Detection of an Explosive Outflow in G34.26+0.15

Abstract

This paper reports on the detection of a likely explosive outflow in the high-mass star-forming complex G34.26+0.15, adding to the small number (six) of explosive outflows detected so far. ALMA CO($2-1$) and SiO($5-4$) archival observations reveal multiple outflow streamers from G34.26+0.15, which correlate well with H₂ jets identified from Spitzer-IRAC 4.5 $\mu$m and [4.5]/[3.6] flux ratio maps. These nearly linear outflow streamers originate from a common center within an ultracompact Hii region located in the complex. The velocity spread of the outflow streamers ranges from 0 to 120 km s^-1. The radial velocities of these streamers follow the Hubble-Lemaître velocity law, indicating an explosive nature. From the CO emission, the total outflow mass, momentum, and outflow energy are estimated to be $\sim$264 M_⊙, 4.3$\times 10^{3}$ M_⊙ km s^-1, and 10⁴⁸ erg, respectively. The event triggering the outflow may have occurred about 19,000 years ago and could also be responsible for powering the expanding UC Hii region, given the similar dynamical ages and positional coincidence of the UC Hii region with the origin of the outflow. The magnetic field lines in the region associated with G34.26+0.15 also appear to align with the direction of the outflow streamers and jets, possibly being dragged by the explosive outflow.

1 Introduction

Highly energetic outflows of explosive nature are a new subclass of molecular outflows detected in massive star-forming regions. Unlike the typical, relatively long-lived collimated bipolar outflows which dissipate the excess angular momentum during the phase of mass accretion onto the forming young stellar objects (YSOs), explosive outflows are impulsive and short-lived and believed to be powered by single, brief energetic events with energy injections of $\sim 10^{47-49}$ erg (Bally & Zinnecker, 2005). A sudden ejection of gravitational potential energy, possibly triggered by the dynamical rearrangement of a non-hierarchical massive young stellar system like a stellar merger or by a protostellar collision can drive explosive outflows (Zapata et al., 2009; Rivilla et al., 2014; Bally, 2016; Bally et al., 2017). Zapata et al. (2017, 2019) have drawn a clear morphological and kinematic distinction between classical protostellar and explosive outflows. Explosive outflows are characterized by the presence of several filament-like molecular gas streamers, isotropically distributed in the sky, each of which follows the Hubble-Lemaître velocity law where the radial velocities of the filaments increase linearly with the projected distance from the central source. The isotropic distribution makes the red and blueshifted streamers to appear to overlap on the plane of the sky. The gas streamers are often traced by CO and SiO molecular lines with the emission reaching radial velocities of up to 100 km s^-1. Aside from the gas streamers, H₂ “wakes” and [FeII] “fingers” can also be observed towards the tips of the filaments as observed in the case of Orion Becklin–Neugebauer (BN)/Kleinman–Low (KL) (Bally et al., 2015; Youngblood et al., 2016). Based on high angular resolution polarization observations, (Orion BN/KL; Cortes et al. 2021 and G5.89$–$0.39; Fernández-López et al. 2021), it is also believed that explosive outflows are strong enough to drag the magnetic field lines and rearrange them in a quasi-radial orientation with respect to the origin of the outflow.

In recent years, a few explosive outflow sources were identified from the molecular gas kinematics, namely Orion BN/KL (Zapata et al., 2009), DR21 (Zapata et al., 2013), G5.89-0.39 (Zapata et al., 2019, 2020), IRAS 16076-5134 (Guzmán Ccolque et al., 2022), Sh2-106 (Bally et al., 2022), and IRAS 12326-6245 (Zapata et al., 2023). From these discoveries, Zapata et al. (2023) have estimated the rate of explosive outflow events to be one every 90 yr in the Milky Way. Interestingly, as noted by these authors, this rate is comparable to the approximate rate of supernovae of one in 50 yr (Diehl et al., 2006) and also the massive star formation rate of one in 50 yr. These indicate that the dynamic interactions in massive young stellar systems like stellar mergers or protostellar collisions might be common occurrences in the initial stages of massive star formation in high-density clustered environments, also leading to the ejection of runaway stars as seen in the case of Orion BN/KL (Rodríguez et al., 2005; Gómez et al., 2008; Bally et al., 2011).

