ALMA Observations of the Extraordinary Carina Pillars: HH 901/902

In Figure 1 we present the HST/ACS image from Smith et al. (2010) of the optical jet HH 901 overlaid with the ALMA high-velocity CO(2−1) and the 1.3 mm continuum emission. The CO emission reveals a bipolar and collimated outflow that is tracing the most inner part of the optical jet in an east–west orientation (with an approximate size of 5000 au). This map shows that the blueshifted emission is located toward an eastern orientation, while redshifted emission is to its western side. We integrate in radial velocities from +6.4 to −1.2 km s⁻¹ for the redshifted emission, and from −8.2 to −14.6 km s⁻¹ for the blueshifted emission for the CO outflow in the HH 901 object. The CO emission between −1.2 km s⁻¹ and −8.2 km s⁻¹ is associated with the dusty pillar as is shown in Figure 2. The LSR systemic velocity of the entire pillar associated with the object HH 901 is about −5.0 km s⁻¹ (Figure 3). The radial velocities of the innermost CO outflow are in agreement to those presented in Reiter & Smith (2014). They found, using slit positions of the optical [S ii] and Hα toward the HH 901, redshifted emission in the west side of the flow with radial velocities about +25.0 km s⁻¹.

The 1.3 mm continuum image traces the disk and possibly part of the envelope at slightly larger scales. From a Gaussian fitting to the continuum source from a map obtained using a ROBUST parameter equal to 0.5 (yielding a better sensitivity), we obtained a deconvolved size of 049 ± 003 × 033 ± 003 with an PA of 100° ± 7° and an integrated flux of 6.8 ± 0.4 mJy Beam⁻¹. The corresponding physical sizes of these deconvolved values are about 700 au, which suggests that we are also seeing part of the envelope and disk. Assuming that the dust emission is optically thin and isothermal, the dust mass (M_d) is directly proportional to the flux density (S_ν) integrated over the source as

$$\begin{eqnarray}&&{M}_{d}=\displaystyle \frac{{D}^{2}{S}_{\nu }}{{\kappa }_{\nu }{B}_{\nu }({T}_{d})},\end{eqnarray} \tag{ 1 }$$

where D is the distance to the Carina Nebula (2.3 ± 0.1 kpc, Smith & Brooks 2008), κ_ν the dust mass opacity, and ${{\rm{B}}}_{\nu }({T}_{d})$ the Planck function for the dust temperature T_d. In reality, thermal dust emission is probably not optically thin, hence the estimated mass is considered to be the lower limit. Assuming a dust mass opacity (κ_ν) of 0.015 cm² g⁻¹ (taking a dust-to-gas ratio of 100) appropriate for these wavelengths (1.3 mm; Ossenkopf & Henning 1994), a typical opacity power-law index β = 1.2 (obtained from our ALMA data), as well as a characteristic dust temperature of 50 K, we estimated a lower limit for the mass of the most compact part (i.e., not interferometrically filtered) of the disk and envelope system associated with HH 901 of about 0.1 M_⊙. The mass uncertainty could be very large (a factor of up to 3 or 4) given the uncertainty in the opacity and in the dust temperature.

We can also estimate the volumetric and column density following the procedure described in Hernández-Hernández et al. (2014). We obtained a lower limit for the column density of 10²⁴ cm⁻² and a volumetric density of 10⁸ cm⁻³ for the dusty source associated with the HH 901 object. These values are observed in very dense regions as compact envelopes and circumstellar disks toward the nearest giant molecular cloud (GMC), Orion (Takahashi et al. 2013; Teixeira et al. 2016).

If we assume that the continuum emission at these millimeter wavelengths is mostly arising from the disk, with a very small contribution of the envelope, the 0.1 M_⊙ could be linked to the disk. If we further adopt a value of M_star/M_disk between 1 and 10 (Bate 2018), we can estimate that the protostar in the middle of the HH 901 object should be a low-mass star. Furthermore, if the envelope is even more massive and the disk mass is smaller, we will have a protostar of lower mass. However, given the uncertainties on the mass estimation for the disks (a factor of 3 or 4), it could also be possible that the star in the middle is an intermediate-mass star. This is in agreement with recent observations (Reiter & Smith 2013, 2014).

