HH 80/81: Structure and Kinematics of the Fastest Protostellar Outflow

Hubble Space Telescope (HST) images obtained in 2018 are combined with archival HST data taken in 1995 to detect changes and measure proper motions in the HH 80/81 shock complex, which is powered by the fastest known jet driven by a forming star, the massive object IRAS 18162-2048. Some persistent features close to the radio jet axis have proper motions of >1000 km s−1 away from IRAS 18162-2048. About 3–5 pc downstream from the IRAS source and beyond HH 80/81, Hα emission traces the rim of a parsec-scale bubble blown by the jet. Lower speed motions are seen in [S ii] away from the jet axis; these features have a large component of motion at right angles to the jet. We identify new HH objects and H2 shocks in the counterflow opposite HH 80/81. The northeastern counterflow to HH 80/81 exhibits an extended but faint complex of 2.12 μm H2 shocks. The inner portion of the outflow is traced by dim 1.64 μm [Fe ii] emission. The full extent of this outflow is at least 1500″ (∼10 pc in projection at a distance of 1.4 kpc). We speculate about the conditions responsible for the production of the ultrafast jet and the absence of prominent large-scale molecular outflow lobes.

The distance to the L291 cloud was constrained by Añez-López et al. (2020), who used both the increase of stellar polarization towards stars near IRAS 18162-2048 and the abrupt increase of extinction as a function of decreasing parallax angle.These polarization data imply a distance of 1248±66 pc.Their extinction measurements imply a distance 1270±65 pc.Zucker et al. (2020) used Gaia DR2 parallax and reddening measurements of stars in the L291 field and found that the extinction increases dramatically at a distance of about 1,400±70 pc .Because the Zucker et al. (2020) distance determination used a larger number of stars, we adopt this distance for the IRAS 18162-2048 / GGD 27 region and its outflow in the analysis presented here.
The radio continuum jet driven by IRAS 18162-2048 is highly collimated (Martí et al. 1993).Several radio knots located along the jet axis imply a projected length of ∼10 parsecs (Masqué et al. 2012).The inner jet close to the source has the fastest speed of any known protostellar outflow.Radio proper motions indicate speeds up to ∼1200 km s −1 for the inner radio knots and ∼400 km s −1 for the outer knots when scaled to a distance of 1.4 kpc (Martí et al. 1995(Martí et al. , 1998;;Masqué et al. 2012Masqué et al. , 2015)).The jet exhibits both thermal and polarized non-thermal emission indicating active acceleration of relativistic particles (Carrasco-González et al. 2012;Rodríguez-Kamenetzky et al. 2017).Lowfrequency (325,610,and 1,300 MHz) non-thermal emission from the HH 80/81 radio knots was found by the Giant Meterwave Radio Telescope with a spectral index −0.7 (Vig et al. 2018).X-rays were detected from HH 80/81 by Pravdo et al. (2004Pravdo et al. ( , 2009)).This is the first outflow from a forming star in which hard x-rays with energies up to 5 keV are detected (López-Santiago et al. 2013).γ-rays have also been detected from a degree-scale region containing HH 80/81 and IRAS 18162-2048 extending to an energy of 1 GeV by the LAT detector on the Fermi gamma-ray observatory (Yan et al. 2022).Spitzer 8 and 24 µm images show a bi-conical infrared cavity (or a quasi-cylindrical cavity with a pinch) centered on the IRAS source and surrounding the radio jet (Qiu et al. 2008).
In single-dish observations of the inner 1.5 ′ region surrounding GGD 27, Qiu et al. (2019) report a two component, wide angle, low-velocity molecular outflow in CO with a spatial extent of only about 1 ′ -much smaller than the extent of the outflow traced by HH objects, molecular hydrogen objects (Mohan et al. 2023), and radio continuum emission.In these CO data, the line wings have Doppler shifts of less than 10 km s −1 with respect to the host cloud.This component of the outflow has a mass of about 2 to 3 M ⊙ .The redshifted CO lobe extends towards the south-southwest.The north-northeast CO lobe is very short with a low radial-velocity and wide opening angle.
High resolution ALMA 1.14 mm observations reveal a cluster of at least 25 compact continuum sources within a 22 ′′ field-of-view centered on IRAS 18162-2048 (Busquet et al. 2019).Sub Millimeter Array interferometric spectral line maps of the central 1 ′ diameter region show at least three monopolar outflow lobes emerging at angles that differ from the IRAS 18162-2048 radio jet (Qiu & Zhang 2009;Fernández-López et al. 2013).These outflows originate from the embedded YSO binary MM2 located about 6 ′′ northeast of IRAS 18162-2048 and possibly from a molecular core located about 2 ′′ north of MM2.The southeast facing jet-like flow from MM2 exhibits bullets with speeds up to 100 km s −1 and a total velocity extend of ∼190 km s −1 .The northwest and northeast facing lobes are only seen at low radial velocities.These observations suggest that IRAS 18162-2048 star forming clump contains several cores in addition to the cluster of young stellar objects.IRAS 18162-2048 (MM1) is the most luminous.Given the sizes of their outflows, MM1 is likely older than MM2 and the molecular core.Heathcote et al. (1998) presented Hubble Space Telescope (HST) images of HH 80/81 and a combination of lowand high-dispersion spectra.These authors used older ground-based images to measure proper motions.They found proper motions up to ∼800 km s −1 in some of the high-excitation knots in the HH 80/81 complex.Thus, this outflow exhibits the fastest known motions in an outflow from a forming star in both radio and visual-wavelength tracers.The high-dispersion spectra show that the HH 80/81 shock complex is redshifted with radial velocities ranging from -100 to over +600 km s −1 .Using the highest radial velocities, proper motions, and a fit to a bow-shock model, Heathcote et al. (1998) derive an inclination angle of the outflow axis with respect to the plane of the sky of about 56±5 • at a distance of 1.7 kpc used in their analysis.At a distance of 1.4 kpc this corresponds to 61±5 • .Using the shape of the disk in a well-resolved 1.14 mm image, Añez-López et al. (2020) derived an inclination angle of the disk axis of 49±5 • .
In this study, we combine the 1995 HST images with new HST observations taken in 2018 to investigate changes in the structure of the HH 80/81 shocks and to measure new proper motions.New ground-based images are used to identify shocks in the counterflow which extends north-northeast of GGD 27.These images also reveal a parsec-scale Hα bubble south-southwest of HH 80/81 beyond the projected edge of the L291 molecular cloud along the outflow axis.The bubble is likely powered by the fast jet driven by IRAS 18162-2048.

HST
HST was used to image HH 80/81 in 1995 using WFPC2 (Program ID 6128; PI Reipurth) and again in 2018 using WFC3 (Program ID 15353;PI Reipurth).In 1995, multiple exposures were obtained in the WFPC2 narrow-band

Apache Point Observatory
Hα images were obtained with the Apache Point Observatory (APO) 3.5 meter telescope between 14 May 2018 and 23 September 2022 using the 4096 by 4096 pixel ARCTIC CCD camera binned 3×3 to give an effective pixel-scale of 0.342 ′′ per pixel at the f/10.3focus of the APO telescope.During the earlier observing runs, we used a narrow-band filter with a 30 Å bandpass centered at 6570 Å (Hα) which excludes emission from the 6584 Å [Nii] emission line.The 30 Å filter is only 4 inches in diameter and caused significant vignetting.In 2021, 5 inch filters with pass-bands of 78 Å centered on the 5007 Å [Oiii], 76 Å centered at 6726 Å to cover the 6717+6731 Å [Sii] lines, and 78 Å centered at 6563 Å to cover the Hα line were acquired.This broader Hα filter also transmits the 6548 and 6584 Å [Nii] lines.The HH 80/81 outflow was re-observed in 2021 and 2022 with these filters.Exposure times, filters, and observation dates are listed in Table 2. Three to twenty frames were acquired at each exposure time and median combined to remove cosmic rays.Standard procedures were used for bias and dark current removal, and flat-fielding was done using twilight flats.
Narrow-band near-infrared images presented here were obtained using the APO 3.5 meter telescope with the NICFPS camera on the dates indicated in Table 2. NICFPS uses a 1024 × 1024 pixel Rockwell Hawaii1-RG HgCdTe detector.The pixel scale of this instrument is 0.273 ′′ per pixel with a field of view 4.85 ′ on each side.The narrow-band filters have band-passes of ∼0.4% of the central wavelength.Narrow-band filters centered off-line were used to obtain an off-line continuum frame to remove stars and reflection nebulosity.Images with 180 second exposures were obtained in the 2.122 µm S(1) line of H 2 .The central-wavelengths and band-passes are listed in Table 2. Separate off-source sky frames in each filter were interspersed with on-source images using the same exposure time at locations offset by at least 5 ′ .
During each observation, a set of 5 dithered images were obtained on both on-source and off-source positions.A median-combined set of unregistered, mode-subtracted sky frames were used to form a master sky-frame that was subtracted from each individual image.The reduced images were corrected for optical distortions.Field stars were used to align the frames, which were median-combined to produce the final images.Atmospheric seeing produced ∼0.9 to 1.5 ′′ FWHM stellar images.The observations are summarized in Table 2. Continuum subtracted images showing only H 2 emission were formed by subtracting the 2.13 µm images from the 2.12 µm images.