In this paper we discuss the likelihood of G34.26+0.15 (hereafter G34) being an explosive outflow source. Located at a distance of 3.3 kpc (Kuchar & Bania, 1994), G34 is a high-mass star-forming complex. It consists of four radio components, two of which are hypercompact (HC) Hii regions, and the others are a cometary ultracompact (UC) and a shell-like Hii regions designated as A, B, C, and D, respectively throughout literature (e.g., Reid & Ho, 1985; Garay et al., 1986; Sewilo et al., 2004). Liu et al. (2013) suggest that the expansion of the Hii region, D is responsible for inducing the sequential star formation in G34, which is also supported in a recent study by Khan et al. (2024). The cometary UC Hii region, C, consists of a “head” and a diffuse “tail” that points (from tail through head; see Figure 1a) in the direction of the supernova remnant (SNR) W44 that lies a projected distance of $\sim$40 pc. However, the SNR shell extends to only $\sim$25 pc in radius and hence is not yet at a distance close enough to influence the gas motion in G34 (Reid & Ho, 1985). This UC Hii region also harbours a chemically rich hot molecular core (HMC) (MacDonald et al., 1995). Mookerjea et al. (2007) propose that the UC Hii region is primarily responsible for energizing the HMC which shows no evidence of internal heating. These authors also found the presence of several nitrogen and oxygen bearing complex organic molecules (COMs). Hajigholi et al. (2016) detected ammonia (NH₃) lines towards the G34 HMC that shows inverse P-Cygni profiles suggesting mass infall onto the central source. G34 has been catalogued as an “outflow-only” candidate by Cyganowski et al. (2008) from the large-scale Spitzer Galactic Legacy Infrared Mid-Plane Survey Extraordinaire (GLIMPSE; Benjamin et al., 2003), where multiple jet/outflow like structures are seen in the IRAC 4.5 $\mu$m band extending in several directions away from the central object. Such emission may arise from the H₂ ($v=0-0$, S(9,10,11)) lines and/or CO ($v=1-0$) bandheads that are excited by shocks from outflows (e.g., Noriega-Crespo et al., 2004; Smith et al., 2006a). According to Cyganowski et al. (2008), the $[3.6]-[4.5]$ color of the extended emission fall within the range of “shocked outflow nebulosity” determined for the DR21 outflow (see Figure 7 of Smith et al., 2006b). In the near-infrared regime, Lee et al. (2013) have identified several isolated H₂ knots from the H₂$-K$ continuum subtracted image of G34 and classified them as candidate H₂ outflows. Furthermore, several SiO outflows have also been detected towards the north-west, south-east, and north-east of G34 (Hatchell et al., 2001).

In presenting the evidence to support the likelihood of G34 being an explosive outflow, we make use of the observations from the archives of ALMA, VLA, and JCMT. The paper is organized as follows. Section 2 outlines the observations and data reduction details. In Section 3 we discuss the identification of outflows and whether G34 is associated with an explosive outflow event. The conclusions from this study are presented in Section 4.

2 Archival Observations

2.1 ALMA data

We make use of the Band 6 Atacama Large Millimeter/submillimeter Array (ALMA) archival data to investigate the gas kinematics of the region associated with G34. The 12-m array observations were carried out on 2019 November 28 using 43 antennas, with baselines ranging from 15 to 313 m (Project ID: 2019.1.00263.S; PI: John Bally). The observations were made in mosaic mode, consisting of 53 pointings distributed in a Nyqusit-sampled grid with a total on-source time of approximately 23 minutes. The average precipitable water vapor during the observations was 1.1 mm. The phase center of the observations was located at the sky position $\alpha_{J2000}$=18^h53^m15^s.871 and $\delta_{J2000}$=01^∘15^′07^′′.131. The largest recoverable scale of this observation is 11^′′.7. Of the four spectral windows (SPWs), we focus on SPW1 and SPW3 centred at 231.065 and 217.648 GHz, covering the transitions ¹²CO($2-1$) at 230.538 GHz and SiO($5-4$) at 217.105 GHz, respectively. We make use of the line-free channels from all the four SPWs to obtain the continuum image. J1924-2914 was used as the flux and bandpass calibrator, while J1851+0035 was used as the phase calibrator.

The data were calibrated and imaged using the Common Astronomy Software Applications (CASA) Version 5.6.1-8. The imaging was done employing the task TCLEAN with the Robust parameter set to +0.5. The continuum map generated has an rms noise level of 0.5 mJy beam^-1 and an angular resolution of 1^′′.5$\times$1^′′.1, equivalent to spatial resolution of $\sim$0.02 pc ($\sim$4000 AU). The CO($2-1$) and SiO($5-4$) cubes have similar angular resolutions of 1^′′.5$\times$1^′′.2 and 1^′′.6$\times$1^′′.3, respectively, and a velocity resolution of 1.5 km s^-1.

