The First Detection of a Protostellar CO Outflow in the Small Magellanic Cloud with ALMA

The color-scale image in Figure 1(a) shows the 0.87 mm continuum emission toward the Spitzer source Y246. We performed 2D Gaussian fitting to the source emission, and obtained the following parameters: a central coordinate of (α_J2000.0, δ_J2000.0) = (0^h54^m03439, −73°19'38528), a beam-deconvolved size of 069 × 048 (= 0.21 pc × 0.14 pc), with a position angle of 27°. The difference between the source position determined by Spitzer and the dust peak is ∼04, which is not a significant offset. The distribution of the continuum emission strongly suggests that it is emitted from the dense envelope gas, mainly as thermal dust emission, associated with the protostellar system. The total 0.87 mm flux (F_{0.87 mm}) is 6.9 mJy. Assuming a uniform dust temperature (T_d) of 20–30 K (Takekoshi et al. 2017), a gas-to-dust ratio of 1000 (e.g., Roman-Duval et al. 2014), and a dust opacity per dust mass of 2 cm² g⁻¹ (Ossenkopf & Henning 1994), the F_{0.87 mm} of Y246 can be converted into 0.7 × 10³ M_⊙ (for T_d = 30 K) and to 1.2 × 10³ M_⊙ (for T_d = 20 K). The velocity width in FWHM of the SO profile at the 0.87 mm peak is 3.4 km s⁻¹. With a radius of 0.17 pc, the estimated virial mass is ∼10³ M_⊙ using the method in MacLaren et al. (1988) assuming a uniform density distribution. Since the virial mass is about the same or smaller than the dust-based mass, it is reasonable to assume that this high-density clump is gravitationally bound. The amount of surrounding gas is large enough to be consistent with the fact that the protostar is in a young evolutionary stage (see Section 2).

Zoom In Zoom Out Reset image size

The CO(3–2) data enables us to probe molecular outflow around the Spitzer source. We visually inspected the presence/absence of the outflow wing emission, whose maximum velocity is apart by more than ∼10 km s⁻¹ from the systemic velocity of 163 km s⁻¹ determined by the Gaussian fitting of the SO profile. The blueshifted and redshifted CO wing emission in Figure 1(b) shows a maximum velocity (${v}_{\max }$) of ∼16 km s⁻¹ and ∼17 km s⁻¹, respectively. We superimposed the high-velocity emission integrated over the velocity range of 147–158 km s⁻¹ (blueshifted) and 168–180 km s⁻¹ (redshifted) on the 0.87 mm continuum map (Figure 1(a)). Although the blueshifted profile extends to lower velocity, ∼160 km s⁻¹, we did not integrate the velocity range due to strong contamination from around the systemic velocity components. The blue and redshifted components roughly compose a northwest to southeast elongation centered at the continuum peak, which corresponds to the position of the embedded source. These spatial distributions and velocity features are quite consistent with known outflow sources in the Galaxy and LMC (see Sections 1 and 4.1) and nicely illustrate the protostellar, molecular outflow from the embedded source. Note that no other high-velocity components similar to those in the 0.87 mm source were detected within the observation field of the ∼25'' diameter.

We estimated the physical properties of the outflow lobes. The projected length of the lobes (R_lobe) is ∼0.3–0.4 pc, based on the distance between the edges of the outflow lobes and the continuum peak. The dynamical time (t_d = R_lobe/${v}_{\max }$) of the lobes is 2 × 10⁴ yr (without inclination angle correction), indicating the protostellar system is in a young evolutionary stage, consistent with the spectroscopic analysis by Oliveira et al. (2013). The CO(3–2) luminosities of the blue and redshifted lobes are ∼4 K km s⁻¹ pc² and ∼2 K km s⁻¹ pc², respectively. Although the CO-to-H₂ conversion factor is not well constrained in the SMC-like low-metallicity environment, we applied the recently derived value of 7.5 × 10²⁰ cm⁻² (K km s⁻¹)⁻¹ (Muraoka et al. 2017) to estimate the flow masses (M_flow) as ∼70 M_⊙ (for the blue lobe) and ∼30 M_⊙ (for the red lobe).

We compare the physical properties of the newly discovered SMC protostellar outflow with those in the Galactic and LMC studies based on an order estimation. The mass ejection rate (M_flow/t_d) is ∼10⁻³ M_⊙ yr⁻¹, which is consistent with the fact that the protostellar source is in a young evolutionary stage of high-mass star formation. The outflow momentum, P_flow = ${M}_{\mathrm{flow}}{v}_{\max }$, is ∼10³ M_⊙ km s⁻¹, and the mechanical force, F_flow = P_flow/t_d, is ∼10⁻² M_⊙ km s⁻¹ yr⁻¹. These outflow properties are roughly consistent with those seen in Galactic protostars (see Figure 5 in Beuther et al. 2002 and Figure 7 in Maud et al. 2015) and in solar-metallicity MHD simulations (see Figures 5 and 6 of Matsushita et al. 2018), at the source luminosity of ∼10⁴ L_⊙. Tokuda et al. (2019) also confirmed this relation in a luminous YSO in the LMC N159W-South region. Thus, we conclude that the outflow properties may be universal across the metallicity range of 0.2–1 Z_⊙, based on current observational knowledge of the Galaxy, LMC, and SMC.

