L_{{\rm H}\alpha}$ the pre-shock neutral gas density as
$n_o \sim (10 {\rm~cm^{-3}}) v^{-1}_2 R_2^{-2}\xi$,
where $R_2$ (in units of $10^2
{\rm~pc}$) is the characteristic radius of the nebula.
The kinetic energy is then $E_k \sim (1 \times 10^{53} {\rm~ergs})
v_2 R_2\xi$ if $n_o$ is spatially uniform. This is likely to be an
overestimate because the bulk of the H$\alpha$-emitting gas is expected
to arise in dense filaments and clouds in which shock velocity tends to be low.
Furthermore, a substantial fraction of the H$\alpha$ emission may also
be due to the ionization by the X-ray source (\S 4.3). Nevertheless,
comparisons with known SNRs suggest that the
X-9 nebula appears to energetic to be due to a single normal supernova (SN).
Could the optical nebula then be a superbubble produced by multiple
SNe and fast stellar winds of massive stars
(e.g., Mac Low \& McCray 1988)?
At the position of X-9, there is indeed a fuzzy blue object
which may represent a stellar cluster ($m_{\rm B} = 20$; Henkel et al. 1996;
Miller 1995). If this is the case, the stellar cluster should then have an age
$t_s \gtrsim 10^7$ yr, as young stars do not appear to
contribute much to the ionization of the optical nebula and
no far-UV emission peak appears at the position of X-9 (e.g., Fig.\ 1).
Comparing $t_s$ with the expansion age of the nebula
$t_e \sim 3R/5v = (1 \times 10^6 {\rm~yr})
R_2 v_2^{-1}$, we find $v \lesssim (5 {\rm~km~s^{-1}})
R_2$. This is less than the turbulence velocity of the ISM
and too small for shock-heating to effectively produce significant
\ha\ emission. Therefore, unless being accelerated recently by an SN (or
a hypernova; see later discussion) the optical nebula is probably not
a superbubble created by a massive stellar cluster.
\subsection{Association of the Optical Nebula and X-9}
Using the log(N)-log(S) function from the \rosat\ All Sky Survey (Hasinger et
al. 1998), we estimate that the probability for a chance projection of an X-ray
source comparable to, or brighter than, X-9 within a circle of
$\sim 10^{\prime\prime}$ radius is only about $\sim 5 \times 10^{-7}$.
Therefore, X-9 is most likely associated with the nebula.
We speculate that the optical nebula may be directly related to
the presence of X-9. One possibility is that they were born together.
The nebula may be an interstellar remnant of a hypernova explosion
that is substantially more energetic than a normal SN.
Hypernovae have been
postulated as the sources of some $\gamma$-ray bursts observed at
cosmological distances (Paczy\'nski 1998; Fryer \& Woosley 1998).
The proposed mechanism for hypernova explosions
is the collapse of certain massive stars and/or their mergers with
compact companions. Such an
event leads to the formation of a BH and provides an extractable
energy of $\sim 10^{54}$~ergs (M\'esz\'aros, Rees, \& Wijers 1999).
Hypernovae may also be responsible for some \HI supershells or holes
observed in the interstellar medium (ISM) of nearby galaxies (Efremov,
Elmegreen, \& Hodge 1998; Loeb \& Perna 1998), as well as
relatively young energetic shell-like nebulae (Wang 1999).
The optical nebula around X-9 could well be such a hypernova remnant.
The X-ray source may represent the resultant BH accreting from the
materials falling back from the explosion. Although a binary may
survive the explosion, the turning-on of a persistent accretion
phase is expected to be long after
the explosion remnant has disappeared. The soft X-ray contribution from the
hypernova remnant may be significant in the PSPC spectrum, whereas the GIS
spectrum is dominated by the bright accretion disk in a higher state,
which makes the detection of the soft component difficult.
Another plausible scenario is that the nebula is currently
powered by an intense outflow or a wind from X-9 as an accreting X-ray binary. If the nebula is purely caused by this outflow,
the required mean energy output over the nebula's expansion age
$t_e$ is then a few times $ 10^{39} {\rm~ergs~s^{-1}}$, which is comparable to
the X-ray luminosity of the source. In principle, the power of
such an outflow could even be greater than the radiation
luminosity of X-9, if it is similar to other types of accreting
BH systems (i.e., Galactic micro-quasars or AGNs). In this case, the IMBH may be a
Population III remnant (e.g., Madau \& Rees 2001) and its companion may be
formed from the collapse of surrounding molecular gas inside
Concentration I
(Henkel et al. 1993; Yun et al. 1994; Brouillet et al. 1992; Fig.\ 1).
X-9 may also contribute to the ionization of the nebula. The MCD model of X-9
predicts an integrated ionizing photon rate of $\sim 4 \times 10^{40}
{\rm~s^{-1}}$ over the 0.016--0.2 keV range; photons at higher energies should
mostly escape from the nebula. This rate is about a factor of 10
greater than that estimated for LMC X-1 (a stellar mass BH candidate),
which is surrounded by an X-ray-ionizing nebula (Pakull \& Angebault 1986).
