A 100 au Wide Bipolar Rotating Shell Emanating from the HH 212 Protostellar Disk: A Disk Wind?

Sulfur is severely depleted in cold molecular clouds, with an abundance being three orders of magnitude less than the cosmic abundance (e.g., Tieftrunk et al. 1994). On the other hand, S-bearing molecules are very abundant in almost all the components associated with the low-mass star-forming process, from dark clouds to protostellar envelopes and slow outflows, as well as hot corinos (van Dishoeck & Blake 1998). This is due to the release of sulfur from the dust mantles that can be caused by either evaporation when the dust is heated to temperatures higher than ∼100 K (e.g., in the hot corinos or at the disk surface), or their erosion in shocks due to gas–grain (sputtering) or grain–grain collisions (shattering), see, e.g., Pineau des Forêts et al. (1993) and Guillet et al. (2011, and references therein). In this context, S-bearing species, and, in particular, those more abundant such as SO and SO₂ can be considered among the best tracers to image the high-density gas components involved in the process leading to the formation of a Sun-like star, namely, the high-velocity collimated jet (Lee et al. 2010; Podio et al. 2015), the accretion shock occurring at the envelope-disk interface (Sakai et al. 2014a, 2014b; Lee et al. 2016; Sakai et al. 2017), the outflow cavity walls, and possibly the disk surface directly illuminated and heated by the central protostar (this paper). The relative abundances of S-bearing molecules depend on the main S-compounds released from mantles, i.e., if S is released in an atomic and/or molecular form such as H₂S or OCS, a still hotly debated open question (Wakelam et al. 2004; Codella et al. 2005; Podio et al. 2014). In any case, SO₂ is expected to mainly form due to the reaction between SO and OH (e.g., Pineau des Forêts et al. 1993). Indeed, the SO₂/SO abundance ratio has been proposed to be an efficient chemical clock to date the gas (Charnley 1997; Hatchell et al. 1998). Unfortunately, Wakelam et al. (2004) and Codella et al. (2005) have shown that this abundance ratio is not very useful, because the SO and SO₂ abundances depend on the initial S-carrier on the grain mantles. In addition, both species may not be tracing the same portion of the gas.

Recent observations of SO in the inner 50 au around the L1527 protostar, located in Taurus, confirmed high SO abundances (up to 10⁻⁷) associated with the dense (≥10^5–6 cm⁻³) gas revealing the accretion shock at the envelope-disk interface (Sakai et al. 2014a, 2014b, 2016, 2017). On the other hand, before the present work, Podio et al. (2015) traced SO high-velocity emission associated with the HH 212, with very high abundances (up to 10⁻⁶). In addition, the authors observed outflow-velocity emission close to the protostar, showing small-scale velocity gradients, indicating that it originates partly from the rotating disk, and partly from the base of the jet, opening the way to investigations at higher spatial scale. Also the SO abundances in the HH 212 jet have been estimated to be very high, up to 10⁻⁶, confirming what was suggested by Lee et al. (2010) for the HH 211 protostellar jet. Also the SO₂ abundance has been measured to be ∼10⁻⁶. On the other hand, Podio et al. (2015) found in the HH 212 disk X_SO = 10⁻⁸–10⁻⁷, i.e., values higher by 3–4 orders of magnitude than those derived in protoplanetary disks (Fuente et al. 2010; Dutrey et al. 2011). These high abundances are plausibly due to either violent shocks caused by jets traveling, or slow shocks occurring close to the protostar, such as the accretion shocks already observed in HH 212 by Lee et al. (2017a).

A compact outflow has recently been detected in SO at ∼013 resolution, extending out from the central source along the jet axis (Tabone et al. 2017). Now at ∼004 resolution, we can resolve the outflow and check for a disk wind component. As shown in the total integrated emission map in Figure 1(b), the SO emission is detected here within ∼±05 of the central source around the SiO jet, extending out from the inner disk. The emission has a velocity with $| {V}_{\mathrm{off}}| \lesssim 10$ km s⁻¹. SO emission is also detected in the disk atmosphere, overlapping with the CH₂DOH emission. No SO emission is detected in the dusty disk, likely because the continuum emission is bright and optically thick there (Lee et al. 2017c); thus, the emission behind it is blocked and the emission in front of it appears absorbed against the bright background.