Overview
Figure 1 shows a mosaic of APO Hα images acquired using a narrow-band filter with a pass-band of 30 Å.In this figure, the two brightest shocks, HH 80 and 81, are saturated.The location and orientation of the fast radio jet core is shown by a blue oval.The location of the walls of an infrared cavity surrounding the radio jet and visible in Spitzer 8 and 24 µm images is shown by orange lines (Qiu et al. 2008).Towards the south-southwest of the IRAS source, a dim chain of compact knots and filaments extends from near the IRAS source towards the bright, high-excitation HH 80 and 81.Beyond HH 81, there is a faint but giant Hα bow shock, the northern portion of which was first noted by Heathcote et al. (1998).Towards the north-northeast where the blueshifted counter-flow to HH 80/81 is expected, the APO observations reveal a chain of dim Hα and [Sii] emission features extending up to 580 ′′ from IRAS 18162-2048.These objects lie close to the axis defined by the radio jet.Because the flux ratio I[Sii]/I(Hα) is larger than 0.5, and since they are located close to the axis of the outflow, we consider them to be Herbig-Haro objects.In Figure 1, dashed red circles show the locations of the new HH objects.H 2 emission is associated with some of these objects.The location and orientation of a prominent infrared dark cloud (IRDC) is marked by a magenta ellipse.This cloud, also known as HH 80N core, was extensively studied by Girart et al. (2001) and Masqué et al. (2011).The most distant radio feature thought to be associated with the IRAS 18162-2018 radio jet, the feature labeled 'radio source 34' in Figure 1 (Masqué et al. 2012), has no visual or near-IR counterpart.
We first discuss the changes in the structure of the HH 80 and 81 shocks in the fields observed during two epochs by HST.Then we present and analyze proper motions in these fields.This is followed by a discussion of the parsec-scale Hα bubble likely inflated by the fast (>800 km s −1 ) flow as it breaks out of the L291 cloud.After that, we discuss the faint shocks located in the counter-flow north-northwest of the IRAS source traced by narrow-band Hα, [Sii], [Fe ii], and H 2 emission.Finally, we discuss some ideas about the production of the ultra-fast jet and large proper motions in this extraordinary outflow.

The 2018 HST Observations; Changes Since 1995
The 2018 WFC3 images cover a larger field of view than the 1995 WFPC2 images with an angular resolution of ∼0.06 ′′ over the entire field (the wide-field chips in WFPC2 have an image scale of 0.1 ′′ per pixel).Of the three filters (Hβ, [Oiii], and [Sii]), the [Sii] images show the most extensive nebulosity.A nearly continuous chain of [Sii]-dominated shocks and filaments extends from the north-northeast corner of the field to the south-southwest corner along the axis of the radio jet.
Figures 2, 3, and 4 show the entire field of view imaged in 2018 in the [Sii], [Oiii] and Hβ filters.Two faint HH objects are located between HH 80/81 and IRAS 18162-2048.These shocks are marked as HH 81 N1 and HH 81 N2.HH 81 N1 is mostly detected in [Sii] and is thus a low-excitation shock.However, the brightest component, which looks like an arc-second diameter bow-shock pointing directly away from IRAS 18162+2048, contains several unresolved [Oiii] knots at its tip.Faint filaments of [Sii] emission connect the bright bow-shock in HH 81 N1 to the bright body of HH 81.
In the main HH objects, HH 80 and HH 81, the Hα, Hβ, and [Oiii] images show bright, compact knots and filaments close to the axis of the radio jet.The brightest features exhibit strong [Sii] emission.Additionally, the [Sii] image shows an extended network of filaments and diffuse features not seen in Hβ or [Oiii].A dim, diffuse region of [Sii] emission extends from HH 81 back towards HH 81 N2 in the upper-left corner of Figure 2. HH 81 N2 consists of a jumble of dim [Sii] knots and filaments over a 10 ′′ diameter region.
Parallel to the chain of bright HH objects extending from HH 80 N2 to HH 81, but about 30 ′′ to the west, there is a dim ∼10 ′′ wide filament of faint [Sii] emission.This feature is marked as 'HH 81 West rim' in Figure 2.After a 45 ′′ gap in this structure between Declination -20:50:30 and -20:51:00, it can be traced for about 130 ′′ to the south-southwest where it connects to the south-southwest edge of HH 80.   .The high-excitation emission traced by [Oiii] and, to some extent, Hβ, is confined to the core of the shock complex close to the extrapolated radio jet axis.In contrast, the lower-excitation [Sii] emission is much more widespread.
Figure 5 shows a close-up view of the HH 81 shock complex in [Sii] (red) and [Oiii] (cyan).Letter designations correspond to features marked in Heathcote et al. (1998).This color image shows a strong excitation gradient from west to east.The brightest and highest-excitation region occurs in feature A which resembled a letter 'X' in 2018 on the [Oiii] image.A series of three arcs of emission suggests shocks moving sideways with respect to the jet axis towards the southeast (features B and C).Both features A and B are [Oiii] dominated at their leading southern edges.The excitation gradient along filament C is reversed with the [Oiii] emission dominating its trailing northeastern edge.eastern wall of a cavity in the [Sii] emission.The western wall is marked by features C, D, E, I, J, and K (note that this is not 'HH 81 West rim' marked in Figure 3).

