Possible Explosive Dispersal Outflow in IRAS 16076-5134 Revealed with ALMA

Figure 1 shows the positions of 14 0.9 mm continuum sources detected with ALMA in I16076. These sources are labeled as MM sources and are sorted by their R.A. We consider a flux density threshold of 5σ ∼ 7 mJy beam⁻¹ as a criterion for visually identifying continuum peaks, which we associate with continuum sources (here, σ is estimated by analyzing the empty background emission). The continuum map shows that the sources are mainly distributed in two small clusters: one small cluster around MM3 and a larger one around MM8. Table 2 presents some parameters for characterizing these continuum sources, including central coordinates, deconvolved sizes, peak intensities, flux densities, and masses. These parameters—except for mass—and their uncertainties were estimated by fitting 2D Gaussians, using the CASA task imfit. Since the sizes of the millimeter sources are of the order of ∼0.1 pc, we assume that they correspond to dense cores. ⁶ MM1, MM3, MM5, MM7, MM8, and MM9 were previously identified by Baug et al. (2021), who also considered these sources as dense cores.

Zoom In Zoom Out Reset image size

We calculate the masses of the dense cores using

$$\begin{eqnarray*}&&{M}_{\mathrm{core}}=\displaystyle \frac{{F}_{\nu }{d}^{2}R}{{B}_{\nu }({T}_{d}){\kappa }_{\nu }},\end{eqnarray*}$$

where F_ν is the observed total flux density, B_ν is the Planck function for a dust temperature T_d , d is the distance to the source, R is the gas-to-dust mass ratio, and k_ν is the dust opacity, which depends on the frequency as ${\kappa }_{\nu }={\kappa }_{o}\cdot {(\nu /{\nu }_{o})}^{\beta }$. We assume that the dust emission is isothermal and optically thin, and neglect any possible effects produced by scattering. Therefore, all the masses presented should be considered as lower limits. We adopt the same assumptions as in Fernández-López et al. (2021), who studied the massive star-forming region G 5.89-0.39, also associated with an explosive dispersal outflow. Hence, assuming that I16076 is a region similar to G5.89, we consider β = 2.0, a dust with thin ice mantles with 10⁶ cm⁻³ density, and a gas-to-dust mass ratio of 100. Interpolating the dust opacity models of Ossenkopf & Henning (1994) to the observed wavelength, we obtain an opacity of 1.80 cm² g⁻¹ at 0.9 mm. Since we do not have a measurement of the dust temperatures in the dense cores of I16076, we fix their mean dust temperature to T_d = 30 K (the measured temperature for the dense clump of I16076, as reported by Urquhart et al. 2018). The estimated masses of the dense cores range from 0.3 to 22 M_☉ (see Table 2). To obtain the uncertainty on the masses, we only considered errors in the total observed flux density. The MM7, MM8, and MM9 cores are among the most massive, with masses greater than 10 M_☉. The masses that we obtained are lower by a factor of 3 than the masses of the dense cores estimated by Baug et al. (2021). The reason for the discrepancies resides in the opacity, which they estimated using an expression of Hildebrand (1983), resulting in their masses being a factor of 3 higher than ours. In addition, they warned that their larger masses had to be taken with caution, as the low angular resolution could hinder multiple systems, due to their being marginally resolved. Such is the case for I16076_O1, in which we detect three possible dense cores (MM7, MM8, and MM9), and I16076_O3, composed of two dense cores (MM1 and MM3).