Figure 3 shows the SO maps at low and high velocities. At low velocity, the outflow can be decomposed into two components, a wide-opening outflow extending out to ∼02 to the north and south from the inner disk, and a collimated jet along the jet axis (Figure 3(a)). The wide-opening outflow is reasonably resolved in the north, appearing as a shell structure opening out from the central source. The shell structure of the outflow can be better seen in the position–velocity (PV) diagrams as discussed below in Section 3.3. In the shell outside the atmosphere, the blueshifted emission overlaps with the redshifted emission (Figure 3(b)), indicating that the shell there is mainly radially expanding. At high velocity, the SO emission is seen around and along the jet axis (Figure 3(c)). As discussed later, since the projected velocity of the expanding shell can reach up to V_off of ±5 km s⁻¹, part of the emission traces the expanding shell projected on the jet axis. The emission peaks that spatially coincide with knots N3, N1, S1/S2, S3, and S4 should mainly trace the jet itself. Like the SiO jet, the blueshifted emission is seen more to the north and the redshifted emission is seen more to the south.

Zoom In Zoom Out Reset image size

A compact outflow has also been detected recently in SO₂ at ∼013 resolution (Tabone et al. 2017). Thus, we can also check for a disk wind component in SO₂. Here 12 SO₂ lines are detected (see Table 4) with $| {V}_{\mathrm{off}}| \,\lesssim$ 8 km s⁻¹. They are weak and thus stacked to produce a mean SO₂ line with a better signal-to-noise ratio. As shown in the total integrated emission map in Figure 1(c), the SO₂ emission is detected within ∼±03 of the central source, not only tracing the outflow but also tracing the inner disk atmosphere. No SO₂ emission is detected toward the dusty disk because of the same reason as discussed above.

Figure 4 shows the SO₂ maps at low and high velocities. At low velocity, unlike the SO emission, the SO₂ emission traces mainly the inner disk atmosphere, as seen in Figures 4(a) and (b). The disk atmosphere is rotating with the redshifted emission in the east and blueshifted emission in the west, the same rotation sense as that found before in CH₂DOH, as indicated by the blue and red arrows in Figure 4(b). Like that seen in SO, the outflow in the north can also be decomposed into a wide-opening outflow and a collimated jet. In the north, the shell structure of the wide-opening outflow can be seen in the east. In the south, the shell structure can only be barely seen in the east, because it is contaminated by the emission around knot S3. At high velocity, like the SO emission, the SO₂ emission is mainly seen around the jet axis, with the blueshifted emission to the north and redshifted emission to the south (Figure 4(c)).

Zoom In Zoom Out Reset image size

Figure 5 shows the position–velocity (PV) diagrams along the jet axis in SO and SO₂ in order to study the expansion velocity of the wide-opening outflow. The PV diagram of the SO₂ emission shows roughly a parabolic structure opening up from the source position to ∼02 to the north and south to V_off ∼ ±5 km s⁻¹, as guided by the white curves. The PV diagram of the SO emission is complicated. Part of the SO emission follows the parabolic structure seen in SO₂ and part of the SO emission is associated with the knots in the jet. The parabolic PV structure is associated with the wide-opening outflow, supporting that the wide-opening outflow is a shell and that the expansion velocity of the shell increases with the distance.

Zoom In Zoom Out Reset image size

Figure 6 shows the intensity weighted velocity (first moment) maps of the SO and SO₂ emission at low velocities in order to search for the rotation in the shell. In SO, the eastern and western outer edges (as roughly delineated by the white curves) of the shell mostly have opposite velocities (Figure 6(a)), indicating that the shell is also rotating around the jet axis, in addition to the expansion mentioned above. The rotation of the shell has the same sense as that of the disk, as indicated by the red and blue arrows. On the other hand, as discussed earlier, the SO₂ emission mainly arises from the rotating disk atmosphere (Figure 6(b)). The outflow edges are too faint in SO₂ to appear in this figure.

Zoom In Zoom Out Reset image size

3.3.1. Rotation and Expansion in the Shell

Figure 7 shows the PV diagrams cut across the jet axis at increasing distance z from the disk midplane with a step of 0025 (or 10 au) to the north and south, in order to study the expansion and rotation in the shell in detail. Since the kinematics in the northern part is better resolved than that in the southern part, our kinematic study here is based mainly on the northern part.