Proper Motions
Proper motions were measured by comparing the HST images taken in 1995 with the new set of images obtained in 2018.Given a proper motion PM (in milli-arcseconds per year) of a feature, the plane-of-the-sky speed is given by V PM (km s −1 ) = 4.74 D kpc PM(mas yr −1 ) where D kpc is in units of 1 kpc.Thus, at D kpc =1.4 kpc, V PM (km s −1 ) = 6.6366PM(mas yr −1 ).Changes and proper motions are most obvious when the images taken in 1995 and 2018 are blinked.Several methods of proper motion analysis and display were utilized: visual inspection and measurement of displacements, use of a Python-based cross-correlation code (Bally et al. 2022), difference images, and images in which the first epoch is shown in red and the second in cyan.Visual inspection suffers from personal bias.The complex structure of HH 80/81, the rich star field, especially in the [Sii] and Hα images, and the large magnitude of the motions over the 23 year interval between images have made the use of the Python code difficult.While intensity difference and ratio images show the motions well, they hide regions where there were no changes between the epochs.We found that color displays provide the best rendering of the complex motions in this shock system.Such color images are presented in the Appendix.
Figures 7 to 12 show the measured proper motions as vectors.Tables 3 to 5 lists the positions, proper motions, proper motion position angles, and the resulting speeds on the plane of the sky assuming a distance of 1.4 kpc.Table 3 presents [Oiii] proper motions, Table 4 presents [Sii] proper motions, and Table 5 presents motions measured on the Hα and Hβ images.In each table the first part presents motions in HH 80, the second part presents motions in HH 81.The numbering starts at 1 in each table section.The fastest motions are in excess of 1,200 km s −1 .
The cooling time of post shock plasma is τ cool ≈ 7000 n −1 H V 3.4 S,100 where n H is the hydrogen volume density (in cm −3 ), V S,100 is the shock speed in units of 100 km s −1 , and τ cool is the cooling time in years (Draine 2011).Thus, at a density n H = 1000 cm −3 , typical for bright HH objects, the cooling time is about 7 years for a shock speed of 100 km s −1 , less than the 23 year interval between the HST images.Thus, it is not surprising that the appearance of the shocks has changed considerably between 1995 and 2018.
A short cooling time of the post-shock plasma in the denser and brighter regions of the HH 80/81 shock system makes the measurement of some proper motions ambiguous.Inspection of the images shows that some features have faded or disappeared altogether while new features have appeared.Nevertheless, in all three species imaged with HST, there are many features whose motion away from GGD 27 is obvious when patterns and complex shapes are considered.
Unlike in some HH objects where the motions of persistent knots can be easily measured by cross-correlation methods, the complex structure, rich field of background stars, and noise in the images led to the failure of our automated technique.Comparison of the 1995 and 2018 images show the existence of some persistent patterns such as partial rings, arcs, and filaments among the jumble of structure.Prior knowledge of the general direction of motion helps in the identification of such features.Although in the future, artificial intelligence programs may be trainable to recognize such structure, here we use visual inspection to measure their apparent motion.In contrast, the very bright feature A (Figure 6) exhibits considerably slower, and more complex, behavior.As discussed above, feature A consists of a series of arcs and filaments which may trace ripples on the surface of a coneshaped bow shock with a high-excitation tip.While the smaller, leading arcs are seen best in [Oiii], the larger trailing arcs are best seen in the lower-excitation [Sii] image.Proper motions indicate that this structure is being compressed, with smaller motion at the leading edge than on the trailing side.This suggests that ejecta is running into a slower moving obstacle located ahead of the leading edge of A. Support for this hypothesis is provided by the presence of non-thermal radio emission and hard X-rays emerging from HH 80 (Pravdo et al. 2004(Pravdo et al. , 2009; López-Santiago et al. 2013; Vig et al. 2018).Analysis of Fermi/LAT data shows the presence of hard γ-rays with energies up to 1 GeV (Yan et al. 2022).However, the source can only be localized to about 1 degree.Thus, it is unclear where the γ-rays originate in the IRAS 18162-2048 outflow complex.
Because feature B has larger proper motions and is located ahead of feature A, we suggest that features A and B are located at slightly different distances along our line-of-sight.Feature A may be interacting with slower moving, or even stationary material along the walls of an outflow cavity drilled by a fast jet while feature B may be closer to the center of the flow channel.There is a relatively isolated arcsecond-scale blob, labeled F in Figure 6, about two-thirds of the way from feature A to the bright features G and H at the southern tip of HH 80.This is one of the few simple, blob-like features where the motion over 23 years is easy to see.  and 2018 (cyan).Figure 8 shows the vector field.Relatively high [Sii] proper motions are seen in the high-excitation core of HH 80 A (features 15 and 16 in Table 4 and in Figure 8), HH 80 knot F (feature 1), and feature 25 farther downstream near the bottom of Figure 8. Feature 25's motion is, however, uncertain; it may represent a case where a knot seen in 1995 had faded or disappeared and a new feature appeared in 2018.But since the apparent motion is along the jet axis at the expected extrapolated position of the radio jet, and close to the location of high proper motion features seen in the high-excitation species, it is included in the figure and associated table.
The most striking aspect of the [Sii] proper motions is the expanding arc of emission in the wake of HH 80-A (features 17 to 23 near the top-left in Figure 8).The ring's motion is not only orthogonal to the jet axis but, near the top, it exhibits motions back towards the IRAS source.This behavior can be understood when the large inclination of the jet axis is considered.Assume the ring is a ripple on the surface of a cone-shaped bow shock receding along the outflow axis.If the ring is expanding with respect to the outflow axis, the portion closest to the IRAS source can exhibit backward proper motion on the plane of the sky (apparent motion towards the source).Conversely, portions of the ring farthest from the source can exhibit proper motions higher than the tip of the bow.
Figures  10, the brightest emission consists of an 'X-shaped' emission region delineated by vectors 2 to 5. The proper motions here are much lower than in the relatively fainter diffuse features located south and east and marked by vectors 6 and 10 through 14.
This pattern is repeated in the Hβ image in Figure 12.Hydrogen recombination line proper motions were measured by comparing the 1995 Hα image with the 2018 Hβ image.Because of more than two magnitudes of visual extinction, features are much dimmer in the Hβ image.Thus, motions of only the brightest features were measured; these are listed in Table 3 and Figures 8 and 11.Overall, the hydrogen proper motions are similar to those seen in [Oiii].
Figure 11 shows the [Sii] proper motions in HH 81.The bright [Oiii] X-shaped feature is absent, and replaced by dim [Sii] knots.Bright [Sii] filaments appear in the wake of the bright [Oiii] X-shaped feature (towards the jet source).However, the proper motions in this filament are relatively low.
The [Sii] emission from HH 81 exhibits even more stunning expansion away from the jet axis.Table 3 lists motions measured in [Sii] at a set of representative locations in HH 81 and indicated in Figure 11.A color representation of the changes and motions between 1995 and 2018 similar to that shown for HH 80 is presented in the Appendix.

A Parsec-scale Bubble Blown by a Fast Protostellar Jet
Beyond HH 80, towards the expected south-southwest terminus of this giant outflow there is a faint but giant Hα bow shock, the northern portion of which was first noted by Heathcote et al. (1998).APO images obtained with an 80 Å filter in September 2022 show that the feature seen by Heathcote et al. (1998) is the northern side of a parsec-scale bubble located where the radio jet from IRAS 18162-2048 is expected to break out into the low density ISM. Figure 13 shows an image formed by subtracting a broad-band image taken with an SDSS i filter from the Hα image.The SDSS i image was scaled to provide the best average subtraction of stellar images in the Hα image.Stars which are relatively brighter in the Hα filter are black; stars which are relatively brighter in the longer wavelength SDSS i filter are white.
The Hα bubble is at least 7.7 ′ (3.1 pc) long with a width ranging from ∼5 ′ (2 pc) at the top of Figure 13 to less than 2 ′ (0.7 pc) at the bottom.Its axis of symmetry lies close to the extrapolated location of the fast radio jet and closely aligned with the direction of the proper motion vectors in HH 80/81.The full extent of the outflow, measured from radio source 34 in the north-northeast to the tip of the Hα bubble is about 1,500 ′′ which corresponds to a physical, projected end-to-end length of about 10.2 pc on the plane of the sky for a distance of 1.4 kpc.Since the redshifted, south-southwest lobe is inclined away from the Sun, this is a lower bound on the true length.A ∼60 • inclination with respect to the plane of the sky implies an astounding physical length of 20 pc.
The outer edge of the Hα bubble is sharp while the inner edge fades into the noise on an angular scale of a few to about 10 arcseconds.Thus, the Hα emission comes from a limb-brightened, filamentary structure.We estimate that the typical line-of-sight path length through the bubble, within an arcsecond or so of the outer edge is of order 0.07 pc.
Field stars with known SDSS r-band magnitudes were used to determine a photometric zero-point for the stacked Hα image of the bubble.This zero-point was used to measure the observed Hα surface brightness.The peak surface brightness is about SB(Hα) ≈ 2 × 10 −16 erg s −1 cm −2 arcsec −2 .The noise level is about 1.5 to 2 × 10 −17 Note-Velocities assume a distance of 1.4 kpc erg s −1 cm −2 arcsec −2 , dominated by diffuse Hα emission and airglow.Assuming a foreground extinction of about A V = 1.4 magnitudes (note that this is lower than the extinction towards HH 80 and HH 81 because the bubble is farther away from the L291 cloud.We here assume a mean extinction of A V =1 mag/kpc).If the extinction at the wavelength of Hα is 0.7808 A V , the extinction corrected peak surface brightness is SB cor (Hα) ≈ 5.5 × 10 −16 erg s −1 cm −2 arcsec −2 .The dimmest parts of the bubble have SB(Hα) ≈ 6 × 10 −17 erg s −1 cm −2 arcsec −2 , which when corrected for extinction corresponds to SB cor (Hα) ≈ 1.6 × 10 −16 erg s −1 cm −2 arcsec −2 .The emission measure is related to the Hα surface brightness by EM(cm −6 pc) = 4.86 × 10 17 I Hα erg s −1 cm −2 arcsec −2 (see Haffner et al. 1998).The emission measure, EM, of the limb-brightened bubble ranges from below 80 cm −6 pc to about 300 cm −6 pc.The emission measure can be related to the mean electron density in the emission region by EM = n 2 e dL pc where n e is the mean electron density and L pc is the line-of-sight path-length through the emission region in units of a parsec.Thus, n e ≈ (EM/L pc ) 1/2 .For an assumed path length L pc = 10 ′′ (∼0.07 pc), the electron densities range from about 30 to 200 cm −3 .A very crude, order-of-magnitude shell mass can be estimated by assuming that the mean shell thickness is about 1 ′′ , that its mean electron density is 100 Note-Velocities assume a distance of 1.4 kpc cm −3 , and that it is a cylinder with a projected surface area of about 4.5 square parsecs (diameter ∼1.5 pc; length 3 pc).This gives about 0.1 M ⊙ as a likely lower bound.If the interior surface brightness is at the noise level, and the LOS depth is 1.5 pc, the emission measure would be 35 cm −6 pc and the mean electron density about 5 cm −3 .The total mass would then be about 1 Mo.This is likely an upper bound on the mass of bubble wall and interior.Thus the mass is likely to be between 0.1 and 1.0 M ⊙ .As discussed below, the interior of the Hα bubble is likely filled with hot plasma whose cooling time may be longer than the dynamical age of the outflow.