2.2 VLA data

To probe the ionized emission associated with G34, we use the 8.46 GHz (3.6 cm) data retrieved from the National Radio Astronomy Observatory VLA (Very Large Array) Archive Survey (NVAS). The observation was carried out on 1991 December 6 using the VLA B/A configuration (Legacy ID: AW303; D. Wood). The 3.6 cm map has an angular resolution of 0^′′.8$\times$0^′′.7 and rms noise level of 0.2 mJy beam^-1.

2.3 JCMT data

The orientation of magnetic field in the region associated with G34 is derived from the archival 850 $\mu$m polarization data (Project ID: M16AD003; PI: Sarah Graves). The observations were done with SCUBA-2/POL-2 (Holland et al., 2013; Friberg et al., 2016, 2018) mounted on the the James Clerk Maxwell Telescope (JCMT) in the POL-2 DAISY scanning mode. The effective beam is $14.1^{\prime\prime}$ at 850 $\mu$m (Dempsey et al., 2013). The data are reduced following the standard procedure for SCUBA-2/POL-2 observations using the STARLINK package SMURF (Chapin et al., 2013; Currie et al., 2014). The final I, Q, and U maps are in units of pW. They are converted to the units of Jy beam^-1 by applying a flux correction factor (FCF) of 725 Jy beam^-1 pW^-1. Due to the flux loss from POL-2, value of FCF is 1.35 times larger than for the standard SCUBA-2 FCF of 537 Jy beam^-1 pW^-1 (Dempsey et al., 2013). The rms noise level of the total intensity (Stokes I) is measured to be 40 mJy beam^-1. The polarization angles are derived following the procedure described in Gu et al. (2024). Only the polarization vectors where the non-polarized intensity ($I$), polarized intensity ($P$), and whose polarization percentage ($p$) satisfying the criteria $I/\delta I\geq 10$, $P/\delta P\geq 3$, and $\delta p\leq 5\%$, respectively were selected. The magnetic field orientation is obtained by rotating the polarization vectors by 90^∘.

3 Results and Discussions

3.1 Identification of outflows

Figure 1: (a) The 1.3 mm continuum map of G34 is depicted in the colourscale. The peak position of HMC is marked with a ‘$\times$’. The contours of the 3.6 cm emission is overlaid in black with contour levels 5, 50, 100, 200, 400, 800, 1200 times $\sigma$ ($\sigma=0.2$ mJy beam^-1). The two HC Hii regions, A and B, and the UC Hii region, C are indicated by ‘+’. The arrow indicates the direction towards the shell-like Hii region, D. The beams of the 1.3 mm and 3.6 cm continuum are shown at the bottom-right and -left corners, respectively. (b) SiO($5-4$) moment-one map of G34 within the velocity range 0 to 120 km s^-1 overlaid with the receding (blue curves) and approaching (red curves) CO($2-1$) outflow streamers identified. The contours of the 1.3 mm emission is also overlaid in gray with contour levels 10, 80, 160, 320, and 640 times $\sigma$ ($\sigma=0.5$ mJy beam^-1). The positions of the HMC and the UC Hii region are marked with ‘$\times$’ and ‘+’, respectively. A portion of the Hii region, D is depicted by the dashed ellipse. The beam of the SiO($5-4$) cube is shown at the bottom-left corner. The green box denotes the field-of-view of (a). (c) Two-colour composite moment-zero maps of CO($2-1$) integrated over the velocity ranges 0 to 48 km s^-1 (blue) and 68 to 120 km s^-1 (red). The red and blueshifted outflow streamers are show in cyan and pink, respectively.

The ALMA 1.3 mm continuum map of G34 is shown in Figure 1(a). The peak position of the continuum emission ($\alpha_{J2000}=18^{\rm h}53^{\rm m}18.56^{\rm s}$,$\delta_{J2000}=+01^{\circ}14^{\prime}57.90^{\prime\prime}$) is determined using the 2D Gaussian fitting tool of CASA viewer. From the fitting to the central dense core emission, we obtain an integrated flux density of 9.3$\pm$0.5 Jy and a peak flux of 4.5$\pm$0.1 Jy beam^-1. The HMC identified by Mookerjea et al. (2007) at 2.8 mm coincides with the 1.3 mm peak. The contours of the radio continuum emission at 3.6 cm is overlaid on this figure. The positions of the two HC Hii regions (A and B) and the extended cometary UC Hii region (C) obtained from Sewilo et al. (2004) are marked. The direction towards the shell-like Hii region (D) is also labelled. The 1.3 mm peak coincides with the UC Hii region, C (within $\sim 0^{\prime\prime}.5$).