Zoom In Zoom Out Reset image size

Near the bottom of the northern atmosphere at z = 005, the PV diagram shows the redshifted emission in the east and blueshifted emission in the west (Figure 7(a)), as expected for a rotating disk atmosphere. In addition, the SO emission shows a velocity increasing toward the center, with the outer boundary of its PV structure well described by the Keplerian rotation velocity profile (magenta curves) due to the central protostar (Lee et al. 2017c). This confirms the presence of a rotationally supported disk orbiting the protostar. The mass of the protostar is ∼0.25 M_⊙ from the kinematics of the disk and envelope in molecular lines (Lee et al. 2017c). At low velocities where the emission traces the outer part of the atmosphere, the PV structure can be roughly described by a tilted elliptical structure as guided by the green ellipse. This outer part is also seen in SO₂ emission. This PV structure confirms that the outer part of the atmosphere is not only rotating but also expanding (Hirota et al. 2017), indicating that the atmosphere there has become a part of the outflow shell. The upper part of the ellipse is much fainter than the lower part, probably because the emission in the nearside is self-absorbed. As we go higher up in the atmosphere to z = 0075, the elliptical PV structure becomes broader in the minor axis (Figure 7(b)), indicating that the expansion motion becomes faster higher up, consistent with the discussion above.

In the shell slightly above the disk atmosphere at z = 01, the PV diagrams also show a tilted elliptical PV structure (green ellipse, Figure 7(c)), indicating that the shell there is also rotating and expanding. As compared to that seen in the atmosphere, it is broader in the minor axis, indicating that the expansion motion is faster. As we go further to the north to z ≥ 0125, the PV structure can be decomposed into two components, one with an elliptical structure for the shell, and the other along the jet axis tracing the emission in the jet (Figures 7(d)–(g)). The major axis of the ellipse is almost aligned with the x-axis, indicating that the gas motion in the shell is dominated by the expansion. In addition, the SO₂ emission of the shell becomes very faint at low velocities, thus the limb-brightened shell (outflow edge) becomes faint and not detected, as seen in Figure 4(a). As a result, the emission of the shell is mainly detected at high velocities, appearing mainly around the jet axis, as seen in Figure 4(c).

In the southern part, similar elliptical PV structures can also be roughly seen for the shell in the atmosphere (Figures 7(h) and (i)). However, no clear elliptical PV structures can be identified easily (Figures 7(j)–(n)) in the PV diagrams above the disk atmosphere. Therefore, we will not attempt to fit the PV structures here because the results would be quite uncertain.

The physical properties (e.g., radius, rotation velocity, expansion velocity, and specific angular momentum) of the shell in the north, where the shell is detected at a higher signal-to-noise ratio and better resolved, can be estimated from the elliptical PV structures seen in the PV diagrams cut across the jet axis in Figures 7(a)–(g). As can be seen, the elliptical PV structures can be roughly described by the green ellipses in the figures. Since the elliptical PV structures are not well defined, we only used an eye-fitting to roughly obtain the green ellipses, which are required to pass through the emission in the shells as much as possible. From these green ellipses, the radius of the shell is given by the maximum position offset from the jet axis, the rotation velocity is given by the velocity at the maximum position offset with respect to the velocity centroid of the ellipse, and the expansion velocity is given by the velocity at the zero position offset with respect to the velocity centroid of the ellipse. The specific angular momentum is then derived by multiplying the rotation velocity with the radius. These physical quantities of the shell are shown in Figure 8, with their uncertainties described in the caption. Since the inclination angle of the outflow is small, no correction is needed for the inclination effect. As expected, the shell becomes wider with the distance (Figure 8(a)). Interestingly, the rotation velocity is roughly the same as the expansion velocity in the shell inside the atmosphere, but becomes much smaller than the expansion velocity in the shell outside the atmosphere (Figure 8(b)). The specific angular momentum in the shell is roughly constant at ∼40 au km s⁻¹ within ∼50 au of the central source, and then decreases to ∼25 au km s⁻¹ at ∼80 au (Figure 8(c)). This decrease in the specific angular momentum is uncertain because the rotation velocity cannot be estimated accurately outside the atmosphere due to the poorly identified elliptical PV structure.