The Counterflow to HH 80/81
An unusual aspect of the outflow from IRAS 18162-2048 is that the brightest HH objects, HH 80 and HH 81, are highly redshifted.These shocks are likely visible because they are seen toward the low-obscuration region beyond the western edge of the L291 cloud.Since HH 80/81 are behind the driving source, the counterflow is expected to be located in the foreground where one might naively expect lower extinction.Yet, no comparably bright Herbig-Haro (HH) objects or molecular hydrogen objects (MHOs) exist north-northeast of the IRAS source.
Our images reveal a faint chain of HH objects and MHOs in the expected counterflow direction.The locations of these features are marked in Figure 1.Their positions are listed in Table 4. Figure 14 shows an Hα image of the sub-field located north-northwest of IRAS 18162-2048.Many of these objects are also visible in the 2.122 µm H 2 S(1) transition.
The first object suspected to be a shock in the counterflow was a relatively bright radio frequency emission knot, designated HH 80N by Martí et al. (1993).HH 80N is located about one arcminute south of the IRDC marked in Figure 1 and was thought to mark the site where the radio jet slammed into the outer parts of this cloud.Molinari et al. (2001) presented far-IR spectra of 63 µm and 145 µm [Oi] and 157 µm [C + ] emission detected by the ISO satellite to show that this visually obscured region is shock-excited and thus likely to be a Herbig-Haro object.Molinari et al. (2001) detected extended [C + ] emission along the entire length of the radio jet.They argue that the radio jet is times larger than the Hα surface brightness, indicating excitation in a shock.The ∼2 ′ offset from the jet axis raises the possibility that it is powered by a source other than IRAS 18162-2048.However, given the giant bubble located south-southwest of the IRAS source, it is possible that HH 80N G also traces a mostly obscured giant bubble wall located in the counterflow.It may be seen through a particularly translucent part of the L291 cloud.We searched the vicinity of radio source #34 for both HH objects and MHOs, but none were detected at our sensitivity limit.Although this portion of the IRAS 18162-2048 flow is expected to be blueshifted and approaching us, it is nevertheless still highly obscured judging from the relatively low-density of stars in our image.This implies that the L291 cloud, if it is a sheet, is even more inclined to our line-of-sight than the giant outflow from IRAS 18162-2048.
Near-IR imaging shows the presence of faint and extended 2.12 µm molecular hydrogen emission between IRAS 18162-2048 and HH 80N. Figure 15 shows a continuum subtracted H 2 image of the field from 2 ′ south of IRAS 18162-2048 to HH 80N A (previously HH 80N). Figure 16 shows a closeup view of the immediate vicinity of IRAS 18162-2048 in a continuum subtracted H 2 image.Figure 17 shows the core region in a continuum subtracted [Fe ii] image.Figure 18 shows the field north of IRAS 18162-2048 containing HH 80N A (HH80N) in a continuum subtracted H 2 image with the various MHOs marked.The continuum subtraction removes most of the bright reflection nebulosity.Thus, these images primarily show pure 2.12 µm H 2 emission in the v=0-0 S(1) line.
In Figure 16, the location of IRAS 18162-2048 is indicated by a cyan circle.The bubble on the upper right traces a cavity created by a moderate mass star at ICRS=18:19:10.457, -20:46:58 also known as GGD 27 IRS4, which is thought to be a Herbig AeBe star with spectral type B2.There is a faint, jet-like feature extending from near the IRAS source through the cavity.The orientation and location is indicated in Figure 16 by a dashed red arrow.This feature lines-up with the 'NW' redshifted CO outflow found by Fernández-López et al. (2013).
The U-shaped cavity in the counterflow opening towards the northeast traces the inner walls of the GGD 27 reflection nebula where the radio jet and surrounding outflow has created a cavity in the clump hosting IRAS 18162-2048.A knot of H 2 emission, designated knot 3 in Mohan et al. (2023), in the middle of the U-shaped cavity lies along the radio jet axis and may mark a shock in the outflow from the IRAS source.
The H 2 images (Figures 15 and 16) show a 45 ′′ long chain of knots extending to the east-southeast of the IRAS source and terminating in a bow shock.This flow may be the H 2 counterpart of the blueshifted 'SE' CO outflow emerging at position angle PA = 126 • detected by Fernández-López et al. (2013).Mohan et al. (2023) give these features the designations MHO 2355 and MHO 2357.MHO 2355 consists of a wiggly chain of five knots.MHO 2357 is the bright H 2 bow shock.MHO 2360 may trace an H 2 knot in the counterflow to MHO 2355 / MHO 2357.Two H 2 knots, marked by small, dashed circles in Figure 15 mark the locations of H 2 knots along the redshifted 'NE' CO outflow lobe found by Fernández-López et al. (2013).Mohan et al. (2023) give these features the designation MHO 2358.A pair of knots designated MHO 2356 are located between these two flows indicating yet another outflow.
Figure 16 shows two wedge-shaped H 2 features close to the radio jet axis, marked by the larger, red dashed circles.These two objects are symmetrically placed about IRAS 18162-2048 defining a line close to the radio jet axis.But the northern H 2 wedge is slightly east of the radio jet axis while the southern wedge is offset slightly west of the radio jet axis.These features are designated MHO 2354-N and MHO 2354-S.MHO 2354-N, designated knot 3 in Mohan et al. (2023), coincides with the northeastern rim of the mid-IR cavity seen in Spitzer 8 and 24 µm images.
Faint 1.644 µm [Fe ii] emission is seen along the jet axis around the location of the H 2 wedges (Figure 17). Figure 16 shows a continuum-subtracted zoom-in view of the region around the IRAS source along with the several dozen YSOs detected by ALMA (Busquet et al. 2019).Along the northern direction, the [Fe ii] emission resembles an elongated bubble.Only dim [Fe ii] emission is seen near the southern H 2 wedge in the form of a pair of parallel streaks.The relative dimness of the southern [Fe ii] feature is consistent with more extinction the redshift of the southern outflow lobe and blueshift of the northern lobe,