In Figure 2, we present the CO(3–2) molecular emission in the range of radial velocities from −62.15 to 83.47 km s⁻¹, avoiding the central velocity channels ([−5.5:6.2] km s⁻¹), which show extended CO emission from the cloud (velocities are expressed with respect to the cloud velocity, v_sys = −87.70 km s⁻¹). ⁷ We perform the identification of the outflow filaments by using Miriad's cgcurs ⁸ task (Sault et al. 1995). We manually obtain the positions and line-of-sight velocities of all half-beam (0.5 arcsec) condensations where the CO emission, in all pixels belonging to the condensation, are greater than 3σ (34 mJy beam⁻¹). Each condensation is marked with a red/blue dot, in case its velocity is redshifted/blueshifted. This procedure is repeated for each channel of the CO data cube, except for the central channels, where the cloud emission was predominant, and the channels with absorption features at 23.4 and 44.5 km s⁻¹ (see Figure 3). These absorption features may be due to colder and overlapping foreground molecular clouds along the line of sight, and are beyond the scope of this contribution. In addition, the ALMA spectrum reveals other weak emission lines from other chemical species (CH₃OH, ³⁴SO₂, HC₃N, and CH₃CHO), which are not focus of the present work. After collecting the positions of the CO(3–2) condensations, we continue to build up a map on which dozens of filament-like structures, emerging quasi-radially from a common center near the MM8 peak, can be distinguished. Two more north–south extended monopolar filaments (one blueshifted to the south and one redshifted to the north) can be seen east of the filaments associated with MM8. The blueshifted filament, with a length of ∼11'', seems to be associated with the MM3 core. The association of the northern redshifted filament (∼18'' in length) with a dense core is not clear, although Baug et al. (2020) attribute it to the I16076_O2 core. In this work, we focus on analysis of the 24 radial filaments identified around MM8. We find 12 blueshifted and 12 redshifted filaments (see Figure 4). Most of them coincide with the outflows reported by Baug et al. (2020) and are associated with the dense core I16076_O1 (resolved into MM7, MM8, and MM9 in our work). The filament parameters, such as the terminal velocity, coordinates of the outer tip, length, and PA, are given in Table 3. Some of the filaments show very high radial velocities (v > 50 km s⁻¹), while others are more quiescent (v < 10 km s⁻¹). They show a wide range of PA values, indicating that, if they are somehow related, they do not have a preferred direction.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 5 shows a diagram with the line of sight velocity as function of the projected distance (a position–velocity, or PV, diagram) of the CO molecular condensations traced by the filaments around MM8 (up to a projected distance <15''). The redshifted and blueshifted filaments are distinguished by the red and blue dots, respectively. As mentioned before, no molecular condensations were marked around 0 km s⁻¹, due to extended emission contamination. In addition, extended emission is also observed at 42.4 km s⁻¹ throughout the field of view. This can be explained by the presence of other emitting clouds along the line of sight, whose extended emission is filtered by the interferometer, hindering the spatial distribution of the filaments from I16076. The CO emission from filaments at higher velocities appears to be more compact (filaments with projected lengths <2'') and close to the origin of the filaments, near the position of MM8. The black and green arrows join the central position (the peak of the MM8 core at 0 km s⁻¹) and the tips of each filament (see Table 3). The current angular resolution of the observations does not allow the kinematics of the filaments to be adequately analyzed from the PV diagram.

Zoom In Zoom Out Reset image size

A 3D representation of the ensemble of the 24 CO(3–2) filaments toward I16076 gives a better idea of the isotropic morphology and its ordered kinematic pattern (Figure 6). The X-axis and the Y-axis in this graphic are the projected distance in R.A. and decl. (in arcseconds), respectively, while the Z-axis represents the radial velocity (in km s⁻¹) of the filaments, which is also displayed by a color scale. The central position is at ${(\alpha ,\delta )}_{J2000.0}={16}^{{\rm{h}}}{11}^{{\rm{m}}}26\buildrel{\rm{s}}\over{.} 546,-51^\circ {41}^{{\prime} }57\buildrel{\prime\prime}\over{.} 23$ (the MM8 core position). For this image to represent the actual 3D distribution of the filaments, the filaments should be undergoing an expansive motion at constant velocity, following ballistic trajectories. Assuming this, Figure 6 would provide a general view of the spatial distribution and kinematics of the ensemble of filaments. As can be seen in Figures 2, 4, and 6, one of the main characteristics of the filaments is their projected quasi-isotropic distribution on the plane of the sky. They also appear to have a common origin in space and velocity. Furthermore, the velocities within most of the filaments become bluer or redder with increasing distance from the origin. This indicates that the filaments present a velocity gradient. Although a velocity gradient similar to Hubble's law is observed, this does not imply that the filaments are currently undergoing acceleration. Since no clear accelerating source is currently increasing the velocity of the filaments, these velocity gradients are likely related to the decay of the velocity of the molecular material, produced by the interactions between the filaments and the material of the environment (see Section 3.3).

