The story in brief
In the following, we report on the detection of a previously unknown extended
low-density nebula of very high temperature and excitation,
TR001507.7-391206, located above the midplane of NGC55. The
nebular emission line spectrum, in which HeII is present, is
consistent with photoionization by about 3 very hot massive O3 or
WR-type stars. There are no indications for shock ionization. The
faint blue optical continuum and XMM-Newton data also support
our assumption that this nebula is not a supernova
remnant. Galactic nebulae harboring massive Wolf-Rayet (WR) stars
may be appropriate examples, in particular the highly ionized nebula
G2.4+1.4 around WR 102 (aka Sk4 or LSS4368). The relatively
large diameter of the inhomogeneously expanding nebula in NGC55 of
about 58pc, compared to 11pc of G2.4+1.4, would also be
consistent with the number of ionizing stars.
Details
This object was discovered during a systematic study on the WR star content of NGC55 with EFOSC1 at the ESO 3.6m telescope in 1990. Narrow band images in Halpha and [OIII]5007Å revealed a faint, round, featureless nebular object. Contrary to the classical giant HII-regions in the body of NGC55, it did not show any central peaking of the nebular emission, nor did any stellar-like central source seem to be present in medium band continuum images.
Fig. 3.1
shows the broad- and narrow band filter images of TR001507.7-391206 obtained with different ESO instruments. Due to better spatial resolution of the VLT Halpha-image, substructure within the nebula is clearly visible. There is no continuum emission in the red spectral region, whereas it is present in the blue. For V- and R-band photometry the location of the object is marked by a circle.
A longslit spectrum was taken with the slit covering the stellar cores of the giant HII-regions as well as the nebula (slit s1). As exposed for 1200s, the observed spectrum was very weak, but showed surprisingly strong [OIII] lines. Even [OIII]4363Å was measurable, and yielded a rough estimate of 18000K for the electron temperature, very high in comparison to values typical in the giant HII-regions in NGC55 (about 12000K,
Tüllmann et al. 2003).
One decade after discovery we now benefit from deeper optical spectroscopy (slit s2), images from the VLT, and X-ray imaging from the XMM-Newton satellite.
In
Fig. 3.2 the VLT Halpha-image of the central part of NGC55, taken from
(Tüllmann et al. 2003), is presented. Slit positions s1 (EFSOC1) and s2 (EFSOC2) are scaled to match the chosen slit width of 2" and the individual slit lengths of both instruments. The magnified image to the right shows the position where both slits cut through the nebula.
Fig. 3.3 displays optical spectra of the WR-nebula
(Tüllmann & Rosa 2004). The upper two spectra (slit s1) have been obtained in 1990 using the 3.6m telescope and EFOSC1 at La Silla, Chile, whereas the lower one (slit s2) was taken 10 years later with EFOSC2 at the same telescope. Nightsky and continuum emission has been removed. Given its emission-line characteristics, a SNR can be most likely excluded. The measured electron temperature is around 18000K and densities are within the low density limit (< 10 cm
-3).
Finally, in
Fig. 3.4 the XMM-Newton view of the same region is shown. X-ray contours (pn-data) at different energies are overplotted on the VLT Halpha-image. In addition, UV broadband contours from the optical monitor (OM) unambiguously reveal hot continuum emission indicating the presence of a hot star younger than type O5.
Gas phase abundances
If projection effects along the line of sight are irrelevant, TR001507.7-391206 is located on top of an extended V-shaped column of gas and dust (seen upside down), that is protruding out of the disk plane of NGC55. This is shown very nicely in
Fig. 3.2 (see also
Fig. 6.1), where the gaseous structure is aligned with slit position s2.
Due to the direct detection of [OIII]4363Å, we can reliably measure the gas temperature and, together with constraints on the density, derive metal abundances of the ionized gas.
Abundances are determined with NAT, a code based on the 5-level-atom program of
De Robertis et al. (1987), which calculates emissivities from the observed emission-line strengths listed in
Table 1 (see
(Tüllmann et al. 2003) for details). A conversion from ionic to element abundances is achieved by applying the ICF scheme of
Mathis & Rosa (1991). Element abundances determined this way are given in
Table 2. For convenience, we consider the metallicity Z to be equal to log(O/H), since oxygen is the most abundant and efficient coolant.
In relation to the solar value, [He/H] appears to be overabundant by about a factor of 2.5 at slit position s1 and by about a factor of 1.5 at position s2. At s2 [Ne/O] also appears to be overabundant by about the same factor. This value is, however, based only on a single Ne
++-transition at 3869Å and therefore less certain.
The ionization of the nebula is high. At slit position s1 (s2) only 21% (34%) of oxygen is expected to be present as O
+ and 79% (66%) as O
++. Therefore, it is remarkable that the oxygen abundance of the nebula is within 0.1dex identical to that found for the extraplanar HII-regions (EHRs) in the disk-halo interface of NGC55
(Tüllmann et al. 2003). Similar to the EHRs, the nebula is also located above the central star forming complex, which would imply that the EHRs and the nebula have been formed from the same gas that has been pushed into the halo by supernova explosions during an early star formation epoch.
