Crescent-shaped Molecular Outflow from the Intermediate-mass Protostar DK Cha Revealed by ALMA

We show the 1.3 mm continuum and ¹²CO(J = 2–1) emission in this subsection. The white contours in the left panel of Figure 1 show the 1.3 mm continuum emission. The emitting region is compact, comparable to the beam size. A two-dimensional Gaussian fit to the image gave a deconvolved size of (503 ± 23 mas) × (306 ± 22 mas) and a position angle of 1288 ± 53. The total flux is 562.4 ± 5.4 mJy. These values are consistent with those reported by Villenave et al. (2021), who used other ALMA archival data with a higher angular resolution of 051 × 028. The cyan and red colors represent the blueshifted and redshifted ¹²CO(J = 2–1) integrated intensity distributions, respectively. The right panels of Figure 1 show the average spectra of blueshifted (top) and redshifted (bottom) components. We defined the ¹²CO integration ranges of the maps as follows: the high-velocity sides are the velocity at which the ¹²CO emission is prominently detected judging from the averaged profiles, and we excluded the lower-velocity range around the systemic velocity of ∼3 km s⁻¹(Yıldız et al. 2013) with our ¹³CO detection, likely containing the infalling envelope around the protostar (see Section 3.2). The blueshifted wing of the ¹²CO spectrum extends from < − 4 to −71 km s⁻¹, and the redshifted component also has high-velocity emission (7.5–55 km s⁻¹). The blueshifted components have arc-like structures with a size of several thousand astronomical units, and they surround the continuum source and have a stripe pattern. In particular, the arc-like structure close to the continuum source resembles a crescent moon. The redshifted components also show a similar stripe pattern.

Zoom In Zoom Out Reset image size

To investigate whether the crescent-shaped structure is an interferometric artifact, i.e., due to the missing flux, we compared our data with the data from a single-dish telescope. Using the 15 meter Swedish-ESO Submillimeter Telescope, Knee (1992) showed that the ¹²CO(J = 1–0) high-velocity emission is less than ∼0.1 K at a velocity of −20 km s⁻¹ with an angular resolution of ∼40''. We measured the CO brightness temperature across the entire field of view (∼40'' × 40'') at −20 km s⁻¹ in our data and confirmed that the brightness temperature in our data is consistent with that in Knee (1992). Thus, the CO crescent-shaped structure seen in the high-velocity component is not likely to be an overemphasized feature due to the interferometric effect. As mentioned in Section 1, recent high-resolution molecular gas observations have reported peculiar structures associated with protostars, such as streamers and dynamical interactions among multiple sources. They are mainly identified in dense gas tracers, e.g., C¹⁸O and HCO⁺, and their relative velocities are close to the systemic velocity within a few kilometers per second. However, we identified the stripe pattern composed of the multiple arc-like structures in the low-density tracer of ¹²CO, and the velocities are too high to be explained by accretion motion. We regard these components as outflow motions, as already suggested in the previous single-dish works (e.g., Knee 1992). The shape deviates significantly from the usually identified conical-type feature and can probably be attributed to the viewing angle effect (i.e., close to the face-on configuration). The interpretation and further discussion of the outflow will be given in Section 4.1.

Figure 2 shows velocity-channel maps of the ¹²CO emission in the range −60–30 km s⁻¹. The range of the velocity integration is 10 km s⁻¹ for each channel. The blueshifted high-velocity component (≲ − 40 km s⁻¹) has a crescent-shaped structure and is distributed slightly northeast of the continuum source. The blueshifted low-velocity component (−20 ≲ v_LSR ≲ 0 km s⁻¹) has a similar shape but surrounds the continuum source, and the emitting region seems to be more extended than the high-velocity component. The redshifted component (≳10 km s⁻¹) has a similar arc-like structure to the blueshifted one, with the opposite orientation.

