Draft version June 21, 2024

Typeset using LATEX twocolumn style in AASTeX631

Why are (almost) all the protostellar outflows aligned in Serpens Main?
Joel D. Green,1 Klaus M. Pontoppidan,2 Megan Reiter,3 Dan M. Watson,4 Sachindev S. Shenoy,1 P. Manoj,5
and Mayank Narang6
1 STScI, 3700 San Martin Dr, Baltimore, MD 21218, USA
2 Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, USA
3 Rice University, Houston, TX, USA
arXiv:2406.13084v1 [astro-ph.SR] 18 Jun 2024

4 University of Rochester, Rochester, NY, USA

5 Tata Institute of Fundamental Research, Mumbai, Maharashtra, IN
6 Institute of Astronomy and Astrophysics Academia Sinica, Taiepei, TW

ABSTRACT
We present deep 1.4-4.8 µm JWST-NIRCam imaging of the Serpens Main star-forming region and
identify 20 candidate protostellar outflows, most with bipolar structure and identified driving sources.
The outflow position angles (PAs) are strongly correlated, and aligned within ±24◦ of the major axis
of the Serpens filament. These orientations are further aligned with the angular momentum vectors of
the two disk shadows in this region. We estimate that the probability of this number of young stars
being co-aligned if sampled from a uniform PA distribution is 10−4 . This in turn suggests that the
aligned protostars, which seem to be at similar evolutionary stages based on their outflow dynamics,
formed at similar times with a similar spin inherited from a local cloud filament. Further, there is
tentative evidence for a systematic change in average position angle between the north-western and
south-eastern cluster, as well as increased scatter in the PAs of the south-eastern protostars. SOFIA-
HAWC+ archival dust polarization observations of Serpens Main at 154 and 214 µm are perpendicular
to the dominant jet orientation in NW region in particular. We measure and locate shock knots and
edges for all of the outflows and provide an identifying catalog. We suggest that Serpens main is
a cluster that formed from an isolated filament, and due to its youth retains its primordial outflow
alignment.

1. INTRODUCTION sociated magnetic field (Konigl & Pudritz 2000), feed-

Star formation is thought to be partly regulated by back may lead to misalignment on the smallest scales
magnetic fields with coherence scales of a few parsec (1000 au) as the protostar evolves (Hull et al. 2013).
(Crutcher 2012) – smaller than Giant Molecular Clouds, One potential tracer of the accretion flow history of
but larger than individual protostars. Magnetic fields star-forming filaments and their cores on parsec scales
likely play a key role in the collapse of cloud cores dis- is whether the angular momentum vectors of stars in a
tributed in elongated structures called filaments (Bally cluster are correlated with each other, and with direc-
et al. 1987; Smith et al. 2016). Star-forming cores tion of the magnetic field along their natal cloud filament
are indeed found to cluster along filamentary density (Nagai et al. 1998).
enhancements (André et al. 2010), however, observa- The spin axes of very young stars may be efficiently
tional confirmation of a direct influence of the mag- traced by their outflows. Indeed, the emergence of en-
netic field has been elusive and there is no consensus ergetic protostellar outflows is a ubiquitous signature of
on the detailed formation mechanism of filaments and early star formation (Frank et al. 2014). Collimated jets
their related young clusters (Hennebelle & Falgarone launching from the innermost regions of low-mass young
2012; Gómez et al. 2018). While theory often assumes stars impact surrounding molecular cloud material and
idealized alignment of protostellar disks, cores, and as- can create striking structures of shocked ionized, atomic,
and molecular gas (Reipurth & Bally 2001; Bally 2016).
Since the jets are likely accelerated and collimated by
Corresponding author: Joel D. Green a rapidly rotating poloidal magnetic field in the inner
jgreen-at-stsci-dot-edu star-disk system, they emerge along the stellar rotation
2 Green et al.