In Figure 1(b) we present the SiO($5-4$) moment-one (intensity-weighted velocity) map towards G34 within the velocity range 0 to 120 km s^-1, overlaid with the contours of the 1.3 mm continuum emission. The velocity range chosen to construct the SiO($5-4$) moment-one includes radial velocities close to the systemic velocity of the ambient gas of the G34 cloud ($V_{\rm sys}=+58$ km s^-1; Hoang et al. 2023). Since SiO emission primarily traces shocks, the contamination from the ambient gas is considered minimal. To trace the large-scale outflows around the G34 complex we use the CO($2-1$) data cube, with which we construct channel maps (Figure 6) having a channel width of 2 km s^-1. Each channel shows several localized emission features. The positions of these features are determined by linearized least-square fits to Gaussian ellipsoids using the task SAD of the Astronomical Image Processing Software (AIPS). Examining these features at consecutive velocity channels within the velocity window 0 to 120 km s^-1, we have discerned 20 filamentary gas streamers with consistent velocity increments and having almost linear structures. The velocity channels in range 48 to 68 km s^-1, where the emission is dominated by the ambient cloud and has spatially extended structures, are excluded while extracting the outflow streamers. Each streamer traces a sequence of CO($2-1$) condensations. Of the outflow streamers identified, 9 are receding (redshifted) reaching radial velocities of up to 62 km s^-1, and 9 of them are approaching (blushifted) reaching up to $-58$ km s^-1 with respect to 58 km s^-1, the systemic velocity of G34. Most of these outflow streamers, depicted by red and blue curves in Figure 1(b), nearly follow straight lines. Both the red and blue streamers appear to overlap on the plane of the sky and seem to be radially distributed from a common centre as their origin. The SiO($5-4$) emission also reveals filaments tracing the CO streamers and pointing back to the origin of the outflow.

The CO($2-1$) moment-zero map of G34 integrated over the velocity ranges 0 to 48 km s^-1 and 68 to 120 km s^-1 is presented in the two-colour composite image in Figure 1(c). The outflow streamers identified are also marked. The figure shows a few possible bipolar outflows likely originating from low-mass protostars. We have taken care to avoid these outflows while extracting the outflow streamers. CO protostellar outflows originating from the HMC have also been detected from the “Querying Underlying mechanisms of massive star formation with ALMA-Resolved gas Kinematics and Structures (QUARKS; Liu et al., 2024)” survey (K. Huang et al., in prep). However, these outflows are not detected at the resolution of the ALMA data presented in this paper. Additionally, there are no large-scale outflow streamers observed along the direction of the protostellar outflows from the HMC. Thus, it is unlikely that any of the outflow streamers identified are misnomered.

Figure 2: Redshifted (a) and blueshifted (b) CO($2-1$) condensations detected in each CO($2-1$) velocity channel towards G34. The velocity of the redshifted emission ranges from 68 to 120 km s^-1 and of the blueshifted emission from 0 to 48 km s^-1. The median position of the intersection points of the outflows which is the possible origin of the outflow is indicated by the brown triangle. The position of the UC Hii region is denoted by the black star. The red and blue dotted lines in (a) and (b), respectively show the different orientations of the streamers with respect to the origin of the outflow.

We plot the CO($2-1$) condensations identified from each velocity channel of the CO($2-1$) cube, with respect to the position of the cometary UC Hii region, C in Figure 2(a) and (b). The red (RF1-RF9) and blueshifted (BF1-BF9) streamers are labeled in this figure. Following Guzmán Ccolque et al. (2024), we find the position of the origin of the outflows. By performing a linear fit on all the CO streamers, we created a dataset of intersection points for each pair of streamers. Excluding the intersection points more than 5^′′ away from the HMC and the UC Hii region, we are left with 7 blue streamers (BF1, BF2, BF5, BF6, BF7, BF8, and BF9) and 6 red streamers (RF1, RF2, RF3, RF4, RF7, and RF8). The origin of the outflow is derived by estimating the median position of the intersection points of these 13 streamers and is found to be located at $\alpha_{J2000}$=18^h53^m18^s.63$\pm$0.06${\rm s}$, $\delta_{J2000}$=01^∘14^′56^′′.56$\pm$0.4^′′. This position lies within the UC Hii region towards the south-east edge at a separation of $\sim 1^{\prime\prime}.6$ from the peak position of the HMC. The red and blue dotted lines in Figure 2, which represent the least-square fits of all the streamers along with the position of the origin, traces the orientation and path of each outflow streamer.