We would like to note that, as mentioned in Section 1, it is nontrivial whether disks form and drive outflows in low-metallicity environments from a theoretical perspective. The detection of the molecular outflow in Y246 suggests that an accretion disk exists under a strong magnetic field in the unresolved innermost region, even in the SMC with 0.2 Z_⊙. Galactic observations indicate that magnetic fluxes and angular momenta in prestellar cores are several orders of magnitude larger than those of protostars (e.g., Goodman et al. 1993; Crutcher 1999), known as the magnetic-flux and angular-momentum problems. Theoretical studies on Galactic star formation have investigated the mechanisms to remove the large amounts of angular momentum and magnetic flux in star formation. An excess angular momentum is transported by protostellar outflow driven by the outer edge of the disk, while nonideal MHD dissipation processes can remove magnetic flux from a star-forming core and promote the formation of a rotationally supported disk (e.g., Dapp & Basu 2010). In addition, the dissipation of magnetic fields is closely related to disk fragmentation, and the formation of binary or multiple stellar systems (e.g., Machida et al. 2008). As the dissipation rate of magnetic fields is suggested to be controlled by the metallicity, the outflow intensity and binary frequency should also depend on the metallicity (e.g., Susa et al. 2015). Moreover, the dust abundance affects radiative heating and cooling rates, changing the gravitational stability of accretion disks (e.g., Matsukoba et al. 2022). Therefore, it is crucial to understand the metallicity dependence of radiative MHD dynamics in star formation. However, so far theorists have primarily assumed solar metallicity, and only a few studies have focused on the outflow driving and disk formation in low-metallicity environments (Higuchi et al. 2018; Tanaka et al. 2018). Thus, to comprehensively understand the star formation process across various environments, we must shed light on low-metallicity star formation in future theoretical studies.

The high-mass protostar Y246 is currently only one example in the SMC, but it will be an important benchmark for future theoretical and observational studies of low-metallicity star formation. To better understand star formation dynamics in the low-metallicity SMC environment, we will need to examine the physical properties of the surrounding dense envelope, e.g., specific angular momentum and infall signature through higher-resolution follow-ups.

It has been seven years since the first extragalactic protostellar outflow detection in the LMC (Fukui et al. 2015). We here discuss why such a discovery has been delayed in the SMC. We mainly point out three possibilities: the sample size, the tracer effect, and scarcity in nature.

As for the sample size, the known YSO targets with spectroscopic confirmation in the LMC and SMC are 277 (Seale et al. 2009) and 33 (Oliveira et al. 2013), respectively. The difference of one order of magnitude is proportional to the sizes of the galaxies and their CO-detectable regions (e.g., Fukui & Kawamura 2010). The large sample size of luminous (≳10⁴ L_⊙) YSOs in the LMC naturally explains the high encounter probability, as higher YSO luminosities generally tend to show energetic, detectable outflows.

Previous CO observations mainly studied J = 1–0 and 2–1 transitions in the SMC's molecular clouds, i.e., the tracer effect may give a reason for the outflow nondetection results. For example, the CO(2–1) observations by Muraoka et al. (2017) could not detect the molecular outflow from the high-mass protostar in the SMC N83C, although the target is as bright as ∼10⁵ L_⊙ and the observational setting was similar to the LMC studies (e.g., Fukui et al. 2015). We spatially smoothed our Y246 CO(3–2) data set into an angular resolution of 17 × 13, which is the same as Muraoka et al.'s (2017) observation on N83C (see Figure 2, upper panel). The resultant brightness temperature of the wing emission is ≳1 K, which can be detected with >3σ at the Muraoka et al. (2017) sensitivity assuming a CO(3–2)/CO(2–1) ratio of ∼1. The nondetection in N83C was probably attributed to a weaker low-J CO emission in high-velocity outflow components. Brightness temperatures of higher-J transitions become higher than those in lower J ones in the case of being optically thin and high temperature (>20–30 K) under local thermodynamical equilibrium conditions. Such environments are rarely seen in typical CO-detectable molecular clouds, but some local regions show stronger CO(3–2) and higher J emission than lower-J CO lines, e.g., at a cloud edge (Tachihara et al. 2012) and protostellar outflows (e.g., Shimajiri et al. 2008; Lefloch et al. 2015). This behavior may be further enhanced in the SMC because CO lines trace denser and warmer gas in lower-metallicity environments (e.g., Muraoka et al. 2017; Tokuda et al. 2021) than in typical giant molecular clouds in the Galaxy (e.g., Nishimura et al. 2015). We thus speculate that CO(3–2) worked favorably as an outflow tracer in the SMC. We note that the available CO(3–2) data in N83C (Muraoka et al. 2017) has a further coarse angular resolution of ∼4'', which did not allow us to detect weak, compact emissions due to the beam-dilution effect. Based on a smoothed spectra of Y246 (Figure 2, lower panel), it becomes more difficult to detect the high-velocity emission with a relative velocity of ≳10 km s⁻¹.

Zoom In Zoom Out Reset image size

As mentioned in the previous paragraph, the behavior of CO as a dense-gas tracer may explain the scarcity of the outflow detection in the SMC. The lobe size of the Y246 outflow (0.3–0.4 pc) is considerably smaller than the pc-scale CO outflow in the outer Galactic region (Fernández-Martín et al. 2017). The time and spatial scales with observable CO outflows may be limited in the SMC, which may further reduce the probability of detection. However, our preliminary analysis found one or two more wing feature candidates (#03, and 33) in the current ALMA sample from the Oliveira et al. (2013) catalog, and thus it is premature to conclude that CO outflows are rare in the SMC.

Further multiobject with multitransition observations will be anticipated to obtain a convincing population of SMC outflows and to investigate their similarities/differences compared to those in the Galaxy and the LMC.