6
3 1

4 2
5

10 11

8
12

Sh1 7
Sh2
9

15
20 16

13
14

19

17

0.042 pc

18 20”

Figure 1. The central location of each outflow (green arrows) and suggested driving sources (blue stars) indicated on a
NIRCam-color image (F140M - blue, F210M - green, F360M - orange, F480M - red). The arrow and source locations are offset
from the outflow for clarity - refer to the coordinates in the catalogue for accurate outflow coordinates. This combined image
is centered at approximately 18:29:55.8 +01:14:34. Image processing credit: Alyssa Pagan.
 Alignment of Outflows in Serpens Main 3

axis and thus trace the angular momentum vector of the sensitivity of the James Webb Space Telescope (JWST;
star itself (Kwan & Tademaru 1988; Ouyed & Pudritz Gardner et al. 2023). Serpens Main is one of the dens-
1997; Banerjee & Pudritz 2006). est sections of the larger Aquila Rift, consisting of two
Jet material ejected from protostellar systems may regions of young stars (Eiroa et al. 2008; Duarte-Cabral
contain sufficient momentum to reach distances com- et al. 2010; Herczeg et al. 2019; Pokhrel et al. 2023),in-
parable to the entire cloud, giving rise to spectacu- cluding some of the densest young stellar associations
lar “parsec-scale” outflows (Eisloffel & Mundt 1997; within 500 pc (Pontoppidan et al. 2004), with an es-
Reipurth et al. 1997). As some protostellar outflows tra- timated age of 105 yr (Harvey et al. 2007). Class 0/I
verse molecular cloud core scales (∼1-2 pc) in less than sources are found primarily in the subcluster/central re-
the cloud life time, they provide an important feedback gions of both the NW and SE regions while Class II/III
mechanism that may act to limit the ability of a cloud to sources are spread out across the region (Winston et al.
form new stars (Hansen et al. 2012; Plunkett et al. 2015). 2007; Lee et al. 2014). The Serpens filament is known
Indeed, molecular clouds are known to form stars at a to display a large coherent magnetic field, possibly re-
relatively low conversion efficiency (Evans et al. 2009; lated to its formation (Kusune et al. 2019), making this
Federrath & Klessen 2012). region a good candidate for connecting alignments of
Previous searches for correlated protostellar spin axis young stars to filamentary structure. However, previ-
alignments have had mixed results. For instance, the ous wide-field imaging of CO outflows in Serpens used
UWISH2 survey (Froebrich & Makin 2016) of Cas- single-dish data at too low spatial resolution (∼15′′ ) to
siopeia/Auriga and Cygnus X (Makin & Froebrich 2018) obtain reliable statistics of outflow alignment (Graves
identified a large number of protostellar outflows and et al. 2010). In this paper, we present a JWST imag-
found uncorrelated outflow position angles (PAs) on ≳10 ing survey of protostellar outflows in the Serpens Main
pc scales. Baug et al. (2020) found no alignment in pro- cluster, and show that the orientations of the outflows
toclusters in H II regions using ALMA. More recently, are highly non-random, and perpendicular to the mag-
using JWST-NIRCam data, Reiter et al. (2022) also netic field lines of the Serpens filament. In Section 2 we
found random orientations of protostellar outflows in describe the NIRCam and ancillary observations. Sec-
NGC 3324 over a field almost 5 pc wide. Hull et al. tion 3 describes the analysis approach and the resulting
(2013) did not find evidence for alignment of the mag- outflow statistics. Finally, we interpret our findings in
netic field and outflow axes in protostars. However, Xu Section 4, and conclude with potential implications for
et al. (2022) found that outflow orientations in nearby the Serpens filament, and other star forming regions.
low mass star forming regions are significantly aligned
with dust polarization vectors at 335 GHz measured by 2. OBSERVATIONS
Planck on size scales > 0.5 pc. Further, the individual 2.1. NIRCam image
outflows are well-aligned with their immediate neighbors
We observed the Serpens Main field with the Near-
on these scales. As predicted by models (Misugi et al.
Infrared Camera (NIRCam; Rieke et al. 2023) on JWST
2023), Kong et al. (2019) found evidence of alignment
as a pre-image preparing for a Near-Infrared Spectrom-
in CO outflows perpendicular to the parent filament.
eter (NIRSpec; Jakobsen et al. 2022) survey of ices (PID
Thus, there is some prior evidence for coherence on core
1611; Pontoppidan et al. 2021). We used four medium-
(or filament) size scales that is not found on molecular
band filters, spanning 1.4 to 4.8 µm, targeting stellar
cloud scales. However, Hull et al. (2017) studied a wide
molecular bands, as well as the 2.12 µm rovibrational
range of scales in a single Serpens protostar and did not
H2 and 4.69 µm rotational H2 S(9) line. The dithering
find that the protostellar structure was aligned with a
strategy used for the JWST Early Release Observations
strong magnetic field.
(Pontoppidan et al. 2022) were used to optimize the uni-
Statistically complete, wide-field observations of the
formity of the depth over as large a fraction of the field as
youngest outflows are challenging because of the high
possible, and to minimize 1/F noise, cosmic rays and bad
dust extinction in the centers of protostellar cores
pixels. Specifically, the image is constructed as a 2×1
(AV >> 10) and the relatively small fields of view of mil-
mosaic with rows offset by 20% and with a combined
limeter interferometers. Thus, while many shock tracers
area of approximately 6.6 × 4.3 arcmin. The maximum
are found in the optical spectrum, these are not visi-
total depth in the field is 1800 s per filter, distributed
ble during the earliest stages of star formation. Con-
on 12 dithers and 7 groups using the BRIGHT2 read-
versely, infrared tracers (particularly rotational molec-
out pattern. The images were obtained in two visits on
ular hydrogen lines like H2 S(9) at 4.8 µm) are much
2023 26 Apr and 2023 12 May. We reduced the data
more accessible, in particular to the high resolution and
using the JWST calibration pipeline (Bushouse et al.
4 Green et al.

