The Explosion in Orion-KL as Seen by Mosaicking the Magnetic Field with ALMA

The maps of total intensity (Stokes I) emission from the 3.1 and 1.3 mm data are shown in Figure 1; these maps comprise the first-ever high-resolution (≲1'') observations of the entire Orion-KL nebula. The mosaics cover regions of ∼220 × 175 (15 pointings) at 3.1 mm and ∼125 × 094 (45 pointings) at 1.3 mm, and are centered on Orion-KL. The ALMA observations resolve the center of the nebula at both 3.1 and 1.3 mm, where BN, source I, source n, and source x are clearly visible, as indicated in the total intensity maps, with BN appearing as an isolated and unresolved object in both maps. We recover all of the continuum sources previously detected by others (Eisner & Carpenter 2006; Friedel & Widicus Weaver 2011; Hirota et al. 2015).

Zoom In Zoom Out Reset image size

The morphology of the dust emission appears filamentary, consistent with previous ALMA observations of Orion-KL (Hirota et al. 2015). Toward the center of the nebula, we detect a cavity from the center of the CO outflow to the North in both the 3.1 and 1.3 mm observations. This cavity has a "U" shape and surrounds the center of the explosion; it is likely that it was created by the explosive event.

We estimate the column density independently from the 3.1 and 1.3 mm data by assuming optically thin dust emission:

$$\begin{eqnarray}&&{N}_{{{\rm{H}}}_{2}}=\displaystyle \frac{{S}_{\nu }}{{B}_{\nu }({T}_{\mathrm{dust}}){\rm{\Omega }}{\kappa }_{\nu }\mu {m}_{{\rm{H}}}},\end{eqnarray} \tag{ 1 }$$

where S_ν is the total flux density of the source, B_ν (T) is the Planck function and T_dust is the dust temperature. The area is given by the solid angle, calculated as ${\rm{\Omega }}\,=\pi /(4.0\mathrm{log}2){b}_{\mathrm{maj}}\,{b}_{\min }$. The dust opacity κ_ν is set as 0.01 cm² g⁻¹ at 1.3 mm and 2.6 × 10⁻³ cm² g⁻¹ at 3 mm (Ossenkopf & Henning 1994). These values assume a gas-to-dust ratio of 100. We use μ = 2.33 as the mean molecular weight of hydrogen; m_H is the mass of a hydrogen atom. The optically thin assumption is justified, as preliminary spectral index estimations from these data suggest a spectral index of ≈3 in the region where we estimate the column density. For this work, we calculate the column density only in the region of the filament relevant to the estimation of the magnetic field strength (see Section 3.4 and Table 2). We use the dust temperature maps from the James-Clerk-Maxwell Telescope (JCMT) reported by Salji et al. (2015). Based on that work, we use a range for T_dust of between 20 and 50 K. By using all the pixels in the two maps shown in Figure 1, we obtained total masses of 35 and 88.2 M_⊙ from the 3 and 1 mm data, ¹⁵ respectively by using T_dust = 25 K as an average dust temperature.

Table 2. Magnetic Field Estimation

λ	Tdust	N	${n}_{{{\rm{H}}}_{2}}$	δϕ	ΔV	B1	B2	B3	$\left\langle {B}_{\mathrm{pos}}\right\rangle$	σB	EB
(mm)	(K)	(1024 cm−2)	(107 cm−3)	(degrees)	(km s−1)	(mG)	(mG)	(mG)	(mG)	(mG)	(1045 erg)
3.1	20	3.0	24.5	19.7	1.4	10.3	16.8	9.5	12.2	3.3	3.3
3.1	30	1.8	14.8	19.7	1.4	8.0	13.1	7.4	9.5	2.6	2.0
3.1	40	1.3	10.6	19.7	1.4	6.8	11.1	6.2	8.0	2.2	1.4
3.1	50	1.0	8.3	19.7	1.4	6.0	9.8	5.5	7.1	1.9	1.1
1.3	20	5.4	44.0	21.1	1.4	12.9	20.9	2.6	12.1	7.5	3.2
1.3	30	3.5	28.2	21.1	1.4	10.3	16.7	2.1	9.7	6.0	2.1
1.3	40	2.6	20.7	21.1	1.4	8.9	14.3	1.8	8.3	5.1	1.5
1.3	50	2.0	16.4	21.1	1.4	7.9	12.8	1.6	7.4	4.6	1.2

