Shaping the Envelope of the Asymptotic Giant Branch Star W43A with a Collimated Fast Jet

The CO(J = 2 → 1) emission spans a total velocity range of ∼200 km s⁻¹ and is distributed over an elongated region (deprojected size ∼300 × 3500 au) whose major axis has a PA ∼ 68° (see Figure 1 and Figure 4 in Appendix B). From the channel map shown in Figure 4 in Appendix B it can be seen that the emission with the highest velocity offset,¹⁰ $| {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \,\sim 100$ (see also Figure 2), is the most collimated and lies closest to the star. For velocity offsets $75\lesssim | {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \lesssim 100$, the emission exhibits a high degree of collimation and it reaches locations farther away from the star than for the highest velocity offsets, i.e., it exhibits a negative velocity gradient, $| {dv}| /| {dr}| \lt 0$. Furthermore, for velocity offsets $0\lesssim | {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \lesssim 75$ the emission is less collimated and the velocity gradient inverts, i.e., the emission reaches locations farther away from the star with increasing velocity offset.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Thus, the CO(J = 2 → 1) emission can be divided into two components characterized by their spatial distributions and velocity gradient. The component with negative velocity gradient has the highest velocity offsets $75\lesssim | {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \lesssim 100$ and it exhibits an extremely collimated morphology all the way from the vicinity of the central star (∼90 au) out to a deprojected distance of ∼1600 au. The negative velocity gradient may indicate a deceleration of the emitting material as it moves away from the central system. This component is shown with gray contours in Figure 1. The other component has lower velocity offsets $0\lesssim | {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \lesssim 75$ and is less collimated, albeit it has a bipolar morphology and is distributed around the high-velocity component, which is shown with green contours in Figure 1. The positive velocity gradient in this component may indicate entrainment of the circumstellar material by a fast jet. The criteria to distinguish these kinematical components are further explained in Section 3.3.

The continuum emission traces a bipolar shell (color map in Figure 1; see also Figure 3) that confines parts of the low-velocity CO(J = 2 → 1) emission. The bipolar shell has a typical deconvolved width of 130 au and it has two bright regions that exhibit point symmetry with respect to the central source. From Figure 1 it can be seen that the CO(J = 2 → 1) emission is absorbed toward these bright regions, suggesting that they are optically thick regions. Therefore, the brightness temperature of the bright regions (≈110 K) is likely to be the actual temperature of the dust. Figure 1 also reveals that the bright regions lie on the sides of the shell located closest to the main axis of the jet. This may be due to the passage of the jet close to the shell causing an increase in the density and/or temperature of the dust. The average density of the shell is estimated to be ${n}_{{{\rm{H}}}_{2},\mathrm{Shell}}\approx 5\,\times {10}^{8}$ cm⁻³ (see Appendix D for the calculations), which favors the physical conditions to produce the observed H₂O masers.

Zoom In Zoom Out Reset image size

There is also a bright emission source at the center of the nebula. A Gaussian fit to this source gives a deconvolved size of 90 × 60 au with a PA ≃ 145°, which is nearly perpendicular to the direction of the jet.

Figure 2 shows a position–velocity (PV) diagram of the CO(J = 2 → 1) emission obtained with a slit along the major axis of the emitting region, PA = 68°. It is evident in this PV diagram that the spatial spread of the CO(J = 2 → 1) emission, delimited horizontally by the solid red and light-blue lines, increases as a function of the velocity offset for $0\lesssim | {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \lesssim 75$ and decreases for $75\lesssim | {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \lesssim 100$. Thus, this PV diagram of W43A exhibits an S-like pattern that resembles the PV diagram of the WF IRAS 16342−3814 (Sahai et al. 2011; Tafoya et al. 2019). Tafoya et al. (2019) interpreted the PV diagram of IRAS 16342−3814 in terms of a jet-driven molecular outflow where the jet decelerates as it interacts with the surrounding material, transferring kinetic energy and momentum. The surrounding material and the jet produce the emission with positive and negative gradients, respectively, in the PV diagram. Given the spatial distribution of the emission seen in Figure 1 and the profile of the PV diagram of Figure 2, it is evident that the spatio-kinematical configuration of W43A is similar to the one of IRAS 16342−3814. Following the analysis done by Tafoya et al. (2019), the launch velocity of the jet can be obtained from the highest velocity offset of the CO(J = 2 → 1) emission, which in Figure 2 is indicated as $| {V}_{{\rm{i}}}| \sim 100\,\mathrm{km}\,{{\rm{s}}}^{-1}$. Considering the inclination of the source of 35° with respect to the plane of the sky, this velocity corresponds to a deprojected expansion velocity ${U}_{{\rm{i}}}\equiv | {V}_{{\rm{i}}}| /\sin 35^\circ =175$ km s⁻¹. The velocity offset at which the velocity gradient changes from positive to negative is $| {V}_{\mathrm{offset}}(\mathrm{km}\,{{\rm{s}}}^{-1})| \approx 75$. This value represents the velocity offset at the tip of the jet, V_f ≈ 75 km s⁻¹, and its corresponding deprojected expansion velocity is ${U}_{{\rm{f}}}\equiv | {V}_{{\rm{f}}}| /\sin \,35^\circ =130$ km s⁻¹.