2023). However, given the lack of high-quality Gaia this band mainly confirms the presence of an outflow
astrometric reference stars, we processed the data in (see Figure 2 for a comparison). The detailed similarity
two steps. The first step processed the F360M filter of outflow candidates in the F360M and F480M bands
with the tweakreg step switched off. We then used the supports that H2 S(9) is the most likely dominant source
photutils package to detect point sources and create of emission in F480M.
an astrometric reference catalog. The remaining three The extended emission in the F210M band is likely
filters were then reduced aligning to the F360M cata- dominated by H2 v=1-0 rovibrational emission from S(0)
log to obtain a high-quality relative registration of the to S(4), with a contribution from reflection nebulosity.
image. The absolute frame was then registered with However, this band may also contain Br γ emission,
the same offset to a new frame manually adjusted to which could come from irradiated cloud edges or disso-
a combination of Gaia and 2MASS stars. The images ciative shocks and is not easily separated from molecular
were processed with version 11.16.21 of the calibration emission. On the other hand, based on the similarity of
pipeline and context jwst 1084.pmap of the Calibration the emission in the F480M band, we assume that the
Reference Database System (CRDS). The spatial reso- H2 dominates both filters. Rovibrational H2 lines are
lution of NIRCam at 4 µm is 0.′′ 13, or about 0.′′ 16 at excited under different conditions than the rotational
F480M. The properties of the filters are summarized in H2 S(9) line, and suffer from greater extinction. Con-
Table 1. sequently, only a subset of outflows appear clearly in
both F210M and F480M (Figure 2). 15 of 20 outflows
Table 1. NIRCam filter summary. are observed in F210M, although 5 of those 15 are only
partially detected compared with the full F480M mor-
Filter Wavelength Tracers phology.
µm Finally, the F140M band is dominated by reflection
nebulosity, with prominent illuminating sources such as
F140M 1.3-1.5 Reflection nebulosity
EC 82 (the Great Disk Shadow; Pontoppidan et al. 2020)
F210M 2.0-2.2 H2 v = 1 − 0 S(1) and EC 90 lighting up the SE region. We use the two
F360M 3.4-3.8 H2 v = 0 − 0 S(14)-S(18) disk shadows seen in the reflection nebulosity to aug-
F480M 4.66-5.0 v = 0 − 0 H2 S(9) ment our sample of protostars with measured position
CO v = 1 − 0 P(1)-P(32) angles (see Section 3.4), and are identified as Sh1 and
CO v = 2 − 1 P(4)-P(25) Sh2 respectively in Figure 1. We summarize the tracers
[Fe II] a4F7/2 - a6D7/2 in each filter in Table 1. Most of the north-eastern core
is not visible in F140M due to extinction.