Zoom In Zoom Out Reset image size

To verify the characteristics of the outflow described above in more detail, we generated ¹²CO moment 1 maps for the blueshifted and redshifted components (Figure 3). The figure clearly shows the two-dimensional velocity structure and/or gradient of the outflow, especially for the blueshifted component. A blueshifted compact high-velocity emission is located on the northeast side of the continuum source, while a more extended low-velocity emission is distributed on the opposite side, forming a layered velocity gradient across the crescent shape as a whole.

Zoom In Zoom Out Reset image size

One of the densest regions (A_V ∼ 20 mag and ${N}_{{{\rm{H}}}_{2}}\,\gt {10}^{22}$ cm⁻²) in the Chamaeleon II cloud is located in the vicinity of DK Cha (e.g., Alcalá et al. 2008; Alves de Oliveira et al. 2014). Compared to the blueshifted outflow, the velocity gradient of the redshifted outflow is not clear. In addition, the redshifted outflow velocity is slower than the blueshifted outflow velocity. Thus, DK Cha seems to be in front of the dense region, and the redshifted components could be decelerated due to the interaction with the surrounding dense gas.

Figure 4 (a) plots the ¹³CO(J = 2–1) moment 1 and the overlaid peak temperature with the gray contours. The ¹³CO image shows an envelope structure with a spatial extent of a few thousand astronomical units. The strong emission on the northeast side is blueshifted (∼1.5 km s⁻¹), and that on the southwest side is redshifted (∼4.5 km s⁻¹). This direction is similar to the expected outflow axis. Figure 4 (b) shows a position–velocity (PV) diagram along that direction; the offset = 0 point corresponds to the position of the protostar. The relative velocity tends to be higher in regions closer to the central protostar within ±5'' (∼1200 au), indicating an infalling motion of the envelope gas (for details, see Section 4.3).

Zoom In Zoom Out Reset image size

The black dashed lines in Figure 4(a) show that the ¹³CO emission exhibits arc-like structures. The blueshifted (∼2 km s⁻¹) and redshifted (∼4 km s⁻¹) arcs extend to the northwest and northeast toward the continuum source and connect to the infalling envelope. Both have a length of ∼1000 au. To further characterize the structures, we estimated their density and mass. Assuming local thermodynamic equilibrium (T = 20 K), we can determine the ¹³CO column density using the following equation (Wilson et al. 2013):

$$\begin{eqnarray}&&{N}_{{}^{13}\mathrm{CO}}\approx 1.5\times {10}^{14}\displaystyle \frac{T\,{e}^{5.3/T}\,\int {\tau }_{\nu }\,{\rm{d}}v}{1-{e}^{-10.6/T}}\,{\mathrm{cm}}^{-2}\,.\end{eqnarray} \tag{ 1 }$$

We obtain ${N}_{{}^{13}\mathrm{CO}}=1.6\times {10}^{15}\,{\mathrm{cm}}^{-2}$. Assuming that the abundance ratio of ¹²CO to ¹³CO is 70 and the ratio of H₂ to ¹²CO is 10⁴ (Frerking et al. 1982), we get a H₂ column density of ${N}_{{{\rm{H}}}_{2}}=1.1\times {10}^{21}\,{\mathrm{cm}}^{-2}$, and arc mass of ${M}_{{{\rm{H}}}_{2},\ \mathrm{arc}}\,=5.7\times {10}^{-4}\,{M}_{\odot }$. We describe the interpretation of the arc-like structures in Section 4.3.

Figure 4(c) is a zoomed-in view of the vicinity of the continuum source, as indicated by the black rectangle in panel (a). As indicated by the magenta dashed line, the ¹³CO emissions of velocities close to the systemic velocity have an S-shaped configuration. For the continuum source, the velocity gradient extends from northwest to southeast, and its extent is about 500 au. The direction of the velocity gradient corresponds to the major axis of the disk and is also perpendicular to the outflow axis and to the large-scale velocity gradient seen in panel (a). Figure 4(d) shows a PV diagram along the arrow shown in panel (c). There are emissions with relative velocities of ≳3 km s⁻¹ in the region within ±2'' (corresponding to ∼500 au) of the continuum source. The region with a negative offset (i.e., northwest of the disk) is blueshifted, and the opposite side is redshifted. The small-scale velocity gradient perpendicular to the outflow could trace the rotating motion of the disk or the envelope (see also Section 4.3).

