DR21 Outflow: Explosive Dynamics in Cygnus-X

• DR21 outflow is a powerful molecular feature in the high-mass star-forming region of Cygnus-X, characterized by extreme mass, momentum, and kinetic energy.
• High-resolution ALMA and SMA observations reveal radial CO streamers with Hubble-like velocity gradients that support an explosive dispersal interpretation.
• The region exhibits diverse chemical, thermal, and dynamical structures, with gamma-ray emission linking the outflow to high-energy processes and cosmic ray acceleration.

The DR21 outflow is the powerful molecular outflow associated with the high-mass star formation region DR21 in Cygnus-X. It has long been regarded as one of the most extreme outflows in the Milky Way in mass and size, but its dynamical classification has remained contested. In one line of work, the outflow is interpreted as an explosive dispersal outflow whose CO streamers radiate from a common center with Hubble–Lemaître-like expansion; in another, it is described as a strongly collimated bipolar protostellar outflow whose red- and blue-shifted overlap can be explained without invoking an explosion (Ccolque et al., 2024, Skretas et al., 2023).

1. Astrophysical setting and nomenclature

DR21 lies in the Cygnus-X star-forming complex at a distance of 1.5 kpc, within a larger filamentary environment that includes DR21(OH), W75S, and additional dense clumps and H II regions (Pandey et al., 2 Sep 2025, Skretas et al., 2023). The broader DR21 filament is not dynamically simple. Large-scale molecular-line studies found widespread infall signatures in optically thick HCO$^+$ and $^{12}$CO, a typical infall speed of $\sim 0.6$ km s$^{-1}$, and mass accretion rates of the order of a few $10^{-3}\,M_\odot\,{\rm yr}^{-1}$ for the two main clumps, supporting a state of global gravitational collapse rather than equilibrium turbulence (Schneider et al., 2010). In the neighboring DR21(OH) clump, high-resolution N$_2$H$^+$ and CH$_3$CN observations were interpreted in terms of convergent flows and low-velocity shocks, not classical high-velocity protostellar outflows (Csengeri et al., 2011).

This nomenclature matters because the phrase “DR21 outflow” is used in two ways in the literature. In the strict sense, it denotes the large DR21 Main outflow with eastern and western lobes and very large energetics (Skretas et al., 2023). In a broader complex-wide sense, it can also refer to distinct outflows in DR21(OH), including the methanol maser outflow and several compact CO/SiO flows associated with MM1/MM2 (Zapata et al., 2011). A common misconception is that every kinematic feature in the DR21 complex is outflow-driven; the filament studies show that infall, convergent flows, and sub-filamentary accretion are also major contributors to the observed velocity field (Schneider et al., 2010, Csengeri et al., 2011).

2. Morphological phenomenology

Historically, DR21 was known as a very energetic bipolar flow with eastern and western extensions seen in molecular gas and shocked H$_2$ emission (Ossenkopf et al., 2010, White et al., 2010). Herschel-SPIRE maps of the DR21 core showed a prominent central peak with extensions to the east and west aligned with the known outflow axis, and the high-$J$ CO ladder was interpreted as tracing warm outflow-associated gas rather than only a compact core (White et al., 2010). Near-infrared echelle spectroscopy of the DR21/W75N region detected 79 discrete H$^{12}$0 components; within that survey, object A9-1 was explicitly identified as a collimated flow corresponding to the Main DR21 outflow, and several DR21-related components exhibited radial velocities in excess of 80 km s$^{12}$1 (Smith et al., 2014).

The morphology changed substantially when interferometric CO($^{12}$2) imaging began to isolate high-velocity structure. SMA observations at $^{12}$3 resolution revealed about 25 high-velocity molecular filaments made up of roughly 300 compact emission features, arranged in a “finger” pattern radiating from a common center (Zapata et al., 2013). The authors argued that the apparent east–west bipolarity seen in H$^{12}$4 was misleading and could be shaped by extinction from a dense north–south dusty lane (Zapata et al., 2013).

ALMA Band 6 observations sharpened the picture. A 1.3 mm continuum mosaic at about 0.74″ $^{12}$5 0.29″ resolution and rms 0.06 mJy beam$^{12}$6 revealed the known compact H II regions plus five previously uncatalogued compact millimeter sources (MM1–MM5) (Ccolque et al., 2024). The decisive dataset was the CO($^{12}$7) cube at about 0.66″ $^{12}$8 0.24″ resolution. By manual channel-by-channel inspection outside the heavily contaminated cloud-velocity range near the systemic velocity, the study identified eighteen CO streamers distributed into narrow, coherent linear chains that radiate quasi-isotropically from one common origin (Ccolque et al., 2024). These streamers are interspersed on the sky, include both red- and blue-shifted structures, and are not organized as a standard bipolar pair (Ccolque et al., 2024).