Note. Here we present the parameters used to estimate the magnetic field strength in Orion-KL, and the results of those estimates. We also show the value for the and energy (E_B ) inside the 12'' radius where the magnetic field has a radial pattern. The magnetic field energy was calculated using the average of the field strength onto the plane of the sky ($\left\langle {B}_{\mathrm{pos}}\right\rangle$) as shown in column 10.

Download table as:  ASCIITypeset image

Figure 2 shows the 1.3 mm mosaic of the magnetic field toward Orion-KL (the 3.1 mm map is shown in Appendix A); these figures show the magnetic field morphology inferred from the polarized dust emission under the assumption that dust grains are magnetically aligned by radiative torques (see Lazarian & Hoang 2007). We discuss the validity of this assumption in Section 4.1. The magnetic field pattern is consistent between both wavelengths. At 1.3 mm, the polarized emission covers most of the filament sampled by the dust total intensity, where the derive magnetic field morphology is in agreement with previous observations done at coarser angular resolution and/or lower sensitivity (Rao et al. 1998; Tang et al. 2010; Hull et al. 2014, see Hull et al. 2020a for more details).

Zoom In Zoom Out Reset image size

The magnetic field pattern can be divided into two distinct morphologies: (1) a component that appears quasi-perpendicular to the major axis of the filament, visible most clearly toward the northern extreme; and (2) a component at the center of Orion-KL, where the magnetic field morphology exhibits a pseudo-radial pattern emanating from the center of the nebula up to a distance of ∼12'', or 5000 au (see Figure 2). By extrapolating the field lines through the center of Orion-KL, we see that they intersect inside the 3'' × 5'' ellipse reported to be origin of the explosive CO outflow (Bally et al. 2017). The extrapolation of the field lines is shown in Figure 3. The transition from the quasi-radial morphology into an orientation quasi-perpendicular to the filament is smooth in both the northern and southern parts of the filament. Moreover, the quasi-perpendicular morphology resembles the hourglass shape seen by JCMT and the Stratospheric Observatory for Infrared Astronomy (SOFIA) at larger spatial scales (Pattle et al. 2017; Chuss et al. 2019). In fact, both morphologies, including the transition from radial to hourglass that we see in the ALMA data, can be seen in the SOFIA data (see the 53 μm map shown in the upper-left panel of Figure 1 in Chuss et al. 2019).

Zoom In Zoom Out Reset image size

Figure 4 shows the comparison between the magnetic field morphology and the spherically symmetric CO outflow. The magnetic field is superposed over the ¹²CO $(J=2\to 1)$ moment 0 maps of both the red- and blueshifted outflowing CO emission (see Bally et al. 2017). In many locations, the magnetic field seems to be aligned with the radial CO streamers, particularly to the West, where the field exhibits prominent radial, "finger"-like structures. Although the radial orientation seems clear, the locations where we detect dust polarization seem to be anti-correlated with the locations of the CO streamers. This anti-correlation is seen in maps of both the red- and blueshifted CO emission (see Figure 4). Outflows carving cavities in the dust have been seen in other star-forming regions; in several of these, the magnetic field has been seen to be elongated along the cavities of the outflow, similar to what we see in Figure 4 (Hull et al. 2017; Maury et al. 2018; Le Gouellec et al. 2019; Takahashi et al. 2019; Hull et al. 2020b).

Zoom In Zoom Out Reset image size

Comparing with the dust emission, we see that the finger-like structures in the CO extend to scales ∼3× larger than the region where we detect the total intensity dust emission (this can be seen by comparing Figures 1 and 4), in particular in directions perpendicular to the filament's major axis. Toward the center of the explosion, the field morphology is more complicated: the field is obviously perturbed, with clear departures from the large-scale field morphology.