Zoom In Zoom Out Reset image size

In light of the findings explained previously in Section 3.2, we hereafter assume that the ensemble of 24 radial filaments constitutes a single outflow entity with a common origin. In this section, we estimate the physical parameters of the ensemble, such as its mass, momentum, and kinetic energy. We follow an analogous procedure to that described in Fernández-López et al. (2020) for G5.89-0.39, and we consider local thermodynamic equilibrium and an optically thin CO emission. To derive the column density of the CO(3–2) transition, we use the following equation:

$$\begin{eqnarray*}\begin{array}{rcl}{N}_{\mathrm{tot}({CO})} & = & \displaystyle \frac{3{\rm{h}}}{8{\pi }^{3}{\mu }_{B}^{2}J}\cdot \displaystyle \frac{{Q}_{\mathrm{rot}}\,{e}^{{E}_{u}/{{kT}}_{\mathrm{ex}}}}{{e}^{h\nu /{{kT}}_{\mathrm{ex}}}-1}\cdot \displaystyle \frac{\displaystyle \int {T}_{B}{dv}}{[{J}_{\nu }({T}_{\mathrm{ex}})-{J}_{\nu }({T}_{\mathrm{bg}})]}\\ & = & \displaystyle \frac{1.195\,\times \,{10}^{14}({T}_{\mathrm{ex}}+0.922){e}^{33.192/{T}_{\mathrm{ex}}}}{{e}^{16.596/{T}_{\mathrm{ex}}}-1}\\ & & \cdot \displaystyle \frac{{T}_{B}{\rm{\Delta }}v}{{J}_{\nu }({T}_{\mathrm{ex}})-{J}_{\nu }({T}_{\mathrm{bg}})}.\end{array}\end{eqnarray*}$$

In the expression above, we use the Planck constant h = 6.6261 × 10⁻²⁷ erg s; the dipole moment μ_B = 1.1011 × 10⁻¹⁹ StatC cm; the partition function Q_rot = kT_ex/(hB₀) + 1/3, with a rigid rotor rotation constant B₀ = 57.635968 GHz; the quantum number of the upper level J = 3; the Rayleigh–Jeans equivalent temperature J_ν (T) = (h ν/k)/(e^{h ν/kT} − 1); the brightness temperature T_B in K; and the velocity interval Δv in km s⁻¹. Regarding the excitation temperature, we take two different values, 40K and 70 K (Felipe Navarete, private communication; see also Navarete et al. 2019), which are appropriate for many massive clumps in the Milky Way. We estimate the mass in each velocity channel as

$$\begin{eqnarray*}&&{M}_{\mathrm{out}}=\mu {mh}{\rm{\Omega }}{N}_{\mathrm{tot}}/{X}_{{CO}},\end{eqnarray*}$$

where μ is the mean molecular weight, which is assumed to be equal to 2.76 (Yamaguchi et al. 1999); m is the hydrogen atom mass (∼1.67 × 10⁻²⁴g); Ω is the solid angle; and a CO abundance of X_CO = 10⁻⁴. We derive the total mass of the explosive dispersal outflow as the sum of the masses of all the filaments in each channel (see Table 3). In order not to overestimate the mass measurement, we exclude the channels that are contaminated by the emission of the parental molecular cloud and the other clouds along the line of sight, as well as the channels that are affected by other molecular line emission (CH₃CHO and unidentified) toward the wing of the CO line. Finally, we add all the channel masses to derive a total mass between 138 and 216 M_☉ for 70 K and 140 K, respectively.

Recently, Raga et al. (2020) have presented an analytical "head/tail model" that could explain the kinematics found in the explosive Orion BN/KL flow. This model considers an ejected plasmon, where the tail of the plasmon decelerates due to the interaction with the medium that it passes through. The PV diagrams resulting from this model are similar to those of the Orion BN/KL filaments (Zapata et al. 2009; Bally et al. 2017), which show a velocity gradient (the radial velocity increases with the projected distance). As a first approximation, we consider a simple scenario, in which the plasmons travel with no friction. Therefore, the plasmons follow constant velocity trajectories. We also assume that the plasmons are ejected isotropically from a common origin. Therefore, the plasmons will follow a roughly constant velocity trajectory, and we also assume that they are ejected isotropically from the origin of the explosion. To better estimate the momentum and kinetic energy, it is necessary to take into account the inclination of each filament (${v}_{\mathrm{filament}}={v}_{\mathrm{rad}}/\cos (i)$, where v_rad is the velocity along the line of sight and i is the inclination with respect to the plane of the sky). But the resolution of the observations does not allow us to obtain information about the inclinations of each filament, so we will take the maximum observed radial velocity (83 km s⁻¹) as the expansion velocity (V_outflow) of all the filaments. With this V_outflow, we can estimate lower limits for the dynamic age, momentum, and kinetic energy of the outflow. We perform this analysis in Figure 7, which shows a PV diagram similar to that in Figure 5. On the PV diagram, we have superimposed models (the green curve and the shaded region) of an isotropic expansion at constant velocity, varying its dynamic age. Models of good qualitative agreement are manually found by using t_dyn = 3500 ± 500 yr. Finally, the derived total outflow are 1.2 × 10⁴ M_☉ km s⁻¹ and 1.8 × 10⁴ M_☉ km s⁻¹, and the total outflow kinetic are 9.9 × 10⁴⁸ erg and 1.5 × 10⁴⁹ erg for 70 K and 140 K, respectively. This implies that the outflow emission in I16076 is associated with a very energetic event.