DISCUSSION
With speeds in excess of 1,000 km s −1 , the HH 80/81 radio jet exhibits the fastest known proper motions of any outflow from a young stellar object.The high-excitation Herbig-Haro objects are mechanically illuminated by this jet.Our proper motion measurements imply that such fast motions persist at a projected distance ∼300 ′′ (2 pc) from the source, the massive young stellar object (MYSO) IRAS 18162-2048.
Rapidly accreting MYSOs with masses between ∼ 10 M ⊙ to ∼20 M ⊙ develop bloated and cool photospheres which prevent them from emitting hydrogen-ionizing, extreme ultraviolet (EUV) radiation (Hosokawa & Omukai 2009).However, for accretion rates below 10 −3 M ⊙ year −1 , as they approach ∼20 M ⊙ , their photospheric radii and temperatures approach zero-ago main sequence (ZAMS) values.They start to ionize their surroundings.The disk rotation curve suggests that IRAS 18162-2048 may have a mass of ∼20 M ⊙ and may be approaching the stage where it emits EUV at a rate expected for a ZAMS star.
If IRAS 18162-2048 is on the ZAMS, its massive disk may trap the Lyman continuum emitted by the star, preventing the growth of an Hii region (Hollenbach et al. 1994).For a stellar mass, M star =20 M ⊙ , and a sound speed in photoionized plasma, c s =10 km s −1 , the gravitational radius is r G = GM star /c 2 s =178 AU .Photo-ionized plasma will be bound by the gravity of the star as long as the ionization front is closer to the star than r G .
The observed peak radio continuum emission at 1.4, 5, and 15 GHz from IRAS 18162-2048 is about S ν = 3.8 ν +0.2±0.1 mJy in beam-matched observations where ν is in units of 1 GHz (Martí et al. 1993).Most of this emission likely originates from optically thin free-free emission from the jet and thus places an upper-bound on the flux from a hyper-compact Hii region.ALMA 1.14 mm (263 GHz) observations at the longest baselines indicate the presence of a compact source less than ∼40 mas (56 AU) in diameter with a flux of 19 mJy, interpreted by Añez-López et al. (2020) to trace ionized gas from a hyper-compact Hii region.Extrapolating the above formula for the flux density from the centimeter regime to 1.14 mm implies a peak flux to be in the range 6.6 to 20.2 mJy with a most likely value of 11.6 mJy.Thus, in addition to the emission from the thermal plasma in the jet, there may be a hyper-compact Hii region surrounding this MYSO which may be optically thick at the centimeter wavelengths.If ∼12 to 19 mJy flux originates from a hyper-compact Hii region which is optically thick at 263 GHz, its radius would be about 8 to 10 AU.Such a compact Hii region would be bound by the gravity of the star and disk.
A 20 M ⊙ ZAMS star has a Lyman continuum luminosity of about 3 × 10 48 ionizing photons per second.A hypercompact Hii region with a spherical radius of 8 to 10 AU, uniformly filled with plasma in photo-ionization equilibrium with this ionizing luminosity will have an electron density of about n e ∼ 10 9 cm −3 .The ALMA 1.14 mm image constrains the hyper-compact Hii region to be smaller than 56 AU diameter.For a radius of ∼56 AU, n e ∼ 7×10 7 cm −3 .
If IRAS 18162-2048 grew via highly variable, episodic accretion, the growing protostar may have experienced periods during which it was accreting at low rates or not accreting at all (Galván-Madrid et al. 2008) .Such periods of quiescence may have allowed the photosphere to collapse and heat-up.The MYSO may have settled onto the ZAMS even at a mass well-below 20 M ⊙ .

The Jet and Disk Axis Inclination Angle Revisited
The inclination angle of the jet or the disk axis is given by i = arctan(V r /V PM ) where V r is the radial velocity and V P M is the proper motion of a feature.Here, i = 90 corresponds to the disk being face-on and the jet axis pointing directly at or away from the Sun.Heathcote et al. (1998) used proper motion and radial velocity measurements to determine the inclination angle of HH 80 and 81.Their spectra found a maximum radial velocity of V r =+600 km s −1 and a mean proper motion of HH 80 and 81 of ∼350 km s −1 .Their estimate of the inclination angle resulted in a value of i=60 • , corrected to a distance of 1.4 kpc.
The high-angular resolution measurements presented here show that the proper motion vector field is complex.Thus, it is unclear if a mean value of the proper motions is an appropriate estimator of the inclination angle of the HH object motions.Using a maximum proper motion, V P M =1,200 km s −1 and the maximum radial velocity of 600 km s −1 from Heathcote et al. (1998) gives i ≈27 • .However, these maximum values were determined for different features in the shock complex.
The high angular resolution ALMA 1.14 mm images of the IRAS 18162-2048 disk presented by Girart et al. (2017Girart et al. ( , 2018) ) and Añez-López et al. ( 2020) allow a measurement of the orientation of the disk axis.The analysis of Añez-López et al. (2020) gives an inclination for the disk axis of i disk =49±5 • .For the remainder of the analysis, we assume that the inclination angle is likely to be between 44 • and 65 • with the south-southwestern lobe of the outflow containing the bright HH objects HH 80 and 81 receding away from the Sun.The fastest proper motions in the jet and HH objects, ∼1,200 km s −1 , imply speeds ranging from 1,670 to 2,400 km s −1 depending on the actual value of the inclination angle.

A Collimated, Fast, Line-driven Wind?
Given the ∼44 • to 65 • inclination angle of the outflow axis from IRAS 18162-2048, the proper motions imply speeds similar to the line-driven stellar winds powered by main sequence O stars, about 2,000 km s −1 (Vink 2022).If IRAS 18162-2048 is a ZAMS O star, it must be so highly embedded that no visual or UV light escapes towards our line-of-sight.However, a fast, line-driven wind could form near the stellar photosphere.
An isotropic, fast wind from the central star may be collimated into a jet by a combination of density gradients and strong magnetic fields.Such a field, anchored to a massive, differentially rotating disk is expected to develop magnetic hoop stress.A lower bound on the strength of a collimating magnetic field can be obtained by setting the magnetic pressure equal to the ram pressure of the wind.For a wind speed of 2,000 km s −1 and mass loss-rate of Ṁ = 10 −5 M ⊙ yr −1 , collimation at distances of 10 to 100 AU from the star requires a magnetic field of order 4 to 0.4 gauss, respectively.
A 20 M ⊙ ZAMS star has a luminosity of about 10 5 L ⊙ , nearly an order of magnitude larger than the measured luminosity of IRAS 18162-2048 assuming an isotropic radiation field.In this scenario, a disk constrained Hii region and jet cavity may allow most of the luminosity to be beamed along the jet axis.This 'flashlight effect' may explain the low measured luminosity of IRAS 18162-2048 from our vantage point (Kuiper et al. 2015).
One problem is that the mass-loss rates from late O-type main sequence stars are typically 10 −8 to 10 −7 M ⊙ year −1 , two to three orders of magnitude lower than the mass loss-rate estimated from the H 2 emission by Mohan et al. (2023) or from the empirical correlation between stellar luminosity and mass-loss rate (Maud et al. 2015).As discussed below, a strong magnetic field anchored to the massive disk may launch a magneto-centrifugal-wind on its own, adding to the mass-loss rate of the system if there is a line-driven component.