To further explore the alignment between the CO streamers and the magnetic field, we use the Rolling Hough Transform (or RHT, see Clark et al. 2014). The RHT is a machine-vision algorithm that is well suited for the identification of linear structures in data (e.g., edge detection). We applied this algorithm to the CO streamers to highlight the linear structures in the data and to correlate them with our polarization map. The linear structures in the CO streamers are clearly identified by the RHT algorithm in both the red- and blueshifted lobes of the outflow (see the right-hand panels in Figure 4). Comparing these structures with the magnetic field, we can clearly see a correlation in the orientations, in particular as we move radially away from the center of the explosion. The RHT maps also confirm the anti-correlation between the dust and CO emission that is seen to the West of the Orion-KL nebula (and also to the East, to some degree), as previously mentioned. In addition, we clearly see radially oriented CO emission in regions where the magnetic field appears to be unperturbed: i.e., toward the ends of the filament where the dust morphology is compact, filamentary, and threaded by a quasi-hourglass field shape.

To estimate the strength of the magnetic field, we use three versions of the Davis–Chandrasekhar–Fermi method (DCF; Davis 1951; Chandrasekhar & Fermi 1953), as previously employed in Cortes et al. (2016, 2019). The DCF technique assumes that the magnetic field lines are perturbed by non-thermal turbulent motions, and thus we cannot directly estimate the field strength from the central region of Orion-KL that has been affected by the explosion. This is because the explosion injects an amount of energy such that the field lines perturbations are likely nonlinear, which makes the assumptions in the DCF technique not applicable. Instead, we use the NE region of the filament, where the field appears to be unperturbed by the explosion (indicated by the gray ellipse in Figure 2). The DCF method requires estimates of the gas number density, the velocity dispersion ΔV, and the dispersion in the magnetic field lines, or δϕ. To estimate the gas density, we assume a thin slab with a depth of half the region's minor axis, or n = N/L with L = 2'', where N is the estimated column density (see Table 2). For the velocity dispersion, we use the NH₃ observations from Wiseman & Ho (1998) performed using the Karl G. Jansky Very Large Array at 8'' resolution. They obtained a value of 1.4 km s⁻¹ at the location of Orion-KL (see their Table 2). Given that their resolution approximately matches the minor axis of our region of interest, we used that value for the line width in our field-strength estimation. The value of δϕ is calculated from our polarization data directly. Because the 3.1 and 1.3 mm data trace different length scales, we derive independent estimations of the field strength from each data set, and obtain average strengths between 7 and 12 mG. By averaging all of the estimates from Column 10 in Table 2, we obtain an average strength at 1.3 mm value of $\left\langle B\right\rangle =9.4$ mG. Note that these three flavors of the DCF method do not take into account the effect of finite angular resolution and integration along the line of sight, both of which tend to smooth the magnetic field and reduce angle dispersions, and thus we consider this estimate as an upper limit of the magnetic field strength. We present the results of these estimations and the parameters used in Table 2.

One of our main assumptions is that grains are aligned by magnetic fields. The other plausible mechanism—mechanical alignment—has two flavors. The classical mechanism requires a significant dust-gas drift (Gold 1952), and appears to be important only for highly irregular dust grains. Tang et al. (2010) discarded the classical Gold mechanism, arguing that the polarization pattern should be aligned with the flow; this is the opposite of the azimuthal polarization pattern that we see in Orion-KL. ¹⁶ The second flavor was introduced by Hoang et al. (2018) and is called "Mechanical Alignment Torques" (or MATs), which are proposed to operate in high-mass regions where radiation pressure drives the outflow. The spatial anti-correlation between the outflow fingers and the regions of radial magnetic field (as traced by the polarized dust emission) suggests that this mechanism is not occurring, as the CO and dust emission would need to be co-spatial for mechanical alignment to be plausible. Therefore, it seems that the most likely mechanism to explain the polarized emission is magnetically aligned dust grains.