In Figures 2 and 3 show the kinematics and distribution of the molecular gas using the C¹⁸O, N₂D⁺, and ¹²CO spectral lines. These figures present the moment-zero emission from ¹²CO together with the moment one of the C¹⁸O, N₂D⁺, and ¹²CO lines. In order to compute these moments we integrate in radial velocities from −14.6 to +6.4 km s⁻¹, which includes the emission from the pillars and the outflows. The molecule that better traces the pillars and the outflows is ¹²CO, see Figure 1. From this figure, one can see how the pillar is very well defined and with a clear east–west velocity gradient going from the blueshifted to redshifted, respectively. In the southernmost side of the pillar, the bipolar outflow of HH 901 is evident. The bipolar outflow is only observed in the area traced by the pillar probably because the critical density for CO(2−1) (∼10⁴ cm⁻³) decreases drastically outside of the pillar and the outflow emission becomes too faint at larger spatial scales to be detected. Both lines, C¹⁸O and N₂D⁺, confirmed the east–west velocity gradient clearly revealed by the CO, as seen in Figure 3. However, for the case of N₂D⁺, this molecule is only observed in a small solid angle coincident with the 1.3 mm continuum image, probably tracing the densest parts of the pillar where the star formation is taking place. The critical density of N₂D⁺ J = 3−2 is about 3 × 10⁶ cm⁻³, which explains the fact that the N₂D⁺ line has better spatial correlation with the 1.3 mm dust continuum emission. Matsushita et al. (2019) reported that the N₂D⁺ line is strongly depleted toward the dusty massive core MM5/OMC-3 and this is explained in terms of chemical evolution. Once CO evaporates in the gas phase, molecules like N₂H⁺ or N₂D⁺ will not have an efficient formation process. This is, however, not observed in our maps probably because a better angular resolution or sensitivity is needed.

As CO(2−1) is probably optically thick, we can use C¹⁸O to estimate the opacity (τ₀) and then to estimate the mass of the pillar and the outflow. Assuming local thermodynamic equilibrium, we can estimate the masses using the equations presented in Scoville et al. (1986) and Palau et al. (2007) for the J = 2−1 transition, obtaining

$$\begin{eqnarray}\begin{array}{rcl}\left[\displaystyle \frac{{M}_{{H}_{2}}}{{M}_{\odot }}\right] & = & 1.2\times {10}^{-15}\,{T}_{\mathrm{ex}}\,{e}^{\tfrac{16.59}{{T}_{\mathrm{ex}}}}{X}_{\displaystyle \frac{{H}_{2}}{{CO}}}\\ & & \times \left[\displaystyle \frac{\displaystyle \int {{\rm{I}}}_{\nu }{dv}}{\mathrm{Jy}\,\mathrm{km}\,{{\rm{s}}}^{-1}}\right]\left[\displaystyle \frac{{\theta }_{\mathrm{maj}}\,{\theta }_{\min }}{{\mathrm{arcsec}}^{2}}\right]{\left[\displaystyle \frac{D}{\mathrm{pc}}\right]}^{2}.\end{array}\end{eqnarray} \tag{ 2 }$$

We used 2.8 as mean molecular weight, an abundance ratio between the molecular hydrogen and the carbon monoxide (${X}_{\tfrac{{H}_{2}}{{CO}}}$) of ∼10⁴ (Scoville et al. 1986), an excitation temperature (T_ex) of 50 K, an average intensity (I_ν) of 0.01 Jy Beam⁻¹ for the outflow, and 0.06 Jy Beam⁻¹ for the pillar; we take a velocity range (dv) of 5.5 km s⁻¹ for the outflow and 2.5 km s⁻¹ for the pillar, and a distance (D) to the source of 2.3 ± 0.1 kpc. We then estimate a gas mass for the outflow of ∼5 × 10⁻⁴ M_⊙ and for the pillar of ∼0.1 M_⊙. For the mass of the outflow, we can correct this value for the inclination, therefore the true velocity range is dv_real = $\tfrac{{dv}}{\sin \theta }$, where θ is the inclination angle taken to be 20° (Reiter & Smith 2013, 2014), so dv_real = 16 km s⁻¹ and the mass is 10⁻³ M_⊙.

The uncertainty in the values of the mass arises mainly from the error in the distance, which is only 4%, the flux density, which is about 10%, and the excitation temperature. Given that the C¹⁸O intensity line is approximately five times weaker than the ¹²CO line, we estimate an average opacity (τ₀) of 0.2, so we corrected the mass value by a factor of $\left[\tfrac{{\tau }_{0}}{1-\exp (-{\tau }_{0})}\right]$. Furthermore, this mass estimation is in agreement with the mass obtained for the HH 901 pillar using a volumetric density of 10⁵ cm⁻³ and a physical volume of 10'' × 2'' × 2'' obtained in Reiter & Smith (2013) and with a value of about 0.1 M_⊙.