Zoom In Zoom Out Reset image size

The outflow found in the high-mass star-forming region I16076 consists of an ensemble of at least 24 radial high-velocity filament-like ejections (ranging from −62 to +83 km s⁻¹), distributed quasi-isotropically. The filaments point back to approximately the same central position, where the MM8 dense core is located (Figure 2). Spatial overlap of the redshifted and blueshifted filaments is observed. The filaments present a linear velocity gradient (see Figure 6). In addition, the estimated energy of the filament ensemble is at least one to two orders of magnitude larger than the typical energies of outflows that are associated with high-mass protostars. All these features point to a direct analogy with those of explosive dispersal outflows: Orion BN/KL, DR21, and G5.89-0.39 (e.g., Bally 2016; Zapata et al. 2017, 2020). The collected evidence strongly suggests that I16076 is a new explosive dispersal outflow candidate. Therefore, the ejection of the filaments would have occurred from a single brief energetic event. However, new and more sensitive observations at higher angular resolution will be needed to confirm this.

As indicated by the elements explained in Section 4.1, the ensemble of filaments in I16076 could constitute a new explosive dispersal outflow. The discovery of another such dispersal event could have further implications, which we explore in this section. For instance, we have no accurate knowledge of how common these dispersal outflows are in our Galaxy , whether they play a role in the energy budget of the or the physical processes from which they originate. Figure 8 shows the spatial distribution of the currently confirmed dispersal outflows in the Milky Way, including I16076 (the red crosses). S106 IR, ⁹ a recently reported possible explosive dispersal outflow (the inverted blue triangle; Bally et al. 2022), is also included in the sample. Making a crude estimate, Zapata et al. (2020) derived a frequency rate of one explosive dispersal event every 130 yr in the Milky Way, considering Orion BN/KL, DR21, and G5.89-0.39. Assuming that I16076 and S106 IR are explosive dispersal outflows, we can refine this estimated rate. First, we assume that the explosive dispersal outflows are evenly spaced over time during the last 15,330 yr (this is the period of time covering all five outflows, taking into account their different distances to Earth). Second, the dispersal outflows are within a projected circle of 5.6 kpc in diameter (the separation between I16076 and DR21, the pair of known dispersal outflows that are the most separated). Finally, we extrapolate the frequency rate to the disk of the our which we assume to be a flat disk with a 15 kpc radius. Since the star formation rate is fairly constant with galactocentric distance (e.g., Nakanishi & Sofue 2006; Urquhart et al. 2014), and we only have a few explosive outflow cases, the frequency rate obtained will be a crude rate of the first-order approximation. Under these assumptions, the estimate of the frequency of the dispersal outflows in the Milky Way should be regarded as a lower limit. In this manner, we obtain a rate of one dispersal outflow every 110 yr. This frequency rate could be even larger, with the availability of a more complete high-resolution and high-sensitivity survey (not just targeted observations), which might lead to the identification of more explosive events within the considered radius.

Zoom In Zoom Out Reset image size

It is worth noting that this frequency rate is comparable to the rate of supernovae (one every 50 yr; Diehl et al. 2006). This fact may have implications for the energy budget of the our even accounting for the lower amount of energy (10⁴⁹ erg), by ∼100 times, that a dispersal outflow deposits in the compared with a core-collapse supernovae, which releases energies of 10⁵¹ erg (Hamuy 2003). The origins of explosive dispersal outflows are probably associated with the disruption of a massive young nonhierarchical stellar system and/or with a protostellar collision (Zapata et al. 2009; Bally et al. 2011; Zapata et al. 2013; Rivilla et al. 2014; Bally 2016; Bally et al. 2017). One of the consequences is the ejection of a small group of runaway protostars (e.g., Gómez et al. 2008; Rodríguez et al. 2020). Hence, there seems to be a link between the origins of explosive dispersal outflows and stellar dynamical interactions. The rate of the explosive dispersal outflows would suggest that dynamical interactions and stellar collisions may be common scenarios in high-density clusters that eject runaway stars. Hence, a more complete survey of these kinds of outflows may be important in advancing our understanding of how massive stars form.