The velocity structure provides information on the protostellar mass, although the angular resolution is not sufficiently high to determine it properly. At an offset of ± 1'' (∼240 au), the relative velocity is ∼2 km s⁻¹ (see Figure 4). To produce such a high velocity at large radii by Keplerian rotation requires at least a 1 M_⊙ source with a mass heavier than typical low-mass protostars. This estimation gives a lower limit of the protostellar mass without considering the inclination angle, which is about a factor of 2 heavier in the case of close to face-on viewing. The dynamically inferred mass is consistent with the mass estimated by Spezzi et al. (2008) based on the position of DK Cha in the H-R diagram.

Numerous observational studies have reported outflows, most of which are nicely shaped bipolar flows when viewed nearly edge-on. It is difficult to spatially resolve the radial velocity distribution within an outflow in such objects because the observed outflow velocity decreases as the line-of-sight direction approaches edge-on. In addition, it is also difficult to investigate the internal velocity structure of an outflow when viewed close to edge-on. We observed the DK Cha outflow from nearly pole-on and could clearly detect its spatial velocity structure, as shown in the Figures 1–3. In this section, we discuss the three-dimensional velocity structure of the DK Cha outflow based on the observed velocity gradient in the plane of the sky.

The spatial distribution of the ¹²CO emission seen in Figures 1–3 is highly complicated. Thus, it is difficult to guess the origin of the emission from the morphology alone. However, as mentioned in Section 3.1, their typical velocities are much faster than the speed of sound at 10 K (∼0.2 km s⁻¹), and their complex structures can only be attributed to outflow origin. Since it is significantly difficult to reproduce blueshifted and redshifted emissions surrounding the continuum source in a typical edge-on view of the outflow, we interpreted the CO emission as a pole-on view of the outflow. Note that previous studies already suggested that the DK Cha outflow is observed nearly pole-on (van Kempen et al. 2009; Yıldız et al. 2015).

4.1.1. Comparison with Another Pole-on Outflow

4.1.2. Three-dimensional Outflow Structure

In the following, we discuss which model could reasonably explain the spatial velocity structure of the outflow observed in the CO emission. As described in Section 3.1, the redshifted component is likely to interact with the surrounding gas, resulting in a complicated structure. On the other hand, the blueshifted component has a relatively simple velocity structure. Thus, we focused only on the velocity structure of the blueshifted outflow (see the left panel of Figure 3).

A simple model proposed by Lee et al. (2000) is often used to interpret the outflow velocity structure emerging on the PV diagram. Their model describes the outflow as a parabolic surface with a velocity structure that accelerates along the outflow axis. The model appears to explain well the outflow velocity structure on the PV diagram along the outflow axis, while it is not obvious that the model can explain the spatial velocity structure in the plane of the sky. We, therefore, present a simplified three-dimensional illustration that provides an overview of the velocity structure in the left panels in Figure 5. We assume continuous acceleration, while for simplicity we represent the outflow as being composed of three different velocities (orange for the low-, green for the intermediate-, and blue for the high-velocity component). The lower left panel of Figure 5 clearly shows that the high-velocity component has a more extended structure than that of the low-velocity component in the plane of the sky. Figures 2 and 3 show that the low-velocity component surrounds the high-velocity component, which is the opposite to what is observed for the model reported by Lee et al. (2000). Thus, it is difficult to reproduce our results with the vertical acceleration model. Note that we do not mean that outflows are not actually accelerating along the outflow axis.