The common origin was refined to about RA $^{12}$9, Dec $\sim 0.6$0 (J2000), within the cometary H II region containing sources A, B, and C (Ccolque et al., 2024). The absence of an embedded protostar or compact millimeter core exactly at that origin is a central morphological argument against a normal steady protostellar jet (Ccolque et al., 2024).

3. Kinematics and energetics

The outflow kinematics are extreme. In the SMA high-velocity CO data, the radial velocities span approximately $\sim 0.6$1 to $\sim 0.6$2 km s$\sim 0.6$3 for redshifted gas and $\sim 0.6$4 down to $\sim 0.6$5 km s$\sim 0.6$6 for blueshifted gas, with most filaments extrapolating back to a common systemic velocity of about $\sim 0.6$7 km s$\sim 0.6$8, matching the ambient molecular velocity of DR21 (Zapata et al., 2013). The ALMA analysis extended this picture: the blue-shifted emission spans roughly $\sim 0.6$9 to $^{-1}$0 km s$^{-1}$1 and the red-shifted emission roughly $^{-1}$2 to $^{-1}$3 km s$^{-1}$4 relative to the parent cloud, giving a total observed velocity spread from about $^{-1}$5 to $^{-1}$6 km s$^{-1}$7 (Ccolque et al., 2024).

A hallmark of the explosive interpretation is that individual streamers show approximately linear velocity gradients along their length, so that radial velocity increases with projected distance from the center (Ccolque et al., 2024). The ALMA analysis expressed the bulk physical estimates through the usual relations

$^{-1}$8

$^{-1}$9

and

$10^{-3}\,M_\odot\,{\rm yr}^{-1}$0

Assuming LTE and optically thin CO emission, and using excitation temperatures of 70–140 K, the total mass of the streamer ensemble was estimated as 120–210 $10^{-3}\,M_\odot\,{\rm yr}^{-1}$1, with total momentum of order $10^{-3}\,M_\odot\,{\rm yr}^{-1}$2 km s$10^{-3}\,M_\odot\,{\rm yr}^{-1}$3 and kinetic energy of about $10^{-3}\,M_\odot\,{\rm yr}^{-1}$4 erg (Ccolque et al., 2024). The adopted most distant extent was about 120″ and the characteristic velocity about 100 km s$10^{-3}\,M_\odot\,{\rm yr}^{-1}$5, yielding a dynamical age of 8600 yr (Ccolque et al., 2024).

The classical DR21 outflow literature also contains larger-scale, tracer-dependent energetic estimates. The 2013 SMA paper reiterated earlier values of mass $10^{-3}\,M_\odot\,{\rm yr}^{-1}$6 and energy $10^{-3}\,M_\odot\,{\rm yr}^{-1}$7 erg, and framed DR21 as about 20 times more energetic than Orion BN/KL (Zapata et al., 2013). By contrast, the CASCADE 3.6 mm study, using HCO$10^{-3}\,M_\odot\,{\rm yr}^{-1}$8 as the main outflow tracer and the separation method, derived a total outflow mass of $10^{-3}\,M_\odot\,{\rm yr}^{-1}$9, momentum of $_2$0, kinetic energy of $_2$1 erg, and dynamical times of roughly 3100–4900 yr (Skretas et al., 2023). In that same analysis, the opening angles of both lobes decreased systematically with velocity, from about $_2$2 to $_2$3 over the range 5 to 45 km s$_2$4 relative to the source velocity (Skretas et al., 2023). This suggests that different tracers and velocity selections isolate different components of the DR21 flow.

4. Explosive dispersal outflow versus bipolar protostellar outflow

The explosive interpretation entered the DR21 literature explicitly with the SMA study titled “A 10,000 Years Old Explosion in DR21” (Zapata et al., 2013). Its argument combined five elements: a radial filamentary morphology rather than broad bipolar lobes, Hubble-like position–velocity behavior, a common kinematic origin near $_2$5 km s$_2$6, the lack of a clear central powering source, and close similarity to the Orion BN/KL explosive outflow (Zapata et al., 2013). The inferred explosion center was given as

$_2$7

and the authors suggested that DR21-D could be a displaced remnant of the disrupted system (Zapata et al., 2013). The proposed physical origin was the disintegration of a massive non-hierarchical stellar system, analogous in kind to Orion BN/KL but larger in scale and older in dynamical age (Zapata et al., 2013).

For several years DR21 remained a debated case. The G5.89 explosive outflow paper described DR21 as the “marginal case” that had stood beside Orion BN/KL as the only suggested but not fully secure example (Zapata et al., 2020). That historical status changed with the ALMA DR21 study, which stated that the morphology and kinematics of the CO streamers confirm the presence of an explosive dispersal outflow at the heart of DR21 (Ccolque et al., 2024). Its strongest evidence was the combination of many narrow CO streamers, a common center, radial rather than bipolar geometry, linear velocity–distance gradients, and very large mass, momentum, and energy concentrated into a compact region (Ccolque et al., 2024).