2.3. Polarization Maps

We use archival SOFIA-HAWC+ data to sample the
2.2. Outflow tracers in NIRCam bandpasses orientation of the cloud-scale magnetic field in Serpens
Protostellar outflows are generally best detected with Main. The Serpens Main region was observed in Band
NIRCam in the F480M bandpass. This bandpass con- D (∼ 154 µm) and Band E (∼ 214 µm) with HAWC+
tains the 4.66 µm H2 S(9) line, the 4.89 µm [Fe II] line, on flight F621, on 10 Oct 2019, as a part of the SOFIA
and 54 CO fundamental P-branch lines, known to be Cycle 7 program 0130 (PI: L. Fanciullo). Serpens Main
strong in protostellar outflows (Ray et al. 2023; Feder- was observed on this flight using the On-The-Fly Map-
man et al. 2023; Rubinstein et al. 2023). Further, this ping (OTFMAP) scan mode of HAWC+. A Lissajous
longest wavelength is the least affected by extinction, scan pattern with scan angle of -30 deg, scan amplitude
with optical depths a factor 2.5 lower at 4.8 µm com- of 220 arcsec with a slew rate of 200 arcsec/s was used
pared to 2.1 µm (Pontoppidan et al. 2024). We conse- to obtain this data. Multiple pointings (4 in Band D
quently use the F480M image to identify candidate out- and 7 in Band E) were used to cover an area of 13 × 13
flows by their morphological appearance and to identify arcmin2 of the Serpens Main star-forming region with
knot and bow shock substructures within each outflow a total integration time of 1952 and 3555 sec in Bands
(see Figure 1 for an overview). D and E, respectively. This resulted in higher signal-to-
We use the F360M band to assist in identifying outflow noise ratio (SNR) in Band E compared with Band D.
parameters, but as the line emission is dominated by the Therefore we used Band E for our best sampled dataset
weaker rotational H2 S(14)–S(18) lines (Ray et al. 2023), to investigate the B-field orientation around our sample.
 Alignment of Outflows in Serpens Main 5

5 Outflow 1 / SH 2-68N F140M 5 Outflow 3 / S68Nc F140M

0 0
5 5
5 F210M 5 F210M
0 0
5 5
5 F360M 5 F360M
0 0
5 5
Distance [arcsec]

Distance [arcsec]
5 F480M 5 F480M
0 0
5 5
20 10 0 10 20 20 10 0 10 20
Distance along outflow [arcsec] Distance along outflow [arcsec]
5 Outflow 10 / EC 53 F140M 5 Outflow 8 / SMM 1a F140M
0 0
5 5
5 F210M 5 F210M
0 0
5 5
5 F360M 5 F360M
0 0
5 5
Distance [arcsec]

Distance [arcsec]

5 F480M 5 F480M
0 0
5 5
20 10 0 10 20 20 10 0 10 20
Distance along outflow [arcsec] Distance along outflow [arcsec]
Figure 2. Bandpass comparison of four prominent outflows. The outflows have been rotated with the position angle in Table 3
to align them with the horizontal axis. At a distance of 430 pc, the extent of each image along the x-axis corresponds to 0.1 pc.

The HAWC+ Band E data was re-processed using the perpendicular to the magnetic field, the direction of the
SOFIA Data Reduction software, SOFIA Redux Version magnetic field in the plane of the sky can be obtained by
1.3.3 (HAWC+ DRP Version 3.2.0). The resulting level adding π/2 to the polarization angle θ and is included in
4 mosaic of HAWC+ Band E polarization maps have a the level 4 mosaic (Hoang et al. 2014; Andersson et al.
pixel size of 4.′′ 55 and effective beam size of 18.′′ 2. The fi- 2015 and references therein). For a detailed calculation
nal level 4 data product includes Stokes parameters I, Q of each of these quantities we refer the readers to the
and U, the polarization fraction P, the polarization an- HAWC+ DRP User’s Manual and Gordon et al. (2018).
gle θ and their uncertainties. Since the thermal emission To ensure the highest quality polarization measure-
from interstellar dust grains is preferentially polarized ments and exclude low SNR pixels, we masked our Band
6 Green et al.