The detection of HeII4686Å emission in both spectra puts some limitations on the ionizing source of the gaseous nebula. Another constraint comes from the lack of stellar continuum in spectra of the "red" wavelength region as well as on the R-band image shown in
Fig. 3.1. Only in the "blue" spectrum obtained with EFOSC2 (s2), at wavelengths below 5300Å, a very faint stellar continuum seems to emerge from the center of the object. This appears to be consistent with the V-band image (
Fig. 3.1), where some continuum emission might be present at the location of the nebula.
EFOSC1 spectra, however, reveal no continuum emission at all, certainly due to exposure times that are too low by a factor of two.
Since 54.4eV are required to ionize helium completely, only the hottest O stars (i.e. WR stars) or SNRs provide sufficient photon energies to produce the HeII4686 line. Further support of an extreme ionizing source is given by the high gas temperature of about 18000K at a metal abundance which is otherwise identical to those of other nebulae in NGC55, and which have gas temperatures of only 12000K.
XMM-Newton data obtained at different energy bands and overlaid onto the VLT Halpha-image (
Fig. 3.4) reveal a significant detection of hard and soft X-rays around the northern part of the nebula. The question whether or not most of the X-ray photons emitted in the southern part of the nebula are absorbed by the extended gas and dust cloud (see
Fig. 3.2) remains unanswered. UV broadband-data collected with the optical monitor (OM) of the XMM-Newton satellite unambiguously shows continuum at the position of TR001507.7-391206 which is strong in the whole UV domain covered.
However, the detection of an optical and an UV continuum within the central part of the nebula and the observed emission-line characteristics make SNRs as ionizing sources rather unlikely.
Beside photoionization of degenerate stars, shock ionization can also
be a possible ionization mechanism of the gas. However, the
weakness of [SII], [SIII], and [NII] emission-lines,
diagnostic diagrams
(e.g., Baldwin et al. 1981,Veilleux & Osterbrock 1987), shock models
(Shull & McKee 1979, Raymond 1979, Dopita & Sutherland 1995), and the absence of MgII emission and strong [OI]6300Å, so characteristic of SNRs, clearly indicate that the nebula is likely photoionized rather than shock ionized.
All this favors an evolved massive star, ionizing its own mass loss bubble.
In order to check this hypothesis, observed emission-line strengths are compared to predictions made by NAT. Therefore, the derived densities, temperatures, and element abundances have been used as input for NAT. As model atmosphere for the WR star, a pure-helium extended atmosphere of T=90,000K
(Wessolowski et al. 1988) has been assumed. Results are shown in
Table 1. They are in very good agreement with observations and support the idea of a WR star driven mass-loss bubble.
Nebular kinematics
Measurements of radial velocities determined from slit s2, 10" south of TR001507.7-391206 (dust cloud), reveal values ranging from
. As this range is identical to velocities found for the nebula, it implies that this object is indeed located on top of a huge gas spike. It can be speculated if the energetic outflows of the WR star are partly responsible for the V-shaped morphology of the gas spike.
Nebular velocities as a function of radius r plotted for both slit positions are shown in
Fig. 3.5. A correction for the systemic velocity has been applied, as suming v
sys= 118 km s
-1 (Puche et al. 1991). Typical uncertainties in velocity are in the range of 4-12 km s
-1 (see errorbars). Negative radial coordinates for s1 (s2) represent the western (southern) part of the nebula.
Data for s2 show that the ionized gas at the outskirts of the nebula meets the
systemic velocity and is therefore mostly at rest. Hence these regions might be coincident with the outer nebula border.
At s1 the systemic velocity is never approached which indicates that TR001507.7-391206 is not expanding symmetrically.
Summary
We have collected multi-wavelength data in order to constrain the
nature of a high ionization, high temperature nebula located in the
diffuse ISM near the main body of NGC55. Its emission line
characteristics, gas temperature, and elevated [He/H] (and possibly
[Ne/O]) abundance are consistent with slightly enriched, plowed matter
as can be found either in SNRs or WR nebulae, ionized by a hot
continuum. Optical and UV broadband data reveal the existence of
continuum emission originating from this object which directly
excludes SNRs. A SNR nature is also made improbable, because the
characteristic emission of MgII and [OI]6300Å
is absent. Moreover, observed emission line intensities compare well
with photoionization models using a hot WR atmosphere of
T=90 000K as input.
The most plausible interpretation of the data is a WR nebula
photoionized by one of the hottest and most massive WR stars (WO)
similar to the nebula G2.4+1.4 in the Galaxy which is ionized by WR
102. Such stars are rather inconspicuous in optical imaging, but very
deep longslit spectra with the VLT across the nebula might well reveal
the characteristic broad emission lines of the central WR object.