Figure 3: (a) Colour-composite image of G34 from IRAC 3.6 $\mu$m (blue), 4.5 $\mu$m (green), and 8.0 $\mu$m bands. (b) IRAC [4.5]/[3.6] flux ratio map with contour levels 1.5, 2.0, and 2.5. The image is smoothed to 3 pixels to improve the contrast against the background. The H₂ jets identified from both these images are depicted by the dashed lines. The red and blushifted CO outflow streamers are also overlaid. The hollow star and triangle mark the positions of the UC Hii region and the possible origin of the outflow, respectively.

In addition to the molecular outflow traced by the CO($2-1$) and SiO($5-4$) lines, we see evidence of multiple jets in the mid-infrared regime in the G34 region. Figure 3(a) is the three color-composite image from the GLIMPSE survey. The emission in the IRAC 4.5 $\mu$m (color-coded as green in the IRAC color-composite images) shows several finger-like diverging structures, resembling multiple jets. Towards the south-east of the extended 4.5 $\mu$m emission in Figure 3(a) is the infrared dust bubble MWP2G0342631+0013065 (Jayasinghe et al., 2019) which bounds the shell-like Hii region, D (Liu et al., 2013). Khan et al. (2024) suggest that this bright infrared emission dominated by the 8 $\mu$m emission results from the expansion of the Hii region that compresses the gas around it, leading to the formation of a shock front. To reduce the contamination from the stellar components at 4.5 $\mu$m, we construct the IRAC [4.5]/[3.6] flux ratio map, presented in Figure 3(b). Along the direction of the jet-like features seen in the 4.5 $\mu$m emission, the flux ratio is $\gtrsim$ 1.5, as opposed to stars with flux ratio $\ll$ 1.5 (Takami et al., 2010). Such emission has been interpreted to be tracing the shock-excited 0-0 S(9) line of H₂ at 4.695 $\mu$m (Noriega-Crespo et al., 2004). Comparing the two images, we have visually identified several jets (jet 1 - jet 8), depicted by the dashed lines in both figures that seem to originate from the G34 complex. The CO outflow streamers are also overlaid on Figure 3. The 4.5 $\mu$m jets and the outflow streamers show a good correlation with the jets also roughly pointing towards the origin of the outflow streamers depicted by the triangle in Figure 3.

3.2 Evidence of explosive outflows

The ¹²CO($2-1$) and infrared observations towards Orion BN/KL have revealed a massive (10 M_⊙) and energetic ($\sim 10^{47}$ erg) outflows produced by a violent explosion likely caused by an N-body interaction resulting in the ejection of the stars BN, “source I”, and “source x” about 550 yr ago (Luhman et al., 2017; Bally et al., 2020). Zapata et al. (2009) suggest that such an isotropic distribution of CO outflows and H₂ finger-like emission is unlike a typical protostellar outflow seen in star-forming regions. Similar results have been reported for DR21 as well by Zapata et al. (2013) and Guzmán Ccolque et al. (2024), wherein they propose that the CO and H₂ emission maybe driven by an explosive event that occurred about 8,600 yr ago. Considering the distribution of CO and H₂ emission around the G34 complex, one can envisage a similar scenario as that of Orion BN/KL and DR21 in G34 as well. However, the shock front from the expansion of the Hii region, D could influence the dynamics of the outflow streamers in G34 and thus affect the isotropy of the outflow distribution expected for an explosive outflow.

Figure 4: Position-velocity diagram of the outflow streamers identified towards G34. Each streamer identified is designated a different colour and labelled as RF1-RF9 and BF1-BF9 for the red and blueshifted streamers, respectively. The gray lines indicate the linear trend between the projected distance and radial velocity starting from the projected distance of 0^′′ corresponding to the origin of the outflow and 58 km s^-1, the systemic velocity G34. The shaded portion indicates the region where the emission is contaminated by the ambient cloud.