A Hybrid, Ionization Initiated, Magneto-Centrifugal Wind?
The massive, ∼5 M ⊙ disk distinguishes IRAS 18162-2048 from most other MYSOs.Continued accretion through a disk requires that the rate at which it flows through the disk toward the MYSO be larger than the rate at which the disk surface layers are photo-ablated by Lyman continuum from the star.Hollenbach et al. (1994) investigated mass-loss from a disk due to photo-ablation by a massive central star.The EUV field can drive a slow wind from the disk surface at radii larger than the gravitational radius, r G .For a ∼20 M ⊙ star, photo-ablation-driven mass loss rates tend be around ∼ 10 −5 M ⊙ year −1 close to what is estimated for the IRAS 18162-2048 outflow.However, the absence of an extended Hii region on scales larger than r G rules out this type of model.Kuiper & Hosokawa (2018) presented models of massive star formation with radiative feedback, including photoionization but they ignored the role of magnetic fields.If the circumstellar disk is threaded by a magnetic field, differential rotation-induced shear can lead to field amplification.Within the disk, or partially ionized disk surface layers, the field is expected to have a toroidal geometry wrapped-up by shear.The expected geometry of open fieldlines above and below the disk is a helical, hourglass-shaped field pinched at its waist by the disk.Oliva & Kuiper (2023a,b) present models with magnetic fields.But, this work ignores fields that might link the circumstellar disk to the forming massive star.Tanaka et al. (2016Tanaka et al. ( , 2017) ) presented models of MYSOs with strong outflows, showing that as they approach the ZAMS, their outflows are the first components to become ionized.
Continued accretion from a magnetized disk onto the MYSO is expected to drag field lines from the disk onto the stellar photosphere, creating closed magnetic loops.These loops will funnel accreting matter onto high latitude regions of the star, in 'funnel flows', commonly thought to occur in low-mass star formation.Accretion mediated by funnel flows are thought to regulate the spin-rates of the forming stars.The kilo-gauss magnetic fields present on some young ZAMS O-type stars such as θ 1 Ori C (Donati et al. 2002) may be fossil remnants of such magnetic structures.
Magnetically confined funnel flows will become ionized by EUV radiation if the MYSO is close to the ZAMS.But, sufficiently dense accretion columns are likely to be optically thick to EUV and may contain neutral cores.In this case, only the surface layers will be ionized.As long as accretion columns don't fully cover the projected surface area of the MYSO, some, or most EUV will escape the inner magnetospheric closed loops to irradiate the disk surface penetrated by open field lines.
Photo-ionization of the disk surface may provide an efficient mechanisms for mass-loading the open magnetic fieldlines.Charged particles injected onto open field lines can be accelerated, forming a magneto-centrifugal wind (Pudritz & Ray 2019).Such winds are thought to play important roles in driving jets and outflows from forming stars.We consider a scenario in which a strongly magnetized disk can support continued accretion onto the MYSO and at the same time drive a magneto-centrifugal wind launched by photo-ionization of the disk surface which loads plasma onto the open field lines.
The electron density and radius of the hyper-compact Hii region implied by the 1.14 mm ALMA data is n e ≈ 10 8 to 10 9 cm −3 at R ii ≈ 60 to 10 AU.If the photo-ablating plasma is lofted off the disk surface as a quasi-spherical wind with a velocity V w the implied mass-loss rate is Ṁii = 2πR ii (3πQ/α B ) 1/2 .In a frame co-rotating with a portion of the disk, V w will likely have a value about 1.5 to 3 times the sound speed in the photo-ionized plasma, c s ≈10 km s −1 because of the acceleration by the large vertical density and pressure gradients.For V w =20 km s −1 , Ṁii ∼ ≈ 9 × 10 −6 to 1.6 × 10 −5 M ⊙ year −1 for R ii =10 to 30 AU, close to the estimated mass-loss rate in the radio jet.
In the rest frame of the star, the plasma loaded onto open field lines will inherit the roughly Keplerian orbit speed at the footprint of the field.For a ∼20 M ⊙ star the orbit speed is ∼420 km s −1 at 0.1 AU, 134 km s −1 at 1 AU, and 25 km s −1 at 30 AU. Plasma loaded onto the field-line at 0.1 AU will reach a a speed of 2,000 km s −1 at a radial distance of 0.476 AU from the axis of the disk.Plasma loaded onto a rigid, open field-line at 1 AU, will reach 2,000 km s −1 at a radial distance of ∼15 AU from the disk axis.But, plasma launched at 25 AU would require a magnetic lever arm of nearly 80 to reach this speed at R∼2,000 AU.
The plasma loaded onto open field-lines is likely to flow away from the disk on a nearly radial trajectory as seen from the star.The hoop-stress of the azimuthal component of the field may collimate this wide-angle flow into a jet at vertical distances of 10s to hundreds (for the launch radii of 0.1 to 1 AU ) to thousands of AU (for the outer launch radii near 30 AU).In this magneto-centrifugal launch scenario, it is likely that the jet is layered with a fast core surrounded by a sheath of slower ejecta reflecting a larger radius of the jet launch point and a slower orbit speed.Only a small fraction of the jet mass-loss-rate is required to have speeds in excess of 1,000 km s −1 to produce the observed ultra-fast motions in the radio jet and HH objects.In a layered jet model, only the jet core is required to have such fast motions.The flow speed may decrease with increasing distance from the jet axis.
A rough estimate of the field strength required to produce the outflow can be obtained by equating the kinetic energy density of the plasma with the magnetic energy density where the magnetic lever-arm causes the photoablation flow from the disk surface to reach the observed jet velocity.Given a mass-loss (or mass-loading) rate Ṁw , the jet velocity, V j , and radius where the magneto-centrifugal wind reaches the jet velocity, R j , the mean density of the flow is ρ(R j ) ≈ Ṁ /(πR 2 j V j ).The mean kinetic energy density is then 0.5ρV 2 j .Setting this equal to j,2000 R −1 j,30 gauss where Ṁ−5 is in units of 10 −5 M ⊙ yr −1 , R j,30 is in units of 30 AU, and V j,2000 is in units of 2000 km s −1 .Thus, a field of a few gauss at 30 AU and less than a kilo-gauss at 0.1 AU is sufficient to accelerate a photo-ablation driven, low-velocity wind from the disk surface to the ∼2000 km s −1 flow observed in the IRAS 18162-2048 jet by means of the magneto-centrifugal mechanism.4.4.Impact of a Fast, > 10 3 km s −1 Jet on the Surrounding Medium The visibility of these fast motions in tracers such as [Sii] and even [Oiii] is surprising in light of these extraordinary speeds.Shocks with speeds in excess of 1,000 km s −1 would lead to the production of considerably higher ionization states.The presence of relatively low ionization states in the fastest moving knots suggests that the observed species are excited by shocks formed where fast ejecta overtake slightly slower, but nonetheless fast moving debris.Alternatively, if the shocks are formed by the collision of fast ejecta with stationary or slow-moving material, the low-ionization states trace forward shocks propagating into a much denser medium than the fast flow.Magnetic fields may also cushion shocks, enabling lower-ionization states to survive and be excited into emission.
The HH 80/81 shocks are located beyond the western edge of the L291 cloud and are thus likely in a relatively low-density environment.The rich, background star-field at visual wavelengths implies our line-of-sight can penetrate several kpc without more than a few magnitudes of extinction.If the inter-cloud medium were uniform, this would imply a density n(H) < 1 cm −3 .But the region just outside obvious obscuration may be somewhat denser, but likely to have a density less than ∼10 to 100 cm −3 .The electron density of the post-shock plasma in the HH 80/81 shocks has been measured to be between n e = 10 3 to 10 4 cm −3 using the [Sii] doublet ratio (Heathcote et al. 1998).Thus, the second model above (dense clumps moving supersonically into a lower density medium) is likely correct.
As discussed in Section 3.3, the cooling time of post shock plasma is τ cool ≈ 7000 n −1 H V 3.4 S,100 where n H is the hydrogen volume density and V S,100 is the shock speed in units of 100 km s −1 (Draine 2011).For n H = 100 cm −3 , likely an upper bound on the density of gas into which fast ejecta from IRAS 18162-2048 is running, τ cool ≈ 1.8 × 10 5 years, sufficiently long to enable the formation of the Hα bubble by fast shocks such as indicated by the large proper motions.The post-shock, low-density plasma will likely expand adiabatically into the surrounding medium.
The temperature immediately behind a 1,000 km s −1 shock is T ps = 3µm H V 2 S /16k ≈ 1.38 × 10 5 V 2 S,100 Kelvin for µ = 0.609 appropriate for fully ionized plasma (Draine 2011).Thus, behind a 1,000 km s −1 shock, the plasma reaches a temperature of about 10 7 Kelvin, sufficiently hot to explain the observed X-ray emission from parts of HH 80 and 81.If driven by the impact of a jet or dense clumps of ejecta interacting with a lower density medium, the hot plasma will expand adiabatically to power an energy-conserving bubble, similar to the early phases of supernova remnant expansion.The expanding bubble sweeps up a shell from the surrounding ISM.The shell can be ionized by fast shocks, by EUV escaping along the jet axis from IRAS 18162-2048, EUV radiation produced by the recombining and cooling X-ray plasma, or the ambient Lyman continuum radiation field pervading the exterior of the L291 cloud.All four mechanisms may contribute to the visibility of the giant Hα bubble.
It is remarkable that no extensive molecular outflow exists around this fast radio jet and associated HH objects.One possible interpretation is that molecules that were swept-up by the younger outflow produced by the IRAS source as it was accreting was subsequently completely dissociated.As discussed above, inside the parent cloud dissociation could have been caused by strong shocks, UV radiation emitted by the MYSO, or by radiation emitted by shocks powered by the ultra-fast jet.The nearly 1,000 km s −1 velocity variation shown by our proper motions shows that post-shock plasmas can reach temperatures in excess of 10 7 Kelvin, a result confirmed by X-ray detection of the brightest HH objects in this outflow (Pravdo et al. 2009).The detection of a parsec-scale bubble provides further evidence that the outflow cavity produced by this outflow is likely filled with low-density EUV-ionized and soft-X-ray plasma, with an emission measure too low to be detected by current methods.The radiation produced as plasma recombines into neutral hydrogen may be capable of dissociating a pre-existing molecular outflow produced during earlier evolutionary phases of IRAS 18162-2048.
The IRAS 18162-2048 outflow may be the best example of the extreme feedback impacts of forming massive stars on their birth environment.As MYSOs grow from under 1 M ⊙ Solar mass to over 10 M ⊙ , they likely create bipolar molecular flows as fast disk winds and jets sweep-up ambient material.Their mechanical power increases with source luminosity and mass (Maud et al. 2015).The size of the entrained molecular outflow lobes will be limited to the size of their parent molecular clump or host cloud.When outflows punch out of their host molecular clouds, the entrained gas may be predominantly atomic or even ionized as is the case with the parsec-scale HH 80/81 flow.
If the shock-ionized plasma produces only one ionizing photon as it recombines, the minimum ionizing photon luminosity of the shock will be given by the rate at which the medium is swept up.The cross-sectional area of the shock, πR 2 s , times the density n of the medium into which it is running, times the shock speed V s gives a minimum on ionizing luminosity.For a shock with a radius R s = 10 17 (5 ′′ at a distance of 1.4 kpc), and speed V s = 10 3 km s −1 , running into a medium with a hydrogen density n = 10 2 cm −3 , we get Q min = πR 2 s nV ∼ 3 × 10 44 photons s −1 .With a post-shock temperature of order 1 to 10 MK, much of this radiation will emerge as X-rays and EUV radiation.Reprocessing of X-ray photons into softer EUV photons by the surrounding medium can increase the Lyman continuum luminosity of the shock by more than an order of magnitude over the above rough estimate.Additionally, multiple shocks will increase the UV luminosity.Finally, fast shocks propagating into a previously swept-up molecular shell may directly contribute to the dissociation of bipolar molecular outflow lobes.
The 2018 [Sii] and 1995 Hα images show that a cylindrical region with a radius of order 10 18 cm and a length of more than a parsec contains shocks emitting in these species.Unfortunately, because of the failure of the 2018 Hα images, the proper motions of fainter features away from the jet axis could only be measured in [Sii].The proper motions in [Sii] show expansion away from the jet axis with speeds of between 100 to 300 km s −1 .As discussed above, it is likely that the Hα and [Sii] emission arises in reverse shocks to keep much of the sulfur in its first ionization stage.The forward shocks must then have higher excitation and likely dissociate any molecules they encounter, including those associated with any previously swept-up bipolar outflow.Using the above formula, the Lyman continuum luminosity of the [Sii] / Hα emission region is likely to be larger than 10 46 s −1 due to its larger area.
Although we do not have any direct measurements of the total outflow mass, a crude estimate of its energetics is possible based on the empirical relations found for other MYSO outflows (Maud et al. 2015).It is likely that the fast speeds found here are what mostly distinguishes this flow from other MYSO outflows.If we assume that over its formation IRAS 18162-2048 ejected a mass of 0.1 M ⊙ in a fast, 10 3 km s −1 jet, the kinetic energy of this component would be E ∼ 10 48 ergs.This is comparable to the energy required to dissociate about 50 to 100 M ⊙ of H 2 by fast shocks and their UV radiation.Thus, this fast outflow could destroy its own fossil molecular outflow.If this scenario is correct, then most of the outflow mass should be atomic or ionized.Future sensitive 21 cm HI , 158 µm C + , or 63 µm Oi observations may be used to measure its mass.
There has been some discussion of outflow-triggered star-formation in the literature with the case of HH 80N being an important potential example of this process (Molinari et al. 2001;Girart et al. 2001;Masqué et al. 2011).However, it remains unclear if the clump (IRDC) ahead of HH 80N A was already forming stars before the IRAS 18162-2048 outflow impacted its environment.Although shocks may alter the chemistry at the cloud surface, it is unclear if they exerted any significant dynamical influence.