The quasi-radial morphology of the magnetic field in Orion-KL suggests significant alignment of the magnetic field (projected onto the plane of the sky) with the outflow. A question that immediately emerges is whether this is because the field lines have been bent by the gas as it moved outward, or whether the field lines have been dragged inward by gravity. Radial magnetic field configurations have been attributed to gravitational collapse in other high-mass star-forming regions such as W51 (Tang et al. 2009; Koch et al. 2018) and W43-Main (Cortes et al. 2016, 2019; Arce et al. 2020). We will consider this question by evaluating the energy balance in the region surrounding the Orion-KL nebula.

Bally et al. (2020) estimated a total mass for the Orion-KL complex of ∼100 M_⊙, which yields a gravitational potential energy of E_W  = GM²/R = 3.52 × 10⁴⁶ erg, when considering a sphere of 12'' (5000 au) in radius enclosing the region where the field appears perturbed. This energy estimate is two orders of magnitude less that the maximum energy estimated in the outflow, or ∼10⁴⁸ erg (Bally et al. 2017). The magnetic field tension must be overcome in order to explain the quasi-radial morphology that we see in our maps. To test whether this is occurring, we must estimate the total magnetic energy in the perturbed region E_B  = ${B}^{2}{R}^{3}/12\pi \sim 2\times {10}^{45}$ erg, where we used a magnetic field strength of 9.4 mG. This energy value is ∼3 orders of magnitude less than the maximum estimated energy of the outflow. Thus, the energy in the explosion is sufficient to overcome even a magnetic field with a relatively large (∼10 mG) strength.

From this simple energetics comparison, it appears that the outflow is the dominant dynamical player in the region immediately surrounding the explosion center. However, it is important to remember that the magnetic field "feels" dynamical interactions through collisions with the charge carriers, as it is likely that the field is "frozen" into the charged fluid (Kulsrud 2005). Both fluids, neutral and charged, are subject to the gravitational field, but for the magnetic field to be dynamically perturbed by the outflow we need to estimate how well the charge carriers are coupled to the neutrals. To quantify this, we estimate the magnetic Reynolds number R_m . Values of R_m  ≫ 1 suggest strong coupling between the gas and the magnetic field; under such a scenario, the gas and the magnetic field should move as a single fluid. We estimate a value of R_m  = 1.7 × 10⁵ ≫ 1 by considering the same quantities used for the estimation of the magnetic field strength (see Appendix B for details). Similar values for R_m have been estimated in the literature toward giant molecular clouds (Myers & Khersonsky 1995). Even if the magnetic field strength were a factor of 10 less (∼1 mG), we would still obtain R_m ≫ 1. Based on this result, we conclude that the field is well coupled with the gas. Therefore, considering the difference in magnitude between the energies in the outflow and in the gravitational field, it is more likely that magnetic field morphology is the result of the field being stretched outward by the outflow rather than being pulled inward by gravity. This is particularly clear at the locations of the CO "fingers," which are located away from the main filament where the gravitational pull should be less important.

Among the plethora of features seen in emission from Orion-KL, the vibrationally excited H₂ and [Fe ii] emissions are the most visually striking (Allen & Burton 1993; Nissen et al. 2007; Bally et al. 2011, 2015, see also Figure 6). Vibrationally excited H₂ emission is a signpost of shock activity because this emission is one of the main coolants for the post-shock gas (Draine & McKee 1993), while [Fe ii] emission is commonly seen toward Herbig-Haro objects. Additionally, observations of emission from complex organic molecules (or COMs) toward the Orion-KL region strongly suggest that shocks have affected the chemistry. Because a shock injects a significant amount of kinetic energy into the medium, there can be increased production of molecules that are not normally abundant in quiescent gas. This is supported by BIMA, SMA, IRAM 30 m, and ALMA observations of a number of shock tracers such as SO₂, ³⁴SO₂, SO, and heated H₂CS, and also by the discovery of chemical segregation in COMs toward Orion-KL (Liu et al. 2002; Feng et al. 2015; Favre et al. 2017; Pagani et al. 2017, 2019; Tercero et al. 2018; Li et al. 2020). Recently, Pagani et al. (2019) found elongated formaldehyde (H₂CO) emission whose origin is consistent with that of the explosive event in Orion-KL. This is critical, because the elongation direction of the H₂CO correlates well with the radial morphology seen in the CO outflow and with the quasi-radial magnetic field morphology we are presenting here. We further note that the maximum elongation of the H₂CO emission (relative to the center of the explosive event) is approximately the same as the radius of the transition from quasi-radial to quasi-hourglass seen in the magnetic field morphology.