The relation between on-the-sky distance and radial velocity of the 20 identified outflow streamers is plotted in Figure 4. The gray lines represent the linear trend between the projected distance and the radial velocity predicted by the Hubble-Lemaître velocity law. The outflow streamers seem to qualitatively follow this linear trend where the radial velocity of the CO($2-1$) condensations increase with the projected distance from the common centre following the Hubble-Lemaître velocity law. The kinematic behavior of a linear increase in radial velocity with the projected distance is regarded as one of the most distinctive signatures of explosive outflows and have been confirmed in the cases of other explosive outflows (e.g., Zapata et al., 2009, 2013, 2020, 2023; Guzmán Ccolque et al., 2024). However, the linear trend is not very clear for some of the streamers which have large velocity dispersion (e.g., BF1, RF6, RF8). A deviation from the linear trend may be attributed to the shocks originating from the interaction of the outflows with the surrounding material (e.g., Guzmán Ccolque et al., 2024). The high-mass star-forming region, G34 is classified as a hub-filament system with ongoing mass accretion through filaments (Khan et al., 2024). Some of the filaments lie along the direction of the outflow streamers (refer to Figure 6 of Khan et al., 2024). The interaction of the outflows with the infalling gas would produce strong shocks resulting in steeper velocity gradients in the position-velocity plot. This scenario is supported by the presence of strong SiO emission, an excellent shock tracer, along the outflow streamers (see Figure 1b).

Assuming local thermodynamical equilibrium and the ¹²CO($2-1$) emission to be optically thin at velocities beyond $\pm$10 km s^-1from $V_{\rm sys}$, we estimate the mass, momentum, and energy of the explosive outflow following Equations (A1-A4) from Li et al. (2020). Taking a distance of 3.3 kpc, excitation temperature of 29 K (Urquhart et al., 2018) and CO abundance of 10^-4 (Blake et al., 1987), these values are calculated for the streamers in each channel and then summed to obtain a total mass, momentum, and outflow energy of $\sim$264 M_⊙, 4.3$\times 10^{3}$ M_⊙ km s^-1, and 10⁴⁸ erg, respectively. This implies that the outflow emission in G34 is associated with a highly energetic event. The estimated outflow energy of 10⁴⁸ erg is similar to the energies derived for the previously identified explosive outflow events. Furthermore, the initial explosion energy would be much larger than the outflow energy because most of the it would have been radiated away by the shocks.

From the emission at 1.3 cm and 2.8 mm, Mookerjea et al. (2007) estimated the spectral index of the UC Hii region to be 0.2, consistent with optically thin emission. Assuming the emission to be optically thin at 3.6 cm as well, we estimate the Lyman continuum flux, $N_{\rm Ly}$ to be 3.4$\times 10^{48}$ s^-1 that translates to an ionizing ZAMS star of spectral type O7-O7.5 (Panagia, 1973). We compute the dynamical age of the UC Hii region from the 3.6 cm map using the following expression.

$$t_{\rm dyn}=\bigg{[}\frac{4\,R_{\rm s}}{7\,c_{\rm i}}\bigg{]}\bigg{[}\bigg{(}\frac{R_{\mbox{H{\sc ii}}}}{R_{\rm s}}\bigg{)}^{7/4}-1\bigg{]}$$ (1)

Here $R_{\rm s}=(3N_{\rm Ly}/4\pi n_{0}^{2}\alpha_{\rm B})^{1/3}$ is the radius of the Strömgren sphere, where $n_{0}=1.6\times 10^{5}$ cm^-3 is the number density of the neutral hydrogen medium. Since a large fraction of diffuse emission is lost in the high-resolution 1.3 mm map due to missing flux effects during interferometric observations and because the 1.3 mm emission is also contaminated by free-free emission from the UC Hii region (Liu et al., 2013), we estimate $n_{0}$ from the single-dish JCMT 850 $\mu$m map. $\alpha_{\rm B}$ is the radiative recombination coefficient taken to be $2.6\times 10^{-13}$ cm³ s^-1 (Kwan, 1997). $R_{\mbox{H{\sc ii}}}=(A/\pi)^{0.5}$ is the effective radius of the UC Hii regions, where $A$ is the area within the 3$\sigma$ ($\sigma=0.5$ mJy beam^-1) contour of the UC Hii region. $c_{\rm i}$ is the isothermal sound speed in the ionized medium, typically assumed to be 10 km s^-1. The dynamical age of the UC Hii region is estimated to be $\sim$17,000 yr.