Comparison with Other Nearby MYSOs
Other nearby (less than 1.4 kpc) massive-star forming complexes producing 15 to 20 MYSOs provide interesting comparisons to the IRAS 18162-2048 radio jet, MHO, and HH outflow complex.We briefly comment on the Orion OMC1, Cepheus A, and Sh2-106 regions.
The BN/KL outflow complex from Orion OMC1 located ∼0.1 pc behind the Orion nebula contains a ∼15 M ⊙ MYSO, radio source I (Ginsburg et al. 2018;Wright et al. 2023), the ∼10 M ⊙ Becklin-Neugeabauer (BN) object, and the ∼3 M ⊙ source x.These three protostars were ejected by a dynamic interaction about ∼550 years ago with speeds of ∼10, 30, and 55 km s −1 , respectively.This event was associated with a ∼ 10 48 erg explosion which launched about 10 M ⊙ of molecular gas.The explosion created hundreds of molecular streamers seen in CO, many of which are associated with shock-excited fingers of molecular hydrogen.The fastest proper motions in the fingertips are in excess of 400 km s −1 .For a recent discussion of this outflow and associated runaway stars, see Bally et al. (2020).Source I (Src I) drives a very young outflow (dynamic age <300 years) which may be powered by an ionized jet launched along the axis of the circumstellar disk which survived the dynamic interaction (Wright et al. 2023).The speed of this jet is not yet measured.
The ∼15 M ⊙ MYSO, HW2 in Cepheus A, drives a radio jet exhibiting proper motions of ∼500 km s −1 (Carrasco-González et al. 2021).Cepheus A contains a spectacular shock-excited molecular hydrogen and CO outflow complex.The multiple chains of MHOs and HH objects originate from the vicinity of HW2 and have been interpreted in terms of a pulsed, precessing jet launched by this massive protostar (Cunningham et al. 2009).
Sh2-106 contains about 3 M ⊙ of plasma, indicating that it is a well developed Hii region.Sh2-106 is ionized by an embedded O9 star, S106IR, which is obscured by over A V ≈20 magnitudes, and may be surrounded by a circumstellar disk (Bally et al. 2022).Instead of a fast wind, S106 IR drives a relatively slow, ∼100 to 400 km s −1 wind and there is no evidence of a jet.The estimated masses of S106 IR and IRAS 18162-2048 are ∼15 and ∼20 M ⊙ , similar to within a factor of two.A key difference between these two stars may be the masses of their circumstellar disks and their evolutionary stages.The S106 IR disk has a mass less than 0.8 M ⊙ while the disk surrounding IRAS 18162-2048 may have a mass of ∼5 M ⊙ .Additionally, the accretion histories of these MYSOs may have been different.

CONCLUSIONS
A comparison of narrow-band images obtained with the Hubble Space Telescope taken in 1995 and 2018 in [Oiii], Hα and Hβ, and [Sii] reveals significant and complex changes in the shock morphologies of HH 80 and HH 81.Some features have disappeared and others have appeared.Where persistent patterns can be recognized in the two epochs, proper motions were measured.Proper motions ranging up to 1,200 km s −1 are found.However, there is a huge dispersion in the measured values.
As a rule, the fastest motions, up to 1,200 km s −1 are seen close to the extrapolated radio jet axis in relatively dim features.The brightest shocks tend to exhibit slower motions in the range of 200 to 400 km s −1 .We interpret this as evidence that the bright shocks are produced where fast flows are impacting slower or stationary obstacles.The images reveal expanding arcs or partial rings around these regions.In these rings, some proper motions are backwards toward the source.We use published spectra to infer the radial velocities of the brightest shocks.The ratio of radial velocity to proper motions of these shocks are used to re-derive the inclination angle of the flow with respect to the plane of the sky.This angle is found to be about 44 to 65 • , consistent with previous measurements.The backwards and sideways motions are thus consistent with the splash of post-shock gas as it moves around a slower obstacle.
Away from the jet axis, [Sii] emission exhibits extensive but slow proper motions orthogonal to the jet, indicating that the outflow is creating an expanding bubble.As this ultra-fast flow bursts out from behind the L291 cloud into the surrounding, lower density ISM, it has inflated a parsec-scale bubble seen in new images as a network of Hα filaments.
We identify a chain of faint HH objects and MHOs in the expected counterflow located on the opposite side of IRAS 18162-2048 to HH 80/81.Dim Hα and [Sii] trace these HH objects.More extensive but dim 2.12 µm H 2 emission is seen between the GGD 27 reflection nebula and star-forming clump located ahead of HH 80N A. Given the 10 parsec projected length of this outflow as traced by radio continuum, HH objects, and MHOs, it is remarkable that only a tiny molecular outflow is seen in the immediate vicinity of the IRAS source.The outflow may have burst out of the host cloud core and is primarily interacting with mostly atomic or ionized inter-cloud or inter-clump gas.Slower moving debris may represent left-over fragments of a bipolar outflow produced when IRAS 18162-2048 was much less massive and drove a less powerful molecular outflow.Shocks could have directly destroyed molecules in such flows.Additionally UV radiation fields produced by the ultra-fast shocks in this outflow may have contributed to the dissociation of molecules.
We briefly discuss possible models for the ultra-fast radio jet and proper motions observed in HH 80 and 81.It is possible that a magneto-centrifugal wind is launched by a strong magnetic field anchored to the ∼5 M ⊙ disk.Photoionization of the inner disk surface layers may load plasma onto open field lines whose footprints co-rotate with the disk.If these field lines rotate rigidly as they expand to larger radii above and below the disk plane, they can accelerate the plasma to the observed >1,000 km s −1 speeds.Hoop stress in the azimuthal component of the field can collimate the accelerated, radial flow into a jet.