Zoom In Zoom Out Reset image size

Another possible approach is to examine the horizontal velocity structure of the outflow (or the velocity distribution in the plane perpendicular to the outflow axis, not along the outflow surface). We consider the horizontal velocity gradient model, in which the gas in the inner regions flows outward at higher velocity (Figure 5 right panels). If we observe such an outflow from a direction close to pole-on, a layered velocity structure should appear, as shown in the lower right panel of Figure 5. In addition, the high-velocity component is expected to be more compact than the low-velocity one. The velocity structure of the observed outflow in the left panel of Figure 3 has a similar layered velocity gradient, with a compact high-velocity component surrounded by an extended low-velocity component. In the panel, the missing part of the ellipse could be due to interactions with the ambient gas, lack of sensitivity outside the observation area, or other factors. Therefore, our observations reveal that outflows with different velocities are driven from different radii. Such a velocity structure can be naturally reproduced by an outflow directly driven by the magnetocentrifugal mechanism (e.g., Machida 2014; Machida & Basu 2019).

The outflow driving mechanism is still under debate. There are two main scenarios. In one, the outflow is produced by ambient gas receiving momentum from the high-velocity jet (e.g., Arce et al. 2007). In the other, the outflow is driven directly from the circumstellar disk by the magnetocentrifugal mechanism irrespective of the appearance of the jet (e.g., Machida 2014). Recent ALMA observations have supported the latter. For example, Bjerkeli et al. (2016) reported that the TMC-1A outflow is driven from the outer edge of the disk. A clear signature of rotation could be detected in the bipolar outflow of Orion Source I (Hirota et al. 2017). The large angular momentum of the outflow indicates that the outflow is driven far from the central protostar. Note that the outflow driven from the region close to the central star cannot have a large angular momentum (or rapid rotation). Matsushita et al. (2019) discovered an extremely high-velocity flow and outflow around MMS 5 in the Orion Molecular Cloud 3, and found that their axes are misaligned, whereas if the low-velocity flow is entrained by the high-velocity flow, their axes should be aligned.

Here, based on the directly driven outflow scenario, we estimate the driving radius from the observed outflow velocities shown in Figures 1–3. Classical theoretical studies (e.g., Mestel 1961, 1968) have shown that the terminal velocity of the outflow (or disk wind) roughly corresponds to the Keplerian velocity at its launching radius. Thus, we simply assume that the observed outflow velocity v_out agrees with the Keplerian velocity v_K at each launching radius r_0,

$$\begin{eqnarray}&&{v}_{\mathrm{out}}\sim {v}_{{\rm{K}}}=\sqrt{\displaystyle \frac{{{GM}}_{\mathrm{star}}}{{r}_{0}}}.\end{eqnarray} \tag{ 2 }$$

With Equation (2), the outflow launching radius r₀ can be described as

$$\begin{eqnarray}&&{r}_{0}\sim {{GM}}_{\mathrm{star}}\,{v}_{\mathrm{out}}^{-2}=0.7\,\mathrm{au}\left(\displaystyle \frac{{M}_{\mathrm{star}}}{2{M}_{\odot }}\right){\left(\displaystyle \frac{{v}_{\mathrm{out}}}{50\,\mathrm{km}\,{{\rm{s}}}^{-1}}\right)}^{-2},\end{eqnarray} \tag{ 3 }$$

where we have used a protostellar mass M_star of 2M_⊙ (Spezzi et al. 2008). As shown in Figures 1–3, the outflow of DK Cha has crescent-shaped structures across the low-velocity (<5 km s⁻¹) to the high-velocity (>50 km s⁻¹) components. The outflow launching region extends over 2 orders of magnitude (∼0.7–70 au) if these outflow components with different velocities are driven directly from their respective radii. Note that the driving radius depends somewhat on the Alfvén radius (i.e., the intensity of the magnetic field and ram pressure). Thus, there is uncertainty in estimating the launching radius within a factor of ≲10 (e.g., Pudritz & Norman 1986; Koenigl & Ruden 1993; Kudoh & Shibata 1997a; Pudritz & Ray 2019). Therefore, the inner (∼0.7 au) and outer (∼70 au) launching radii can be changed according to the extent of the Alfvén radius. However, the large extent of the outflow driving region (over 2 orders of magnitude in radius) would not change significantly, because the outflow velocity at each shell (see Figure 5 right panels) is primarily determined by the Keplerian velocity at each launching radius (Kudoh & Shibata 1997a, 1997b; Machida et al. 2008), and the difference in the outflow velocity (∼5–50 km s⁻¹) gives the difference in the launching radius proportional to the outflow velocity (see Equation (3)).