This shows that the material at the tip of the jet has decelerated during its journey from the launching region to its present location. This behavior is expected if turbulent entrainment along the jet is a dominant process since, for such a case, the average velocity of the jet would decrease as momentum is transferred to the entrained material (Chernin et al. 1994). On the other hand, the overall velocity spread of the emission at a given position in the PV diagram, delimited vertically by the solid red and light-blue lines, decreases with the distance from central system as an effect of the lateral expansion and turbulent motion of the gas (Chernin et al. 1994; Balick et al. 2013). From Figure 2 it can be seen that beyond the radius of the outer edge of the H₂O masers clusters, $r={r}_{{{\rm{H}}}_{2}{\rm{O}}}$, the maximum velocity offset of the CO emission begins to drop significantly. Thus, it is likely that the largest deceleration occurs when the jet interacts with material of the shell, creating the physical conditions (T ≈ 400 K; ${n}_{{{\rm{H}}}_{2}}\approx {10}^{9}$ cm⁻³) that favor H₂O maser emission (Yates et al. 1997). This would explain the high concentration of H₂O masers in the interaction region. After the jet interacts with the shell, it continues streaming outward and decelerates even further.

The collimation factor of the CO structure that traces the jet, defined as the ratio between the length and the width, is very high (∼20). Surprisingly, the degree of collimation of the jet is not affected after passing the region of interaction with the bipolar shell. The persistence of the high degree of collimation of the jet after the interaction with the shell is likely due to the pressure of the magnetic field, since it is expected to dominate the gas pressure by a factor of a few hundred in the interaction region (Vlemmings et al. 2006). Even though from these observations there is no clear evidence that the jet is precessing, as opposed to what was thought from the H₂O maser observations earlier (Imai et al. 2002), it is possible that there have been some slight variations in the PA of the jet that may explain the changes seen in the distribution of the H₂O masers over the past couple of decades (Chong et al. 2015).

Considering that the velocity of the jet remains roughly constant from the launching point to the H₂O maser region, the kinematical timescale for the jet to reach this location is ∼30 yr. In addition, assuming that the jet suffers a constant deceleration in the region beyond the location of the H₂O masers, the kinematical timescale from the H₂O maser region to the tip of the jet is ∼35 yr. Thus, the total kinematical age of the jet is of just τ_jet = 65 yr.

In addition to CO(J = 2 → 1), the p-H₂S (${J}_{{K}_{{\rm{a}}},{K}_{{\rm{c}}}}$ = 2₂₀ → 2₁₁) line was also detected from the observations of this project. The H₂S emission is shown in Figure 5 in Appendix C superimposed on the continuum emission. As can be seen from Figure 5 in Appendix C, the H₂S emission has a spatial distribution that is significantly different from that of the CO emission. Particularly, while the later one seems to be confined by the shell in the direction perpendicular to the outflow, the former one extends beyond the shell. Figure 6 in Appendix C shows a PV diagram of the H₂S emission where it can be seen that it exhibits the characteristic morphology of a double-lobe structure, indicated with solid lines, similar to the morphology of the bipolar shell traced by the continuum emission. H₂S is known to trace shocked material (e.g., Holdship et al. 2016), and in the case of W43A it is most likely tracing the shock produced by the expansion of the shell. Thus, we assume that the H₂S emission traces not only the spatial distribution but also the kinematics of the bipolar shell. The kinematical age of the bipolar shell can be obtained as τ_shell ≈ R_shell/V_shell, where V_shell is the expansion velocity of the shell at any point located at a distance R_shell from the central source. The expansion velocity of the H₂S emission at the location of the H₂O masers (∼1100 au) is ≈50 km s⁻¹/sin(θ_inc) = 90 km s⁻¹, where θ_inc = 35°. The resulting timescale for the shell is τ_shell ≈ 60 yr.

Thus, the kinematic ages of the jet and the bipolar shell, traced by the continuum emission, are basically equal, strongly suggesting that the former created the latter as it interacted with material in the vicinity of the driving source. The difference in size between the jet and the bipolar shell, despite their common kinematic age, can be explained in the following way. When the highly collimated jet was launched for the first time, around 65 yr ago, it interacted with the dense material close to the central source creating the bipolar shell. As circumstellar material was swept up, the expansion of the bipolar shell slowed down. Additionally, since the jet is launched with a velocity higher than the resulting velocity of the shell, it opens its way out through the shell, creating the physical conditions that favor the H₂O maser emission in the region where the two components interact. Since the edges of the CO(J = 2 → 1) emission are clearly defined and there is no evidence of collimated emission beyond the structures seen in Figure 1, it is likely that W43A is indeed undergoing the earliest phase of collimated mass loss, as suggested previously (Imai et al. 2002). It is worth noting that recent numerical simulations presented by Balick et al. (2020) predict morphologies very similar to that of W43A for certain combinations of initial conditions of the density, velocity, and magnetic field. Therefore, these observational results are of utter importance to constrain such kind of numerical simulations.

The CO(J = 2 → 1) emission associated with the jet also exhibits a clumpy structure. The redshifted jet has nine knots, while the blueshifted jet exhibits only seven knots (see Figure 3). The knots are roughly associated in pairs, although there are some variations in their separations from the central source. The missing knots in the blueshifted jet coincide with the location of one of the continuum bright regions. Thus, it is likely that they are obscured by dense material in the dusty shell. The average knot spacing is 180 au. Similar clumpy structures are commonly found in jets associated with young stellar objects (Hirano et al. 2006, 2010; Lee et al. 2007; Santiago-García et al. 2009) and they are interpreted as due to periodic events of increase of the velocity within the jets (Santiago-García et al. 2009; Hirano et al. 2010). For the case of W43A, this could be explained in terms of a jet-launching mechanism associated with accretion in a binary system that hosts an AGB star and a companion with eccentric orbits. When the components are in periastron the accretion disk is perturbed and the mass-loss rate increases (Bonnell & Bastien 1992; Clarke & Syer 1996; Davis et al. 2013), leading to internal shocks in the jet that produce the observed clumpy structure (Hirano et al. 2006; Santiago-García et al. 2009). Given the knot spacing and the speed of the jet, the time interval between them is around 5–7 yr, which would be equal to the orbital period of the binary system.