For the outflow, the resulting kinetic energy (E_k = $\tfrac{1}{2}{{mv}}^{2}$) is about 10⁴² ergs and the momentum (p = mv) is 1 × 10⁻² M_⊙ km s⁻¹. These values are slightly higher than those estimated by Lumbreras & Zapata (2014) and Zapata et al. (2015, 2018) for outflows associated with low-mass protostars, but possibly consistent if the correction for outflow inclination is taken into account in these studies. If we estimate the mechanical force for the outflow F_CO, we obtain a value of 1.5 × 10⁻⁵ M_⊙ yr⁻¹ km s⁻¹. This value is low compared to those obtained for intermediate-mass young stars reported in Table 7 of van Kempen et al. (2016), again suggesting that this outflow is associated with a low-mass or in the limit of an intermediate star. If we add the momentum and energy from the optical counterpart it could be an intermediate-mass star. Here, it is important to mention that the outflows in Carina are strongly irradiated, and maybe most of the energy budget is already in the optical irradiated counterparts, thus a direct comparison with the outflows of van Kempen et al. (2016) could be not so straight.

As already mentioned in Section 1, Reiter & Smith (2013, 2014) suggested that the HH 901 jet is energized by a relatively young intermediate-mass star (2–8 M_⊙). However, this estimation of the mass (and evolutionary stage) is very poorly constrained because they do not have direct detection of the IR flux from the driving source of HH 901 (Reiter & Smith 2014). This assumption is only inferred by the large mass-loss rates of a few 10⁻⁶ M_⊙ yr⁻¹ (these are about of two orders of magnitude larger than those observed in T-Tauri stars) obtained from the optical ionized lines. These large mass-loss rates can be later transformed to mass-accretion rates and then to luminosity of the exciting star. However, the possibility that the powering source of HH 901 is of intermediate-mass is in agreement with our ALMA data, but this intermediate-mass star must be on the low range of mass as suggested by the outflow energy.

Given the high mass-loss rate reported in the HH 901 object (10⁻⁶ M_⊙ yr⁻¹), the protostar in the middle could be a Class 0/I object, so this should be very obscured even in the IR wavelengths. This is in agreement with the fact that the protostar is still surrounded by an envelope traced by the 1.3 mm and N₂D⁺ emission.

Following Mesa-Delgado et al. (2016), we can estimate the photoevaphoration timescale for the pillar or globulette associated with the HH 901 object. Taken Equation (2) (already discussed in that paper), which allows us to estimate the theoretical value of the globule/disk mass photoevaporation rate $\dot{M}$ by extreme ultraviolet radiation (EUV),

$$\begin{eqnarray}\begin{array}{rcl}{\dot{M}}_{\mathrm{EUV}} & \simeq & 4\times {10}^{-9}\ {\left(\displaystyle \frac{{F}_{\mathrm{EUV}}}{{10}^{10}{{\rm{s}}}^{-1}{\mathrm{cm}}^{2}}\right)}^{1/2}\\ & & \times \ {\left(\displaystyle \frac{R}{100\mathrm{AU}}\right)}^{3/2}\ {M}_{\odot }{\mathrm{yr}}^{-1},\end{array}\end{eqnarray} \tag{ 3 }$$

where F_EUV is EUV flux and R is the radius of the globule or disk. As the massive cluster Tr 14 is the closest (1.0 pc) to the HH 901 object, we ignored at this point the radiation for the massive cluster located also in this region Tr 16. The Tr 14 cluster has an UV luminosity (Q_H) of $2\times {10}^{50}\,\mathrm{photon}\ {{\rm{s}}}^{-1}$ (Smith & Brooks 2008), implying ${F}_{\mathrm{EUV}}={10}^{12}\,\mathrm{photon}\ {{\rm{s}}}^{-1}\ {\mathrm{cm}}^{-2}$ at the distance of the HH 901 object. For a radius of 1000 au, and using Equation (3), we obtain a mass photoevaporation rate $\dot{M}$ equal to 1 × 10⁻⁶ ${{\rm{M}}}_{\odot }$ yr⁻¹. Therefore, taking a mass of 0.1 M_⊙ for the pillar of HH 901, we obtain that the pillar will be photoevaporated in around 10⁵ yr. This timescale is shorter than that estimated, for example, for the massive evaporating gas globule (EGG) or globulette 105–600 (Mesa-Delgado et al. 2016) of about 10⁶ yr. However, the 105–600 globulette is further away from the Tr 14 massive cluster (17 pc), and is more massive (0.1 to 0.3 M_⊙), which explains the large difference.

In Figure 4 we present the HST/ACS image from Smith et al. (2010) of the optical jet HH 902 overlaid with the ALMA high-velocity CO(2−1) and the 1.3 mm continuum emission. As in the HH 901 object (Figure 1), the CO emission also reveals a bipolar and collimated outflow that is tracing the most inner part of the optical jet in a northwest and southeast orientation. This map reveals that the blueshifted emission is located toward an eastern orientation, while redshifted emission is to its western side, similar to the HH 901 object. We integrate in radial velocities from +3.7 to −5.6 km s⁻¹ for the redshifted emission, and from −10.8 to −15.2 km s⁻¹ for the blueshifted emission for the CO outflow in the HH 902 object. The CO emission between −5.5 km s⁻¹ and −10.7 km s⁻¹ is associated with the dusty pillar as is shown in Figure 2. The LSR systemic velocity of the entire pillar associated with the object HH 902 is about −8.5 km s⁻¹. The radial velocities of the innermost CO outflow are in agreement to those presented in Reiter & Smith (2014).

