EXPANDED VERY LARGE ARRAY OBSERVATIONS OF THE NEBULA AROUND G79.29+0.46

Massive stars play a fundamental role in the evolution of galaxies. They are major contributors to the interstellar UV radiation and, via their strong stellar winds, provide enrichment of processed material (gas and dust) and mechanical energy to the interstellar medium. Despite their importance, the details of post-main-sequence evolution of massive stars are still poorly understood. Recent evolutionary models suggest that luminous blue variables (LBVs) and related transition objects may play a key role in the massive star evolution, representing a crucial phase during which a star loses most of its H envelope (Lamers et al. 2001). More recently, it has also been pointed out that LBVs might be direct supernova progenitors (Kotak & Vink 2006; Smith et al. 2008), enhancing their importance in the framework of stellar evolution.

LBVs are luminous (intrinsically bright, L ∼ 10⁶ L_☉) objects, exhibiting different kinds of photometric and spectroscopic variabilities. They are massive (M ∼ 20–120 M_☉), characterized by intense mass-loss rates (10⁻⁶ to 10⁻⁴ M_☉ yr⁻¹), which can also occur in the form of eruptive events. Although eruptive events have been witnessed very rarely (e.g., η Car and P Cyg) the presence of extended, dusty circumstellar nebulae around LBVs (LBVNe) suggests that they are a common aspect of LBV behavior (Weis 2008). There are, however, many aspects of LBV phenomenon that are not completely understood. Among these are the total mass lost during the LBV phase (a key parameter necessary to test evolutionary models), the origin and shaping of the LBVNe, and how the mass-loss behavior (single versus multiple events, bursts) is related to the physical parameters of the central object.

The mass-loss archaeology of the central object can be recovered from an analysis of its associated nebula. A successful approach is based on a synergistic use of different techniques, at different wavelengths, that allows one to analyze the several emitting components coexisting in the nebula. In particular, a detailed comparison of mid-IR and radio maps, with comparable spatial resolution, has provided estimates of both ionized gas and dust masses and allowed us to sort out morphological differences in the maps which can be associated with mass-loss behavior during the LBV phase (Buemi et al. 2010; Umana et al. 2010).

1.1. The Nebula Surrounding G79.29+0.46

2.1. EVLA Observations

2.2. Spitzer Data

As we intend to compare the spatial distribution of the ionized gas component, traced by the radio observations, with the morphology of the dust component, traced by the mid-IR observations, in the following we will analyze and discuss only the 5 GHz data set, as that image provides details with a spatial resolution comparable to that of Spitzer observations. The morphology of the radio nebula is evident in our 6 cm EVLA multi-configuration image, which reveals a well-defined shell-like structure whose overall shape was previously reported by Higgs et al. (1994). However, the improved quality of our image, shown in Figure 1, allows us to discern finer details of the ionized gas distribution, most notably the highly structured texture of the nebula. There is evidence for interaction between G79.29+0.46 and the surrounding interstellar medium, i.e., the bright frontal structures in the northeast and the southwest regions of the nebula that delineate regions where the shell structure is locally modified. In particular, the brighter filaments in the southwest region appear to frame the shocked southwestern clump observed in CO (Rizzo et al. 2008; Jimenez-Esteban et al. 2010).

Zoom In Zoom Out Reset image size

The central object is well detected, as is the (probably extragalactic) background object to the northwest. The position, flux density, and angular sizes of these components have been derived by fitting two-dimensional Gaussian brightness distributions to the map. We obtained a flux density of 1.51 ± 0.08 and 1.88 ± 0.09 mJy for the central and background components, respectively. The uncertainty associated with the flux density estimation is given by

$$\begin{equation} \sigma =\sqrt{({\rm RMS}_{{\rm tot}})^{2}+(\sigma _{{\rm cal}}S_{\nu })^{2}}, \end{equation} \tag{ 1 }$$

where RMS_tot is the rms noise in the map and σ_cal is the systematic error due to the flux calibrator (typically on the order of 3%).

Within the errors, the derived flux densities are in agreement with those determined by Higgs et al. (1994), who derived a spectral index of 1.39 ± 0.14 between 5 and 8.4 GHz. They pointed out that, even though this spectral index is steeper than the canonical 0.6, it is still consistent with mass loss from a stellar object.