The competing view is represented most clearly by the CASCADE analysis of DR21 Main (Skretas et al., 2023). Using combined IRAM 30 m and NOEMA 3.6 mm data at $_2$8 resolution, that study found integrated intensity maps of the HCO$_2$9 emission revealing a strongly collimated bipolar outflow with substantial red/blue overlap, hollow cavities in both lobes, and no evidence for elongated, filament-like structures of the type expected in confirmed explosive outflows (Skretas et al., 2023). It emphasized that red/blue overlap is not diagnostic by itself, because a bipolar flow close to the plane of the sky can show the same behavior (Skretas et al., 2023). The paper also argued that several weak structures are better interpreted as cavity walls than explosion debris, and that north–south HCO$^+$0 emission is ridge material rather than a second orthogonal outflow (Skretas et al., 2023).

The literature therefore contains two explicit and incompatible characterizations of DR21 Main. A plausible implication is that the disagreement is sensitive to tracer choice, angular filtering, and the morphological criteria used to distinguish explosive ejecta from cavity walls. That inference is consistent with the fact that the explosive case is built primarily from high-velocity CO streamer identification, whereas the bipolar case is built primarily from HCO$^+$1, HCN, and dense-gas/shock tracers over the full outflow extent (Ccolque et al., 2024, Skretas et al., 2023).

5. Thermal, chemical, and shock structure

The DR21 outflow is chemically and thermally stratified. Herschel-SPIRE spectroscopy over 196–671 $^+$2m detected 19 molecular lines plus atomic lines, including the CO ladder from $^+$3 through $^+$4 (White et al., 2010). LVG/RADEX modeling of the SPIRE CO and $^+$5CO data yielded a warm component with $^+$6 K, $^+$7, $^+$8, and filling factor $^+$9, interpreted as a warm, moderately dense, clumpy component associated with the outflow (White et al., 2010). A lower-excitation component with $_3$0 K and $_3$1 emerged when the SPIRE data were combined with ground-based CO observations, indicating that the outflow environment is multiphase (White et al., 2010).

Herschel-HIFI observations refined the heating problem. One HIFI study concluded that, despite the known shock-excitation of H$_3$2 and the clearly visible strong outflow, the emission of all lines up to $_3$3 THz can be explained by purely radiative heating in two clumpy PDR ensembles, with hot dense clumps close to the central cluster probably dynamically affected by the outflow and a more widespread distribution of cooler dense clumps (Ossenkopf et al., 2010). This is an important corrective to the assumption that a powerful outflow must dominate the thermal budget of every warm molecular component. Another HIFI study focused on water abundance and found broad outflow emission in both $_3$4CO $_3$5 and H$_3$6O near 1100 GHz, with the H$_3$7O broad emission having FWHM $_3$8 km s$_3$9 centered at $_2$0 km s$_2$1 (Tak et al., 2010). The outflow para-H$_2$2O abundance was derived as

$_2$3

far above the abundances inferred for the dense core and the foreground cloud, and interpreted as the result of warm gas liberating water from icy grain mantles (Tak et al., 2010).

The ionized and irradiated interfaces are also well constrained. HIFI detection of ortho-H$_2$4O$_2$5 absorption at 1115 GHz toward DR21 yielded a lower limit

$_2$6

and the velocity distribution closely matched OH 6 cm absorption and [C II] $_2$7m, indicating a thin layer of hot gas on clump surfaces directly facing the blue-shifted blister outflow (Ossenkopf et al., 2010). At $_2$8 pc resolution, para-H$_2$9CO mapping of the DR21 filament found 24–114 K kinetic temperatures with a mean of $J$0 K; it found no direct evidence that dense gas in DR21 is heated by shocks from the putative $J$1-yr-old explosion, but it did identify DR21W1 as a shock-dominated region associated with the western DR21 flow, with temperatures in the range 33–88 K, exceptionally high non-thermal linewidths, and high Mach numbers (Zhao et al., 2024).

Spatially resolved far-infrared spectroscopy with SOFIA FIFI-LS added a lobe-by-lobe view (Karska et al., 19 Mar 2025). The western lobe is dominated by high-$J$2 CO, [O I] $J$3m, and OH, while the eastern lobe is dominated by [C II] $J$4m, [O I] $J$5m, and [O III] (Karska et al., 19 Mar 2025). Non-LTE modeling of [O I] line ratios suggested $J$6 in the west and $J$7 in the east, consistent with a denser shock-dominated western interaction region and a more PDR-like or ionized eastern cavity (Karska et al., 19 Mar 2025). Assuming that the bulk of the far-infrared line emission arises in shocks, the study derived an outflow power of $J$8 and a mass-loss rate of $J$9 (Karska et al., 19 Mar 2025).