E array, including only pixels with SNR ≥ 150 in total 3.3. Measuring position angles
intensity (Stokes I), < 50% in percent polarization, and The outflow PAs are measured relative to the candi-
a SNR of > 3 in polarization fraction. We measured date driving source, or a central position within the out-
the average polarization angle in a half beam (9.′′ 1 ra- flow itself if no unambiguous driving source can be iden-
dius circle) around each of our targets and include it in tified. For outflows without an obvious driving source,
Table 2. the central position is based on the orientation and posi-
tion of knots and bow shocks. The central positions are
3. ANALYSIS listed in Table 2. The PA is estimated by calculating a
3.1. The Serpens Main cluster separate PA from the driving source to each identified
knot in the flow (see Figure 4). These are then averaged
Figure 1 shows a color-composite of the NIRCam im- to produce a single value. To estimate the uncertainty
age. The NW and SE regions together form a flow axis in PA, we take the width of the outermost bow shock
that constitutes the Serpens Main region; considerably edge and calculate the range of allowable angles rela-
off the south edge of the mosaic is Serpens South (Guter- tive to the central/driving source. For outflows with
muth et al. 2008). It is clear that the NW and SE re- clearly defined morphologies, this uncertainty varies be-
gions contain the densest and most opaque material in tween 1 and 10◦ , but is as high as ∼ 20◦ for nebulous,
this region. wide angle, or overlapping flows. The longer an outflow
is, or the narrower the morphology appears, the better
3.2. Identifying outflows constrained the PA becomes. Thus, outflows or tightly
It is visually apparent from Figure 1 that most out- collimated jet-like structures with clear driving sources
flows in the region appear to be aligned in position an- have the lowest uncertainty.
gle. However, to quantify the alignment, we identify An example of the identified knot structures used for
outflows in the NIRCam images based on a hierarchy the PA determination is provided for one outflow in Fig-
of criteria. Using the F480M image, which includes the ure 4 and Table 3. In this case, some of the change
strongest and least extinguished outflow lines, we visu- of PA knot-to-knot appears systematic, perhaps due to
ally searched for extended structure with a “bow shock” precession, suggesting that our PA uncertainty estimate
type morphology, defined as a ∼180 degree ‘C’ shaped is slightly conservative.
arc. Since the bow shocks are directional, we tracked
each backward until locating either: 1) another bow
shock with similar orientation, 2) a series of compact
knots indicative of a jet along the same orientation, or 3) 3.4. Position angles for edge-on disks
a continuum source that could plausibly be driving the There are two edge-on disks in the field that supple-
outflow. Any system meeting this criteria is collectively ment the source position angles indicated by the out-
considered an outflow candidate (lowest confidence class flows: EC 82 and “Shadow Jr.” (or “Shd 2”, as re-
C). For each candidate, the F360M and F210M images ferred to in this work; see Figure 1). The disk around
were inspected for counterparts to the bow shocks seen the intermediate-mass young star EC 82 casts a large
in the F480M image. If the outflow is recovered in at shadow on surrounding reflection nebulosity, giving rise
least one of the F210 or F360M filters (but not F140M, to the so-called “Great Serpens Disk Shadow”, first ana-
which does not typically reveal outflows due to extinc- lyzed using data from the Hubble Space Telescope (Pon-
tion and lack of H2 lines), the outflow candidate is given toppidan et al. 2020). The expansive shadow is most
confidence class B. Finally, if 1) a driving source can noticeable in the F140M image. Because the disk po-
plausibly be identified, or 2) another bow shock oriented sition angle is well-established, it represents a comple-
in the opposite direction, and along the outflow axis is mentary star for which the rotation axis is likely known,
detected, the outflow candidate is given the highest con- assuming it is traced by its disk. Additionally, a second,
fidence class A. Although the catalog includes outflow much smaller, disk shadow, noted in Pontoppidan et al.
candidates from all confidence classes, only those with (2020), is also visible east of EC 82. The orientation
confidence A are included in our statistics in the follow- of this second disk shadow is similar to that of EC 82.
ing analysis. The location of each outflow is shown in Although we do not clearly detect jets/outflows around
Figure 1, an aligned gallery is shown in Figure 3, and these two sources in the F480M data, they cannot be
the catalog itself is presented in Table 2. ruled out. Both angles are provided in Table 1, rotated
by 90◦ to match the outflow axes for the rest of the
sample, assuming these are perpendicular to the disk.
 Alignment of Outflows in Serpens Main 7