We also estimate of the dynamical age of the outflow assuming that all the streamers move with the same velocity. This gives a range of maximum radial velocities for the outflows with varying inclinations relative to the line-of-sight. Taking the maximum observed radial velocity with respect to the systemic velocity of G34, $\sim$62 km s^-1, representing the outflow closest to the line-of-sight, and a maximum projected distance of $\sim 74^{\prime\prime}$ which is the farthest outflow, the dynamical age of the outflow is estimated to be $\sim$19,000 yr. It is to be noted that this gives an order of magnitude estimate at best, considering we have not taken into account the inclination angle of the streamers with the line-of-sight and have assumed an isotropic distribution of the outflow streamers.

The positional coincidence between the centre of the outflow streamers and the UC Hii region (Section 3.1) along with their similar dynamical timescales suggest a possible relationship between the expansion of the UC Hii region and the outflow, where the two were probably triggered by the same explosive event. Nonetheless, we cannot dismiss the possibility that the UC Hii region predates the explosive event, given the large uncertainty in the age estimation of the explosive outflow.

Similar expanding Hii regions are also seen within the explosive outflows in DR21 (Zapata et al., 2013; Guzmán Ccolque et al., 2024), G5.89$-$0.39 (Zapata et al., 2019, 2020), and IRAS 12326$-$6245 (Zapata et al., 2023) where the expanding shell is centred at the origin of the outflows. These studies suggest that the explosive event that drives the outflows is also responsible for powering the expanding Hii regions. Currently there are no known massive young stellar objects (MYSO) at the peak of the UC Hii region in G34. It is possible that the protostar may have been ejected during the explosive event, as in the case of Orion BN/KL (Bally et al., 2011; Rodríguez et al., 2020) and DR21 (Zapata et al., 2013; Guzmán Ccolque et al., 2024). An in-depth proper motion study would be required to confirm this. Furthermore, several compact hot cores are detected along the periphery of the cometary head of the UC Hii region (K. Huang et al., in prep) which could indicate triggered star formation post explosion.

All the characteristics including the Hubble-Lemaître-like expansion motion of the streamers and high outflow energy suggest that G34 is a likely explosive outflow candidate.

3.3 Outflow-dragged magnetic field

Figure 5: The 850 $\mu$m Stokes I map of the region surrounding G34 from JCMT SCUBA-2/POL-2 with contour levels 5, 10, 25, 65, 100, 200, and 500 $\sigma$ ($\sigma=0.04$ Jy beam^-1). The purple line segments represent the magnetic field direction as inferred from the polarization data. The red and blushifted CO outflow streamers (red and blue curves) and the H₂ jets identified from Figure 3 (dashed lines) are also overlaid. The hollow star and triangle mark the positions of the UC Hii region and the possible origin of the outflow, respectively. The beam is shown on the bottom-left corner.

In Figure 5 we present the 850 $\mu$m Stokes I map from the SCUBA-2/POL-2 archival observations towards G34. The magnetic field lines inferred from the polarization data are overlaid on this figure, along with the CO outflow streamers and H₂ jets. Even with the coarse resolution of the single-dish polarization observations, one can see that the magnetic field is approximately oriented along the direction of the outflows streamers and jets. To quantify the relative alignment of the magnetic field lines to the H₂ jets, we calculate the angle difference between the directions of the two (see Appendix B). Figure 7 shows a plot of the same, where we see that the angle difference for most magnetic field lines and H₂ jets fall within $-20^{\circ}$ to $20^{\circ}$. It is to be noted most of the magnetic field lines along jet 1, jet 2, and jet 6 have angle difference $>|20^{\circ}|$. Nonetheless, the median value of the angle difference between the magnetic field lines and the H₂ jets is $\sim-5^{\circ}$. This prompts us to infer that the explosive outflow could be responsible for aligning the magnetic field along the outflow direction.

Of the six explosive outflow sources identified, magnetic fields have been reported towards Orion BN/KL (Cortes et al., 2021) and G5.89$–$0.39 (Fernández-López et al., 2021) from high angular resolution dust polarization observations using ALMA. From the 1.3 and 3.1 mm dust polarization observations Cortes et al. (2021) found the magnetic field to be orientated quasi-radially within $\sim$5000 au from the origin of the explosive outflow. The outflows carve cavities in the dust resulting in the polarized dust emission to have an anti-correlation with the outflow streamers and the magnetic field is aligned in the direction of the cavities. Evaluating the energy balance, they estimate that the explosive outflow may be energetic enough to propel a shock from the centre of the Orion BN/KL nebula which can drag the magnetic field lines and rearrange them in a quasi-radial orientation in the inner radius of the outflow ($\sim$5000 au). In the case of G5.89$–$0.39, Fernández-López et al. (2021) have found that the magnetic field follows a radial pattern towards the ‘Central Shell’ at 1.2 mm which coincides with the shell-like UC Hii region, similar to Orion BN/KL. These authors suggest that such a radial distribution of magnetic field could be a signpost of explosive outflow events. These studies lend support to our inference that the magnetic field lines in G34 are also dragged by the explosive outflow event. While the magnetic field may be dragged by the explosive outflow, we cannot rule out the possibility that the magnetic field dominates the orientation of the outflows in this region. However, to obtain a better insight requires evaluating the energy balance between the magnetic field and the outflow energies (e.g. Cortes et al., 2021). This advocates for higher angular resolution polarization observations to enable a reliable estimate of the magnetic field energy.