Figure 1 .
Figure 1.Overview of the HH 80/81 giant outflow in an Hα mosaic from the APO 3.5m telescoe obtained with the 30 Å bandwidth filter.This image has had a large-scale intensity gradient removed.The blue oval shows the location of IRAS 18162-2018 (GGD 27) and the approximate orientation of the radio jet emerging from this source.Yellow lines surrounding GGD 27 show the approximate location and spatial extent of the outflow cavity walls as traced in Spitzer 8 µm and 24 µm images.Dashed red circles mark the locations of various shocks and radio features.A blue circle near the top marks the location of radio source 34 at the suspected end of the radio jet.The magenta oval marks the location of the IRDC discussed in the text.A red 'X' symbol near the source region marks the B2/B3 star located at the center of the circular near-IR H2 bubble discussed in the text.The small, unlabeled, dashed circle near the bottom along the radio jet axis marks the location of a diffuse Hα knot which is also shown in Figure13.

Figures 5
Figures5 and 6show closeup views of the main HH 80/81 shock complex in [Sii] (red) and [Oiii] (cyan).The high-excitation emission traced by[Oiii]  and, to some extent, Hβ, is confined to the core of the shock complex close to the extrapolated radio jet axis.In contrast, the lower-excitation [Sii] emission is much more widespread.Figure5shows a close-up view of the HH 81 shock complex in [Sii] (red) and [Oiii] (cyan).Letter designations correspond to features marked inHeathcote et al. (1998).This color image shows a strong excitation gradient from west to east.The brightest and highest-excitation region occurs in feature A which resembled a letter 'X' in 2018 on the [Oiii] image.A series of three arcs of emission suggests shocks moving sideways with respect to the jet axis towards the southeast (features B and C).Both features A and B are [Oiii] dominated at their leading southern edges.The excitation gradient along filament C is reversed with the [Oiii] emission dominating its trailing northeastern edge.

Figure 6 .
Figure 6.The HH 80 field observed in 2018 with HST showing [Sii] (red) and [Oiii] (blue and white).The features identified by Heathcote et al. (1998) are marked at their 2018 locations.

Figure 7 .
Figure 7. HH 80 in [Oiii] showing proper motions as vectors superimposed on the 2018 image.The vector lengths correspond to the motion in a 23 year interval.Feature B discussed in the text consists of the three knots whose proper motions are labeled as 1, 2, and 3 in this figure.The A complex consists of features 4 through 8.An interactive version of this figure is available.Clicking on the image will switch between the 1995 and 2018 images to show the changes and motions between the 23 year span.

Figure 8 .
Figure 8. HH 80 in [Sii] showing proper motions as vectors superimposed on the 2018 image.The vector lengths correspond to the motion in a 23 year interval.An interactive version of this figure is available.Clicking on the image will switch between the 1995 and 2018 images to show the changes and motions between the 23 year span.

Figure 9 .Figure 10 .
Figure 9. HH 80 in Hβ showing proper motions as vectors superimposed on the 2018 image.The vector lengths correspond to the motion in a 23 year interval.An interactive version of this figure is available.Clicking on the image will switch between the 1995 and 2018 images to show the changes and motions between the 23 year span.

Figure 11 .
Figure 11.HH 81 in [Sii] showing proper motions as vectors superimposed on the 2018 image.The vector lengths correspond to the motion in a 23 year interval.An interactive version of this figure is available.Clicking on the image will switch between the 1995 and 2018 images to show the changes and motions between the 23 year span.

Figure 12 .
Figure 12.HH 81 in Hβ showing proper motions as vectors superimposed on the 2018 image.The vector lengths correspond to the motion in a 23 year interval.An interactive version of this figure is available.Clicking on the image will switch between the 1995 and 2018 images to show the changes and motions between the 23 year span. Figures 7 and 10 show this feature marked with vectors (number 9 in the [Oiii] image and number 1 in the [Sii] image) superimposed on the 2018 [Oiii] and [Sii] images of HH 80.For most features, the [Sii] proper motions are much slower and more chaotic than those in [Oiii] or Hα/Hβ.Blinking the HH 80 [Sii] images shows overall expansion of the [Sii] emission orthogonal to the extrapolated radio jet axis and the high-excitation core of the shock complex.The Appendix presents color images of the [Sii] emission in 1995 (red)

Figure 14 .Figure 15 .Figure 16 .
Figure 14.An Hα image taken with the 30 Å filter showing the HH objects in the counterflow direction north of IRAS 18162-2048 The magenta oval marks the location of the IRDC north of HH 80N A (previously HH 80N) which is discussed in the text.

Figure 17 .Figure 18 .
Figure 17.The immediate vicinity of IRAS 18162-2048 in a continuum-subtracted [Fe ii] image.The red circle marks the location of IRAS 18162-2048..(2011) who called it the 'HH 80N core'.The minor axis of this cloud is towards the northeast at PA∼30 • .Masqué et al. (2011) found a total mass of ∼20 M ⊙ for the clump.They found three young stellar objects embedded within which are potential sources for exciting HH objects in this region.Girart et al. (2001) found a bipolar molecular flow emerging from the HH 80N core.This flow is oriented nearly east-west at position angle ∼80 • .It is possible that HH 80N D is associated with this flow.However, the orientation of this outflow is inconsistent with being the driver of HH 80N G.

5. 1 .
Appendix: Proper Motions Shown as Color and Animated gif Images In this Appendix, we show proper motions as color images with the 1995 epoch images shown in red and the 2018 epoch image shown in cyan.
Figures 25 and 26 are two colored images showing the changes and motions in HH 80 and HH 81 between 1995 and 2018 in which the [Sii] emission is shown in red and the [Oiii] emission is shown in cyan.

Figure 25 .
Figure 25.HH 80 with [Sii] in red and [Oiii] in cyan in 1995 (left) and 2018 (right).An interactive version of this figure is available.Clicking on the image will switch between the 1995 and 2018 images to show the changes and motions between the 23 year span.

Figure 26 .
Figure 26.HH 81 with [Sii] in red and [Oiii] in cyan in 1995 (left) and 2018 (right).An interactive version of this figure is available.Clicking on the image will switch between the 1995 and 2018 images to show the changes and motions between the 23 year span.

Table 1 .
HST Observations Used For This Analysis Sii]) filters were used.[Oiii]and the Hα and Hβ hydrogen recombination lines trace fast shocks with speeds in excess of 100 to 150 km s −1 .[Sii]tracesslower shocks with speeds between 10 to 100 km s −1 .Unfortunately, a guiding failure resulted in the loss of the Hα image during the 2018 observations.The Hα images from 1995 are combined with the new Hβ image from 2018 for the analysis of nebular evolution and proper motions traced by hydrogen line emission.Table1lists the observations used in this analysis.

Table 2 .
Apache Point Observations Used For This Analysis 10, 11, and 12 show proper motion vectors within HH 81 in [Oiii], [Sii], and Hβ superimposed on the 2018 epoch HST image.The brightest [Oiii] emission is associated with relatively slow proper motions compared to fainter, faster features to the south and east.In Figure

Table 4 .
Counterflow HH Objects and MHOs