Considering the conservation of the (specific) angular momentum ($j={r}_{\mathrm{out}}{v}_{\mathrm{rot}}=\sqrt{{{GM}}_{\mathrm{star}}{r}_{0}}$), it is valuable to estimate the rotational velocity of the outflow v_rot,

$$\begin{eqnarray}&&{v}_{\mathrm{rot}}=0.4\,\mathrm{km}\,{{\rm{s}}}^{-1}{\left(\displaystyle \frac{{r}_{\mathrm{out}}}{1000\,\mathrm{au}}\right)}^{-1}{\left(\displaystyle \frac{{M}_{\mathrm{star}}}{2{M}_{\odot }}\right)}^{0.5}{\left(\displaystyle \frac{{r}_{0}}{70\,\mathrm{au}}\right)}^{0.5},\end{eqnarray} \tag{ 4 }$$

where r_out is the observed outflow radius (the half width of the outflow lobe). Thus, the rotational velocity of the outflow is clearly less than the velocity resolution in this observation. Since the outflow rotational velocity becomes smaller due to the inclination angle, it is natural that we could not detect any sign of outflow rotation.

In this subsection, we discuss the gas motion expected from the ¹³CO(J = 2–1) emission. As described in Section 3.2, two velocity gradients around DK Cha are detected in the ¹³CO emission (Figure 4), one perpendicular and one parallel to the outflow axis. Relative velocities seem to be higher in regions closer to the central protostar. This velocity structure is similar to the infalling and rotating gas motion reported in recent ALMA observations (e.g., Aso et al. 2015; Sakai et al. 2016; Yen et al. 2017). In particular, the velocity structure (moment 1 map) of HL Tau (Yen et al. 2017, Figure 1) is very similar to our results (Figure 4(a)) in that both sources have a 1000 au scale envelope gas, velocity gradients parallel and perpendicular to the outflow axis, and an S-shaped structure with the gas having velocities close to the systemic velocity in the ¹³CO(J = 2–1) emission. Thus, it seems that the ¹³CO emission has traced the rotating infalling envelope.

We can also confirm the existence of ¹³CO arc-like structures in Figure 4(a). There is no sign of a binary star in DK Cha, and thus the structure is unlikely to be the result of gravitational interactions in a multiple system (Matsumoto et al. 2015) like MC27/L1521F (Tokuda et al. 2014, 2016). Rather, it is more natural to interpret these structures as nonaxisymmetric accretion flows toward the central protostellar system since these are connected to the rotating infalling envelope described above. In other words, we expect that the structures are the same phenomenon as the streamers recently observed (Pineda et al. 2022, and references therein). Among Class I objects, streamers have been detected in HL Tau and Per-emb-50 with lengths of 520 and 3000 au, respectively (Yen et al. 2019; Valdivia-Mena et al. 2022). The streamers detected in DK Cha have a similar size (∼1000 au) to these objects. Our observations may be a case in which the rotating infalling envelope and the streamers were detected simultaneously. HL Tau, Per-emb-50, and DK Cha are all intermediate-mass protostars (∼2M_⊙). Therefore, it is interesting to investigate whether streamers are associated with a broader range of protostellar mass. In addition, since shock tracers such as SO have been detected together with streamers (Garufi et al. 2022), it will be important to observe different molecular emission lines for DK Cha.