As previously mentioned, the magnetic field exhibits a quasi-radial morphology only up to scales of 12'' (∼5000 au), whereas both the outflow and the dust emission in the filament extend well beyond that radius. At this ∼5000 au radius, the field seems to continuously transition to the quasi-hourglass morphology seen on large scales by JCMT and SOFIA, far from the center of the explosion. We propose that this transition marks the position of a shock front that is propagating from the origin of the explosion in Orion-KL. To visualize this transition better, we plot a profile for the magnetic field as a function of distance from the center of the explosion. We present this in Figure 5, where we show the difference between a purely monopole magnetic field (i.e., a radial profile) and the field position angle from the 1.3 mm data. The quasi-radial morphology within the proposed shocked region is clearly seen, having an almost constant average difference value of ∼25°; the transition at a radius of ∼12'' manifests itself as a sharp slope in the difference profile. Furthermore, using the time and location of the explosive event as initial conditions, we obtain a speed of ∼50 km s⁻¹ for a shock traveling along the dust filament, up to 12'' in Orion-KL. Although the area in which the field appears perturbed (i.e., primarily radial) is clearly within a circle of ∼12'' in radius, the shock does not have to be spherically symmetric because the shock speed will depend on the ambient density in which the perturbation is propagating. Bally et al. (2015) presented H₂ and [Fe ii] emission that show "fingers" and wakes that are ≳30'' away from the explosion center and the dust filament. They suggests a perturbation traveling at high speeds (between 200 and 300 km s⁻¹), as measured from the proper motions of the H₂ and [Fe ii] "bullets" whose origin is also consistent with the origin of CO explosive outflow. We argue that this perturbation, or shock moving at ∼50 km s⁻¹, is slower because it is traveling through the much denser environment probed by the ALMA observations.

Zoom In Zoom Out Reset image size

To further support our proposal, we compare our 1.3 mm magnetic field map with the aforementioned H₂ emission map from Orion-KL (Bally et al. 2015). Figure 6 shows the magnetic field map superposed on the H₂ emission with semi-transparent ellipses indicating the different regions affected by the shock. The brightest region of H₂ emission is located outside the dust filament, where dust emission levels are below the ALMA detection threshold. When we look at our proposed post-shock region (as indicated by the orange semi-transparent circle in Figure 6), we do see diffuse H₂ emission in this post-shock region, which is expected, as vibrationally excited H₂ emission is a major coolant in shocked regions. However, beyond the shock front (i.e., the ∼12'' boundary), the H₂ emission appears significantly attenuated, in particular as we move NE along the main axis of the dust filament and away from the post-shock region into the regions of unperturbed magnetic field, which are shown as green semi-transparent ellipses in Figure 6.

Zoom In Zoom Out Reset image size

Unfortunately, from our data alone, we cannot conclude what type of shock is manifesting itself through the magnetic field morphology in Orion-KL. Furthermore, to the best of our knowledge, current models are only one-dimensional, with the field orthogonal to the shock propagation direction (with the notable exception of Pilipp & Hartquist (1994), who computed models using oblique magnetic fields with respect to the shock front); therefore, we do not have a model to which we can refer for comparison. In particular, explaining the smooth transition seen in the magnetic field morphology and the field's relationship with the shock front will require such models. Nonetheless, we hope that these findings will motivate much-needed multi-dimensional numerical work.