6. Related outflows in DR21(OH) and the surrounding complex

A full account of “DR21 outflow” usage must distinguish DR21 Main from the neighboring DR21(OH) outflows. DR21(OH), about 3 arcmin north of DR21, hosts a cluster of compact millimeter sources and multiple compact outflows rather than one dominant flow (Orozco-Aguilera et al., 2018, Zapata et al., 2011). SMA continuum observations at $^{12}$00 resolution resolved nine compact sources in MM1/MM2 and found that the most distinctive outflow there is a well-collimated east–west low-velocity methanol/formaldehyde outflow with blueshifted emission to the east and redshifted emission to the west, centered on SMA4 in MM2 (Zapata et al., 2011). Its characteristic velocities are near the systemic value, roughly $^{12}$01 to $^{12}$02 km s$^{12}$03, and the integrated CH$^{12}$04OH and H$^{12}$05CO fluxes of 80 to 150 Jy km s$^{12}$06 correspond to isotropic luminosities of about $^{12}$07 (Zapata et al., 2011). The same SMA study showed that high-velocity CO and SiO trace other compact outflows from MM1/MM2 and are not related to the low-velocity methanol maser outflow (Zapata et al., 2011).

The maser phenomenology in DR21(OH) is correspondingly rich. EVLA and SMA imaging showed that the distribution of 36 GHz class I methanol masers in the DR21(OH) outflow is similar to that of the other class I transitions, with numerous multitransition overlaps; all 14 of the 36 GHz masers in the outflow region are coincident, within positional uncertainty, with 44 GHz masers, and class I masers at 36 and 229 GHz occur in virtual overlap with class II 6.7 GHz masers at the main continuum source (Fish et al., 2010). A separate SMA line survey of the CH$^{12}$08OH maser outflow derived a methanol rotational temperature of $^{12}$09 K and column density of $^{12}$10 at the best-constrained outflow spot, while not detecting complex molecules such as CH$^{12}$11OCHO, $^{12}$12, and CH$^{12}$13CH$^{12}$14CN (Orozco-Aguilera et al., 2018). More recently, wide-band LTE and RADEX modeling of DR21(OH)/N44 identified a distinct broad outflow component in CH$^{12}$15OH and H$^{12}$16CO at $^{12}$17 km s$^{12}$18 with $^{12}$19 km s$^{12}$20, reinforcing the view that DR21(OH) is a highly dynamical, actively star-forming environment with multiple flows and shocks (Freeman et al., 2024).

These DR21(OH) results are not measurements of the DR21 Main outflow itself. Their importance is classificatory: they show that the DR21 complex contains several physically distinct outflow systems, so shorthand references to “the DR21 outflow” must be interpreted with care.

7. Galactic context and high-energy significance

DR21 now occupies a central place in the emerging Galactic class of explosive dispersal outflows. The 2024 ALMA study stated that five dispersal explosive outflows associated with massive star-forming regions have been confirmed in the Galaxy: Orion BN/KL, G5.89-0.39, S106-IR, IRAS 16076-5134, and IRAS 12326-6245, with DR21 added as another clear case (Ccolque et al., 2024). It also stressed that their frequency of occurrence in the Galaxy and the originating nature are still uncertain (Ccolque et al., 2024). In the G5.89 ALMA paper, DR21 served as the benchmark older system whose kinematic age helped motivate an occurrence estimate of about one explosion every $^{12}$21 yr, close to the supernova rate; in that narrative DR21 moved from being an isolated marginal exception to a member of a small but growing empirical class (Zapata et al., 2020).

The broader significance of DR21 expanded further with the report of gamma-ray emission spatially coincident with the outflow (Pandey et al., 2 Sep 2025). Using 15 years of Fermi-LAT data in the 0.2–500 GeV range, the study found a detection at greater than $^{12}$22, best described spatially by an extended radial Gaussian with $^{12}$23 and spectrally by a power law with an exponential cutoff with

$^{12}$24

Assuming the 1.5 kpc distance, it inferred a gamma-ray luminosity

$^{12}$25

in the 0.1–500 GeV band, and estimated that no more than about 15\% of the outflow’s kinetic power is converted into cosmic-ray acceleration (Pandey et al., 2 Sep 2025). The gamma-ray morphology overlaps dense gas tracers and allWISE mid-infrared emission, and the paper interpreted the emission in terms of hadronic $^{12}$26 collisions in a dense star-forming environment (Pandey et al., 2 Sep 2025). If that interpretation holds, explosive dispersal outflows become not only a dynamical subclass of massive-star feedback, but also a newly recognized class of Galactic high-energy sources.