Outflow 1 / SH 2-68N Outflow 2 / SMM 1

Outflow 3 / S68Nc Outflow 4 / OO Ser

Outflow 5 / EC 37 Outflow 6 / S68Nb2

Outflow 7 / SMM 1b Outflow 8 / SMM 1a

Outflow 9 / SSTc2d J182950.5 Outflow 10 / EC 53

Outflow 11 / Serpens 20 Outflow 12

Outflow 13 Outflow 14 / Serpens 56

Outflow 15 Outflow 16 / SMM 3

Outflow 17 / Serpens 9 Outflow 18 / SMM 11

Distance [arcsec]

5 Outflow 19 / Ser-emb 4E Outflow 20 / 2MASS J18300491+0114393

0
5
20 10 0 10 20
Distance along outflow [arcsec]
Figure 3. A gallery of the F480M images of each outflow. The scale of each image is identical, and outflows have been rotated
by the PA provided in Table 3. Orange circles indicate the position of the driving source candidate, when known. The images
are scaled using an arcsinh function to emphasize faint, extended emission.
8 Green et al.

Table 2. Average position angle and uncertainty, and likely driving source for each outflow in this work. RA/Dec are given
for the central/driving source coordinates. Pol. is the dust polarization angle as measured in the HAWC+ Band E (216 µm
archival data (see text).

ID RA Dec PA Length Length Ratioa Pol.b Conf. Driving Source Cand.

degree degree degree arcsec degree

1 277.45017 1.27892 141.2 ± 9.3 0.39 1.12 119.0 ± 3.5 A SMM 9 (SH 2-68N)
2 277.45025 1.26917 129.6 ± 3.5 0.59 1.11 110.5 ± 3.1 B SMM 1 (S7)
3 277.45296 1.28233 112.0 ± 2.1 0.91 1.23 118.5 ± 4.6 A S68Ncc
4 277.45471 1.27225 108.3 ± 21.4 0.9-2.8 1.18 110.8 ± 3.5 A OO Ser
5 277.45521 1.275431 115.6 ± 2.5 1.4 1.06 109.3 ± 11.1 A EC37 (V370 Ser)
6 277.45663 1.28506 151.6 ± 2.7 1.3 – 259.2 ± 5.2 A S68Nb2
7 277.45704 1.24914 158.8 ± 6.9 0.51-1.7 – 166.0 ± 4.5 A SMM 1b
8 277.45742 1.25581 135.2 ± 6.2 1.5-5.1 1.03 131.3 ± 4.5 A SMM 1a
9 277.45946 1.23919 135.6 ± 9.2 0.89-1.4 1.04 179.8 ± 5.3 A SSTc2d J182950.5+01141
10 277.46321 1.27800 138.9 ± 5.2 1.7 – 116.1 ± 9.8 A EC 53
11 277.46742 1.26347 83.4 ± 0.6 1.3 1.29 265.1 ± 6.7 A Serp 20
12 277.46833 1.25169 132.7 ± 6.4 1.2 1.02 106.6 ± 4.1 A No identification
13 277.47400 1.22158 123.2 ± 15.8 0.59 5.56 161.7 ± 6.2 B No identification
14 277.47996 1.22283 68.1 ± 5.0 0.63 – 177.3 ± 6.2 A Serpens 56
15 277.49504 1.24622 156.3 ± 11.9 0.33 – 240.1 ± 6.3 B No identification
16 277.49642 1.23522 160.8 ± 0.7 1.4 1.32 239.1 ± 5.1 A SMM 3
17 277.49646 1.21064 2.7 ± 4.8 1.2 – 228.9 ± 3.7 A Serpens 9
18 277.50167 1.19583 76.1 ± 2.0 1.1 1.02 259.0 ± 6.6 B SMM 11
19 277.50296 1.21603 130.4 ± 9.1 1.7 – 266.9 ± 4.3 C Ser-emb 4E
20 277.51067 1.24542 216.8 ± 10.1 1.7 – 197.3 ± 13.2 A 2MASS J18300491+0114393
21 277.48688 1.24633 134.0 ± 5 – – 177.3 ± 5.1 A [EC92] 82
22 277.50621 1.25431 140.4 ± 5 – – 241.0 ± 39.5 A Shd 2
a The ratio of the lengths of two outflow lobes. This is only available for bipolar morphologies.
b The position angle of the polarization vector.
c The driving source position (S68Nc) presented here is the center of the central knot, as indicated in Figures 2 and 3.
 Alignment of Outflows in Serpens Main 9