Zoom In Zoom Out Reset image size

With 12 SO₂ lines detected, we can construct a population diagram (Goldsmith & Langer 1999) to estimate the rotational temperature and the column density of the molecule in the disk atmosphere, the wide-angle outflow shell, and the collimated jet. It is a diagram that plots the column density per statistical weight in the upper energy state in the optically thin limit, ${N}_{u}^{\mathrm{thin}}/{g}_{u}$, versus the upper energy level E_u of the lines. Here ${N}_{u}^{\mathrm{thin}}=(8\pi k{\nu }^{2}/{{hc}}^{3}{A}_{{ul}})W$, where the integrated line intensity $W=\int {T}_{B}{dv}$ with T_B being the mean brightness temperature.

The integrated line intensity of each transition can be measured for the SO₂ emission in the south, where the emission is detected at a higher signal-to-noise ratio. The resulting population diagram is shown in Figure 9. Fitting the data points with E_u > 70 K, we estimate a rotational temperature of 81 ± 20 K and a column density of 8.2 ± 2.0 × 10¹⁵ cm⁻². Notice that the data points with E_u < 70 K are not used for the fitting because their emissions are only barely detected, with their intensity slightly over 3σ. These estimated values are mainly for the disk atmosphere and the outflow shell, because SO₂ emission is mainly detected there.

Zoom In Zoom Out Reset image size

The SO emission traces mainly the outflow shell and the collimated jet. The excitation temperature is unknown. Along the jet axis, the emission has a peak brightness temperature of ∼150 K and a mean integrated line intensity of 700 K km s⁻¹. Thus, we assume a temperature of ∼200–300 K in the collimated jet, which is consistent with the temperature range discussed in Podio et al. (2015). The column density of SO there is then estimated to be ∼(1.5–1.8) × 10¹⁶ cm⁻². In the shell, the mean integrated line intensity is ∼400 K km s⁻¹. The excitation temperature can be assumed to be ∼100 K, which is an approximation to that derived above in SO₂ for the disk atmosphere and the outflow shell. Thus, the SO column density is estimated to be ∼6 × 10¹⁵ cm⁻² in the outflow shell.

The abundances of SO and SO₂ have been estimated before at lower resolution by comparing their column density to that derived from the CO J = 3−2 emission (see Table 3 in Podio et al. 2015) and thus can be used here to derive the mean H₂ column density. With a SO abundance of (1–10) × 10⁻⁷, the mean H₂ column density is estimated to be (6–60) × 10²¹ cm⁻² in the SO shell. With a SO₂ abundance of (5–12) × 10⁻⁷, the mean H₂ column density is estimated to be (5–12) × 10²¹ cm⁻² in the SO₂ shell. Given that the thickness of the shell is ≲50 au, the mean H₂ density is ≳(4–40) × 10⁶ cm⁻³ in the SO shell, and (3–8) × 10⁶ cm⁻³ in the SO₂ shell. As can be seen, the SO shell could have a higher mean H₂ density than that in the SO₂ shell.

In star formation, one important topic is to search for a low-velocity disk wind that can carry away the disk angular momentum, allowing the disk material to accrete. Previously, a rotating outflow was detected in CB 26 in CO J = 2−1 emission and was argued to trace a disk wind launched from a radius of ∼30 au in the disk (Launhardt et al. 2009). Later, a rotating outflow was also detected in Orion BN/KL Source I in SiO maser, appearing to be a disk wind launched from a radius of ∼20 au in the disk (Matthews et al. 2010). In recent follow-up observations of the same object, a rotating outflow was also detected further out in Si¹⁸O, and argued to trace a disk wind launched from a radius of >10 au (Hirota et al. 2017). Recently, another rotating outflow was detected in TMC1A in CO J = 2–1, and also argued to trace a disk wind launched from a radius of >5 au in the disk (Bjerkeli et al. 2016). Another example is HH 30, where a rotating CO (J = 2–1) shell is observed to surround one side of the bipolar optical jet (F. Louvet et al. 2018, in preparation).

In HH 212, we can search for a disk wind emerging from the well-resolved disk. As discussed later, the wind's source can be on the disk 2–3 au from the protostar. At this radius, the density and temperature on the disk surface (at one scale height) are estimated to be ∼(2–4) × 10¹² cm⁻³ and 770–1050 K, respectively, based on the disk model in Lee et al. (2017b). Thus, our SO and SO₂ lines, with a critical density of ∼10⁷ cm⁻³, are better tracers of a disk wind than the CO J = 2–1 line, which has a critical density of ∼10⁴. However, note that, since only charged particles are directly coupled to the magnetic field, neither CO and SO/SO2 can be directly accelerated by a magnetic field; they must be coupled to magnetically accelerated charged particles through collisions.