Assuming that the central radio source is related to the stellar wind from the LBV, we can derive its current-day mass-loss from the standard formula (Panagia & Felli 1975):

$$\begin{equation} \dot{M}= 6.7 \times 10^{-4} v_\infty F_{\nu }^{3/4} D_\mathrm{kpc}^{3/2} (\nu \times g_\mathrm{ff})^{-0.5} \;{M}_{\odot }\;\mathrm{yr}^{-1}, \end{equation} \tag{ 2 }$$

where full ionization and cosmic abundances have been assumed, F_ν is the observed radio flux density, in mJy, and v_∞ is the terminal velocity of the wind in km s⁻¹. The free–free Gaunt factor g_ff is approximated by $g_\mathrm{ff}=9.77(1+0.13 \log {\frac{T^{3/2}}{\nu }})$ (Leitherer & Robert 1991). From the radio flux density observed at 5 GHz, assuming as stellar wind velocity a value of v ∼ 110 km s⁻¹ (Voors et al. 2000), a wind temperature of 10⁴ K, and a distance of 1.7 kpc (Jimenez-Esteban et al. 2010), we derive a mass loss rate of $\dot{M}= 5 \times 10^{-7}\;{M}_{\odot }\;\mathrm{yr}^{-1}$. Vink et al. (2008) pointed out the possibility that G79.29+0.46 is associated with the nearby DR15 region rather than with Cyg OB2. If this is the case, G79.29+0.46 is likely to be closer, at about 1 kpc. Since mass-loss scales with the distance as D^3/2, this would imply a reduced mass loss of a factor 0.45. These values are smaller, but still consistent, with previous evaluations obtained by other authors using different techniques (e.g., Waters et al. 1996), but it is at least an order of magnitude smaller than the current-day mass-loss rates derived from radio measurements for other LBVs (Umana et al. 2010).

The 8 μm IRAC and 24 and 70 μm MIPS images of G79.29+0.46 are shown in Figure 2(a). Previous versions of these images were presented by Kraemer et al. (2010) and Jimenez-Esteban et al. (2010), who both pointed out the presence of a second larger radius shell in the 24 μm image. Our reprocessed maps confirm the existence of the second 24 μm shell and, thanks to our improved reduction, also provide a hint of its presence at 70 μm. Both the 24 and 70 μm maps show the same overall distribution of the dust, consistent with at least two nested dusty shells surrounding the LBV. On the contrary, in the 8 μm map where a major contribution from warm dust is expected, the dust appears to be more concentrated in two southwest and southeast regions. The new EVLA radio observations allow, for the first time, a morphological comparison between the ionized gas and the dust. In Figure 2(b), the same maps are shown but now with the 6 cm EVLA map superimposed using white contours. Once again there is a difference between the IRAC and MIPS maps: while the nebular emission at both 24 μm and 70 μm is more extended than the ionized gas (radio nebula), the 8 μm emission appears well contained within the ionized part of the nebula. Moreover, while the radio emission from the nebula is incomplete in the north region, the emission at 24 μm and 70 μm has a more uniform distribution. The apparent lack of dust in the southern regions in both 8 μm and 24 μm are perhaps due to an absorption effect caused by the infrared dark cloud (IRDC) which may be located in front of the southern region of the nebula.

Zoom In Zoom Out Reset image size

To better visualize the spatial distribution of ionized gas relative to the dust, surface brightness profiles have been extracted along cuts through the nebula from the 6 cm, 24 μm, and 70 μm maps. However, a direct comparison can be performed only between the 6 cm and the 24 μm maps, as they share a comparable angular resolution, ∼5'' for the 6 cm EVLA versus ∼6'' for the 24 μm MIPS map. A superposition of the 6 cm EVLA map on the 24 μm map is shown in Figure 3. We extracted 18 cuts from each map and determined an azimuthally averaged source profile, shown in (Figure 4). The radio shell has a smaller thickness and a sharp decreasing trend in the outer part of the nebula. The dust nebula is more extended and shows a smoother distribution in the outer regions. The second shell is clearly evident in the 24 μm profile, at a distance of about ±200'' from the center. At the same distance, less defined peaks are visible in the 70 μm profile, smoothed out by the coarser spatial resolution.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In this Letter, we present new radio EVLA observations of the nebula surrounding the LBV G79.29+0.46. In particular, we have analyzed the 6 cm map and, for the first time, compared the radio free–free, which traces the spatial distribution of the ionized gas, with maps recently obtained in the mid-IR, which traces the spatial distribution of the dust component.