3.4 Rate of explosive events in the Milky Way

From the six explosive outflows reported in literature (Orion BN/KL, DR21, G5.89-0.39, IRAS 16076-5134, Sh2-106, and IRAS 12326-6245), Zapata et al. (2023) have estimated the rate of explosive events to be one in every 90 yr in the Milky Way. These authors have also noted that this rate is comparable to the approximate rate of supernovae of one in 50 yr (Diehl et al., 2006) which is also similar to the massive star formation rate of one in 50 yr. With the new detection of an explosive outflow in G34, we update the rate of events in our Galaxy following the same method described in Zapata et al. (2023). Assuming that the explosive events are evenly spaced over a time span of 31,560 yr (the time period covering all the seven outflows and taking into consideration their different distances to earth) and are distributed within a projected circle of radius 2.8 kpc (the separation between IRAS 16076-5134 and DR21, the farthest separated outflow sources), we extrapolate the frequency of occurrence of explosive events to the disk of our Galaxy which is taken to be a thin disk with a radius of 15 kpc. This gives a total number of 200 explosive events in the Galaxy and the rate of explosive events to be one in every 160 yr. Our estimate is higher than that reported by Zapata et al. (2023), owing to the larger dynamical age of the explosive outflow in G34 compared to the other six explosive outflows detected so far. This, however, is a crude estimate based on several assumptions, including the size of the Galaxy and an approximate dynamical age of the outflow. The rate of occurrence explosive events in the Milky Way can be refined as more explosive outflows are detected, and hence, can draw a better correlation with the rate of supernovae and massive star formation.

4 Conclusions

The ALMA CO($2-1$) and SiO($5-4$) archival observations have revealed the presence of multiple outflow streamers associated with the high-mass star-forming complex, G34. Along with molecular outflows at mm wavelengths, several H₂ jets have also been identified from the IRAC 4.5 $\mu$m and [4.5]/[3.6] flux ratio maps showing a good correlation with the CO outflow streamers. The molecular outflow streamers have nearly linear structures and seem to emanate from a common centre within the UC Hii region. The radial velocity of each streamer follows the Hubble-Lemaître velocity law, indicative of the explosive nature of the outflow. The explosive event that initiated the outflow appears to have occurred about 19,000 yr ago. This event may also be the mechanism that powers the expanding UC Hii, as indicated by their similar dynamical ages and the positional coincidence of the UC Hii region with the origin of the outflow. Additionally, these explosive outflows might be responsible for aligning the magnetic field along the outflow direction. Our results add to the small sample of rare explosive outflows observed in our Galaxy.

Appendix A Channel maps

The CO($2-1$) channel maps of width 2 km s^-1 in the velocity range 0 to 50 km s^-1 and 70 to 120 km s^-1 are shown in Figure 6(a) and (b), respectively.

Figure 6: The ALMA ¹²CO($2-1$) channel maps from in the velocity range 0 to 50 km s^-1 (left) and 70 to 120 km s^-1 (right) towards G34. The CO condensations in each channel having a nearly linear structure with velocity increment are identified as the outflow streamers.

Appendix B Direction of magnetic field lines v/s H₂ jets

Figure 7: The angle difference between of the magnetic field lines ($\theta_{B}$) and the H₂ jet direction ($\theta_{\rm jet}$) along each H₂ jet identified in G34.

To quantitatively evaluate the alignment of the magnetic field lines with the H₂ jets, we compute the angle difference between the two. We select the magnetic field lines along each H₂ jet and determine the difference in angle between the direction of each magnetic field line ($\theta_{B}$) and the direction of the corresponding H₂ jet ($\theta_{\rm jet}$). This is repeated for each H₂ jet and the resulting difference in orientation between the magnetic field lines and each H₂ jet is plotted in Figure 7.