Recently, a compact outflow has been detected in HH 212 at ∼52 au (013) resolution, extending out from the disk (Tabone et al. 2017). It is rotating and thus may trace a disk wind. Now at ∼16 au (004) resolution, the compact outflow is found to consist of two distinct components: a wide-angle outflow shell and a collimated jet. The collimated jet is aligned with the SiO jet and thus likely traces the shock interactions along the jet axis. Interestingly, the outflow shell lies within the large-scale molecular outflow detected in C³⁴S (Codella et al. 2014) and HCO⁺ (Lee et al. 2017c), and is thus not the rotating envelope material pushed away by the underlying wind or jet (see Figure 10 for various components). Moreover, it is rotating and seen extending out from the inner disk at a radius of <20 au within the centrifugal barrier (Lee et al. 2017c), and thus may trace a disk wind coming out from there. The specific angular momentum is roughly constant at ∼40 au km s⁻¹ within ∼50 au of the central source. This specific angular momentum corresponds to that at ∼7.2 au in the disk, where the rotation velocity is ∼5.5 km s⁻¹, assuming a central mass of 0.25 M_⊙ (Lee et al. 2017c). It decreases to ∼25 au km s⁻¹ at a distance of ∼80 au. This decrease in specific angular momentum could be inconsistent with that expected for a disk wind; however, further observations are needed to confirm this. On the other hand, it is also possible that the SO/SO₂ outflow shell is the disk/atmosphere material pushed up and out by a wind launched at the innermost disk, e.g., by a wide-angle radial wind component surrounding the fast jet as in the X-wind model.

Zoom In Zoom Out Reset image size

The shell kinematics could be used to differentiate the above two origins of the outflow shell. If the shell traces the disk wind itself, the velocity vectors are expected to be parallel to the shell structure. If the shell traces the disk or atmosphere material pushed away by a wide-angle radial wind, then the velocity vectors in the shell are expected to be radially directed. For simplicity, the shell here can be assumed to roughly have a parabolic structure with z = CR²+002 in the cylindrical coordinate system and C ∼ 14 arcsec⁻¹ (as indicated by the white curves in Figure 6). If the shell material is moving along the shell, then we have v_R = v₀R and v_z = 2v₀z in the cylindrical coordinate system. If the shell material is radially expanding, then we have v_R = v₀R and v_z = v₀z (Lee et al. 2000). Figure 11 shows the model PV structures of these two models at an inclination angle of ∼4°, in comparison to the observed PV structures. The solid curves are from the radially expanding shell model with v₀ ∼ 42 km s⁻¹ arcsec⁻¹ and the dotted curves are from the material moving along the shell model with v₀ ∼ 45 km s⁻¹ arcsec⁻¹. The velocity vectors and magnitudes in these two models are shown in Figure 10, with the black arrows for the former and gray arrows for the latter. As can be seen, both model PV structures can match the observed PV structure reasonably well. Therefore, unfortunately, we cannot determine which model is better based on the kinematics here in HH 212. Having said that, it is unclear whether a disk wind can have an expansion speed that increases with the distance from the disk and at the same time a specific angular momentum that decreases with distance. For a shell accelerated by an underlying wide-angle radial wind, it is reasonable to have an expansion velocity increasing with the distance (Shu et al. 1991; Lee et al. 2000), because the disk atmosphere has a density decreasing with height. Whether the specific angular momentum can naturally decrease with distance in this scenario is uncertain. In addition, the shell in the north is seen connecting to knot N4 further north, which is likely also affected by it (see Figure 3).

Zoom In Zoom Out Reset image size

Another question we can ask is the following: does the wide-angle wind component of an X-wind, if it exists, have enough momentum to drive the outflow shell? The mass of the shell can be roughly estimated from the mean H₂ column density derived from the SO emission in Section 3.4. Thus, the total mass of the shell within ∼02 of the central source in the north and south is ${M}_{s}=1.4{m}_{{{\rm{H}}}_{2}}{N}_{{{\rm{H}}}_{2}}A\,\sim$ (0.4–4) × 10⁻⁴ M_⊙, where A is the area of the shell. Assuming a mean expansion velocity of ∼5 km s⁻¹ in the shell, the momentum is then (0.2–2) × 10⁻³ M_⊙ km s⁻¹. In the X-wind model, the wide-angle wind component could have a mass-loss rate and a velocity similar to those of the jet component (Shu et al. 1995). The jet component was found to have a mass-loss rate of ∼10⁻⁶ M_⊙ yr⁻¹ and a velocity of ∼100 km s⁻¹ (Lee et al. 2015). According to our simple radially expanding model for the shell, the shell has a dynamical age of 1/v₀ or ∼45 years, indicating that the shell could have been driven for that long by an underlying wind. Thus, the wide-angle wind component can, in principle, provide an accumulated momentum of ∼4.5 × 10⁻³ M_⊙ km s⁻¹, enough to drive the shell.