The nebula around X-9 shows indications of
X-ray ionization. First, the H$\beta$ to H$\alpha$ flux ratio ($\sim 0.6$)
of the nebula is high, indicating a high electron temperature of $\sim 10^5$~K
(Miller 1995). Second, the shell-like morphology of the nebula appears
to be rather diffuse, as is expected from the relative long absorption
path-length of soft X-rays and its strong energy dependence
(e.g., Rappaport et al. 1994).
Further scrutiny of the above scenarios and their relative importance
is both desirable and possible. The nature of the blue object is yet to be
determined: is it the optical counterpart of the accreting system or the
outflow?
The outflow may also be probed by observing its nonthermal radio emission,
which should show a persistent flat or inverted radio spectrum
(e.g., Fender 2001). The 20 cm radio continuum map of Bash \& Kaufman (1986),
which is centered on M81, does show a 3$\sigma$ contour
at the position of X-9.
The total flux of $\sim 1$ mJy, if indeed associated with the outflow,
is substantially greater than that observed from Galactic micro-quasar-like objects.
Detailed optical spectroscopy of the nebula will be especially important
for determining the expanding velocity of the nebula and will provide
important constraints on the extreme UV to
soft X-ray radiation properties of X-9. Such radio/optical
observations, combined with future high resolution X-ray imaging and spectroscopy,
should allow for a firm determination of the relation
between the nebula and X-9.
In summary, the association of X-9 with
both the blue object and the shell-like $H\alpha$ nebula as well as
Concentration I provides an excellent opportunity for studying the nature
of ultraluminous X-ray sources and for characterizing their radiation and outflow as well as their effects on the interstellar medium.
\acknowledgements
The author thanks B. W. Miller and M. Yun for the H$\alpha$ and
\HI\ images used here, A. D'Onofrio for her help in analyzing the
archival \asca\ GIS data mentioned in the text, and the referee for
useful comments. The author also appreciates
comments from S. Immler, R. Williams, K. Wu, and S. N. Zhang on the work,
which was funded by NASA under the grants NAG5--9429 and NAG5--8999.
\begin{references}
\reference{} Adler, D. S., \& Westpfahl, D. J. 1996, AJ, 111, 735
\reference{} Bash, F. N., \& Kaufman, M. 1986, ApJ, 310, 621
\reference{} Brouillet, N., et al. 1992, A\&A, 262, L5
\reference{} Colbert, E. J. M., \& Mushotzky, R. F. 1999, ApJ, 519, 89
\reference{} Cox, D. P., \& Raymond, J. C. 1985, ApJ, 298, 651
\reference{} Day, C. S. R., et al. 1995, The ABC Guide to \asca Data Reduction, \asca\ Guest Observer Facility, NASA/GSFC
\reference{} de Vaucouleurs, G., et al. 1991, Third reference catalogue of bright galaxies, New York: Springer-Verlag
\reference{} Efremov, Y. N., Elmegreen, B. G., \& Hodge, P. W. 1998, ApJL, 501, 163
\reference{} Ezoe, Y., Iyomoto, N., \& Makishima, K. 2001, PASJ, 53, 69
\reference{} Fabbiano, G., 1988, ApJ, 325, 544
\reference{} Fender, R. P. 2001, MNRAS, 322, 31
\reference{} Freedman, W. L., et al. 1994, ApJ, 427, 628
\reference{} Fryer, C. L., \& Woosley, S. E. 1998, ApJL, 502, 9
%\reference{} Fryer, C. L., Woosley, S. E., \& Heger, A. 2000, ApJ, 550, 372
%\reference{} Haiman, Z., \& Loeb, A. 2000, ApJL, 552, 459
\reference{} Hasinger, G., et al. 1998, A\&A, 329, 482
\reference{} Henkel, C., et al. 1993, A\&A, 273, L15
\reference{} Hill, J. K., et al. 1992, ApJL, 395, 37
\reference{} Immler, S., \& Wang, Q. D. 2001, ApJ, 554, 202
\reference{} Loeb, A., \& Perna, R. 1998, ApJL, 503, 35
\reference{} Mac Low, M.-M., \& McCray, R. 1988, ApJ, 324, 776
\reference{} Madau, P., \& Rees, M. J. 2001, ApJL, 551, 27
\reference{} Makishima, K., et al. 2000, ApJ, 535, 632
\reference{} M\'esz\'aros, P., Rees, M. J., \& Wijers, R. A. M. J. 1999, New Astronomy, 4, 303
\reference{} Miller, B. W. 1995, ApJL, 446, 75
\reference{} Miller, B. W., \& Hodge, P. 1994, ApJ, 427, 656
\reference{} Paczy\'nski, B. 1998, ApJL, 494, 45
\reference{} Pakull, M. W., \& Angebault, L. P. 1986, Nature, 322, 511
\reference{} Rappaport, S., Chiang, E., Kallman, T., \& Malina, R. 1994, ApJ, 431, 237
%\reference{} Vancura, O., et al. 1992, ApJ, 394, 158
\reference{} Wang, Q. D. 1999, ApJL, 510, 139
\reference{} Yun, M. S., Ho, P. T. P., \& Lo, K. Y. 1994, Nature, 372, 530
\reference{} Zhang, S. N., Cui, W., \& Chen, W. 1997, ApJL, 482, 155
\reference{} Zhang, X., Zhang, S.N., \& Yao, Y. 2001, Gamma-ray conference, Baltimore, Maryland
\end{references}
\end{document}
\end