The dust distribution observed at 8 μm appears different than the distribution at 24 and 70 μm. The warmer dust, traced by the 8 μm emission, is more concentrated in the southwest and southeast regions. Moreover, when compared with the radio nebula, the 24 and 70 μm emission appears more extended, while the 8 μm emission is well contained inside the radio nebula. This is consistent with the existence of two dust components: a cooler component traced by the 24 and 70 μm emission and a warmer dust component, probably consisting of smaller grains, traced by the 8 μm emission. This warmer component, with morphological properties quite different from those observed at longer wavelengths, could be related to the polycyclic aromatic hydrocarbon emission that has been claimed by Jimenez-Esteban et al. (2010).

An improved MIPS 70 μm map, constructed using the most updated pipeline from MIPSGAL team, detects the presence of a second shell very similar to that seen at 24 μm. Nebular lines and dust continuum are the main possible contributors to the 24 μm emission from objects embedded in a dusty nebula with a hot central component. The possibility of such a combination, with one kind of emission being predominant, has been suggested by Mizuno et al. (2010) to explain the 24 μm emission detected in the MIPSGAL Bubbles. In the case of G79.29+0.46, the presence of similar shells at 24 μm and at 70 μm, together with the lack of any prominent emission line falling within the response curve of the MIPS 24 μm band, as evident from the low-resolution Infrared Spectrograph spectrum of the shell (Jimenez-Esteban et al. 2010), leaves very little doubt that the shell emission is entirely due to thermal dust emission.

The fact that there are at least two nested shells provides strong constraints on the origin of the nebula, as it is difficult to explain the presence of multiple shells in the hypothesis of a nebula consisting of swept-up interstellar medium material, as suggested by Higgs et al. (1994). It is evident that the dusty shells consist of material ejected by the central object in different mass-loss episodes.

From the 24 μm emission profile, we have determined that the inner shell peaks at 100'' and the second one at 200'' from the central object. Such dust emission peaks can be related to the epoch when enhanced mass loss took place. Assuming a distance of 1.7 kpc (Jimenez-Esteban et al. 2010), this corresponds to linear distances of 0.82 pc and 1.64 pc, respectively. This implies, assuming a shell expansion velocity of ${\sim} 30\;{\rm km\;{\rm s}}^{-1}$ (Waters et al. 1996), that the two mass-loss episodes occurred 2.7 × 10⁴ and 5.4  ×  10⁴ years ago, or 1.6 × 10⁴ and 3.1 × 10⁴ years ago, if a distance of 1 kpc, as suggested by Vink et al. (2008), is used. Our results point out that mass loss can occur in different episodes, for which we derived characteristic timescales, and pose quite strong constraints that stellar evolution models must take into account.

The direct comparison of the 6 cm and 24 μm maps (that share comparable spatial resolution) indicates that only part of the inner, brighter nebula is ionized (i.e., the nebula is ionization-bounded). The radio nebula shows a sharper profile, depicting regions where probable interaction between the second and the first dusty shells is taking place. This is evident in several regions, most notably in the northeast and in the southwest part of the nebula, where the overall quite regular shell-like morphology is locally disturbed. The mid-IR dust emission linearly depends on density and as a modified blackbody on temperature; the mid-IR emission is therefore most sensitive to dust temperature. On the other hand, assuming a uniform temperature, the thermal radio emission depends on the square of the ionized gas density. This different dependence on density means that the radio emission will emphasize features that have the largest density. This is consistent with the spatial coincidence of the bright radio features with the CO emission reported by Rizzo et al. (2008) and Jimenez-Esteban et al. (2010). In particular, the high critical density of the J = 3–2 CO line (n₃ ∼ 10⁴ cm⁻³) indicates that the CO emission traces higher density regions, confirming that the brighter radio features delineate interactions between the shells where shocks can occur.

This work is based in part on observations made with the Spitzer Space Telescope, which is operated by the Jet Propulsion Laboratory, California Institute of Technology under a contract with NASA. Support for this work was provided by NASA through an award issued by JPL/Caltech.