One more question to ask is the following: why is there only one shell? In the case of a radial expanding shell, the shell has a dynamical age much younger than the age of HH 212. Since the jet is episodic, the inner wind that drives the shell should also be episodic. Since the semi-periodic spacing between two bow shocks in the jet is ∼3'' (Lee et al. 2007), the period for the major ejection of the jet/wind is ∼50 years and thus slightly longer than the dynamical age of the shell. Therefore, it is possible that the shell was produced recently and the earlier shells are lost in the outflow cavity probably because the density becomes too low to excite the SO and SO₂ lines. Or as suggested in Tabone et al. (2017), the shell traces an onion-like shell in the disk wind coming out from the disk around the SiO jet (see their Figure 3), but the part further away from the central source is lost in the cavity.

Independent of the exact mechanism for producing the outflow shell, if it is really rotating with a specific angular momentum of ∼40 au km s⁻¹, it is likely coming from the disk at a relatively large radius of at least several astronomical units. If the outflow is indeed an MHD disk wind from the disk, then using Equation (4) in Anderson et al. (2003), and an outflow velocity of 5–10 km s⁻¹, the launching radius is estimated to be 2–3 au. The specific angular momentum is much larger than the mean value for the SiO jet (∼10 au km s⁻¹), indicating that the outflow shell is not simply the jet angular momentum and is much larger than the mean value for the SiO jet (∼material that is swept into a shell). One may argue that the shell material may come from the infalling protostellar envelope that extends all the way to less than several astronomical units above and below the disk. However, the large-scale outflow cavity wall, which presumably marks the inner boundary of the envelope, is located well outside of the outflow shell (as indicated in the cartoon of Figure 10). Furthermore, in the case of a much more evolved T Tauri star HH 30, there is little evidence for a large-scale envelope, and yet a rotating, expanding shell is still observed surrounding the jet (F. Louvet et al. 2018, in preparation), indicating that an envelope is not needed for producing such a structure. This leaves the material coming directly from the disk as the most likely material for the outflow shell.

The key question is whether that disk material in the outflow shell is launched from the several astronomical unit radius by some mechanism internal to the local disk, such as through magnetocentrifugal acceleration or magnetic pressure gradient, or by some external agents, such as interaction with a wide-angle wind surrounding the SiO jet. Indeed, the two mechanisms are not mutually exclusive: outflow launched from the several astronomical unit radius is expected to interact with the wide-angle wind interior to it, if such a wind exists. In any case, the rotating outflow shell carries away angular momentum from the disk, most likely from the several astronomical unit radius, which is in contrast with the rotating SiO jet driven from the ∼0.05 au radius.

Previous observations of HH 212 at an angular resolution of ∼05 have shown that the SO and SO₂ abundances in the outflow are very high, up to ∼10⁻⁶ (Podio et al. 2015). Now at higher resolution, the outflow is resolved into a small-scale wide-opening outflow shell and a collimated jet. The high SO and SO₂ abundances in the collimated jet could be due to violent (internal) shocks in the jet. The origin of high abundances in the wide-angle outflow shell is uncertain. If the shell traces a disk wind launched from the disk, then the high abundances could be due to the thermal sublimation of icy mantles near the source and the gradual heating by ambipolar diffusion during MHD acceleration (Panoglou et al. 2012). If the shell is the disk/atmosphere material pushed away by an inner wind, then the high abundances could be due to shock interactions with the inner wind.

As for the disk, Podio et al. (2015) have found a high SO abundance of 10⁻⁸–10⁻⁷. Our observations also show high SO and SO₂ column densities (see Section 3.4) and thus high SO and SO₂ abundances there in the disk atmosphere. These high abundances in the disk atmosphere could be due to the accretion shocks already observed in HH 212 by Lee et al. (2017a), radiation heating by the central protostar, or shock interactions with an inner wind.