The 1.3 mm continuum emission in Figure 4 is also tracing, as in Figure 1, the disk+envelope associated with the HH 902 object. However, in this figure two objects, named A and B, are revealed. Component A is associated with the HH 902 object, while component B is to its northwest. Component B is not associated with any molecular outflow, maybe because it is still in its pre-stellar phase. This source is extended with a size of about 2000 au and has a flux density of 38 ± 2.0 mJy Beam⁻¹. Assuming a dust temperature of 30 K, this flux density corresponds to a mass of 1.5 ${{\rm{M}}}_{\odot }$, which is likely associated with an extended envelope. For the millimeter component A, from a Gaussian fitting to the continuum source from a map obtained using ROBUST equals to 0.5 (with a better angular resolution), we obtained a deconvolved size of 034 ± 002 × 028 ± 002 with an PA of 107° ± 13° and an integrated flux of 6.5 ± 0.7 mJy Beam⁻¹. The corresponding physical sizes of these deconvolved values are about 700 au, which again suggests that we are also seeing part of the envelope and disk.

Following Equation (1) and taking similar values to the ones assumed for the object HH 901, we estimated a lower limit (because again the thermal dust emission is probably not optically thin) for the mass of the most compact part of the disk and envelope system (of component A) associated with HH 902 of about 0.1 M_⊙. In a similar way as for the HH 901 continuum object, this mass for the dust for the HH 902 object corresponds to a low-mass protostar.

In Figures 5 and 6, the kinematics and distribution of the molecular gas using the C¹⁸O, N₂D⁺, and ¹²CO spectral lines are presented. These figures present the moment-zero emission from ¹²CO together with the moment one of the C¹⁸O, N₂D⁺, and ¹²CO lines. Similar to Figures 3 and 4, in order to compute these moments we integrate in radial velocities from −15.2 to +3.7 km s⁻¹, which includes the emission from the pillars and the outflow. The molecule that better traces the pillars and the outflow is again ¹²CO, as seen in Figure 5, but now for the HH 902 object. From this figure, one can see how the pillar is very well defined and with a clear east–west velocity gradient going from the blueshifted to redshifted, respectively. In the most southern side of the pillar, the bipolar outflow of HH 902 is evident. Both lines, C¹⁸O and N₂D⁺, confirm the east–west velocity gradient clearly revealed by CO, as seen in Figure 5. However, for the case of N₂D⁺, this molecule again is only observed in a small solid angle coincident with the 1.3 mm continuum image, probably tracing the densest parts of the pillar, as in the HH 901 object.

Using Equation (2) and assuming similar values for the excitation temperature, an abundance ratio, and distance, we can obtain the mass for the outflow and the pillar associated with the HH 902 object. We take an average intensity (I_ν) of 0.015 Jy Beam⁻¹ for the outflow and 0.07 Jy Beam⁻¹ for the pillar, and we also take a velocity range (dv) of 5.0 km s⁻¹ for the outflow and 3.0 km s⁻¹ for the pillar. For these values, we estimate a gas mass for the outflow of 7 × 10⁻⁴ M_⊙ and for the pillar of ∼0.2 M_⊙. We also corrected these mass values by the opacity. Moreover, correcting these values by the inclination, we obtain a mass for the outflow of 10⁻³ M_⊙.

Taking again Equation (3) and similar values for the EUV radiation (${F}_{\mathrm{EUV}}={10}^{12}\,\mathrm{photon}\ {{\rm{s}}}^{-1}\ {\mathrm{cm}}^{-2}$) at the distance of the HH 902 object and a radius of 4 × 10³ au, we obtain a mass photoevaporation rate $\dot{M}$ equal to 1 × 10⁻⁵ M_⊙ yr⁻¹. Therefore, taking a mass of 0.2 M_⊙ for the pillar HH 902, we obtain that the pillar will be photoevaporated in around 2 × 10⁴ yr, which is a lower value than the timescale of the HH 901.

Given the short photoevaporated timescales of the molecular pillars and that the protostars in these pillars are still very embebed, we suggest that the disks inside of the pillars will be quickly affected by the radiation of the massive stars (in about 10⁴—10⁵ yr), forming proplyds like those observed in Orion (O'Dell et al. 1993, 2008; Henney and O'Dell 1999).