Table 3. Location, PA, and distance from center (bright cavity can no longer be easily distinguished from back-
source) position of Outflow 3. ground nebulosity; for an illustration of these parame-
ters, see the top part of Figure 4. In this example, the
Knot RA Dec PA Dist Shock2W position represents the point of the bow shock.
degree degree degree arcsec We measure the full width of the bow shock by visual
identification of where each side is detected above the
Shock2W 277.44954 1.28342 -72.4 (107.6) 13.18
background. We perform a similar estimation for each
S68Nc 277.45296 1.28233 – 0 outflow knot. We may observe a weak but positive cor-
W4 277.45392 1.28203 107.9 3.78 relation between outflow length and width, but in gen-
W3 277.45433 1.28183 110.0 5.31 eral conclude that these parameters are not predictive
W2 277.45479 1.28164 110.8 7.10 of each other.
W1 277.45513 1.28156 109.8 8.38 The length of the outflows with bipolar morphology
C 277.45521 1.28136 113.6 8.96 varies considerably, from ∼ 9 − 65′′ . At a typical 430 pc
E1 277.45546 1.28142 110.4 9.73 distance to Serpens Main (Herczeg et al. 2019), assum-
E2 277.45571 1.28131 110.5 10.75 ing a shock speed of 100 km/s (Reiter et al. 2022) we
E2.5 277.45583 1.28125 110.7 11.26 find that the dynamic age of the outflows ranges from
E3 277.45600 1.28108 112.5 12.10 200 - 1400 yr, considerably younger than many of the
E4 277.45633 1.28083 114.1 13.50 outflows in the NGC 3324 study, which generally found
ShockE 277.45675 1.28058 117.5 15.30 kinematic ages of 1000–10000 yr.
Shock2E 277.45813 1.27928 120.7 21.25
4. DISCUSSION
4.1. Outflow Density
The surface density of young stars of all classes in Ser-
pens Main has been estimated at 79 YSO per pc2 (scaled
to the correct distance to Serpens; Harvey et al. 2007).
The 20 outflows we are identify are contained in a region
measuring approximately 0.6 pc × 0.5 pc, or about 66
Shock2W

outflows per pc2 . This is considerably higher density of

C W1, W2, W3, W4 S68Nc
flows than in other star forming regions observed with
ShockE E4, E3, E2.5, E2, E1

Shock2E NIRCam. Carina (NGC 3324) included about 31 identi-

fied outflows in a roughly 3 pc × 2 pc region (Reiter et al.
2022), or about 5 outflows per pc2 , more than a factor
of ten lower than in Serpens. This may be attributable
to a number of effects, including differences in resolution
(NGC 3324 is eight times the distance of Serpens), age of
the clusters, and prevalence of nearby massive stars. In
NGC 1333, a comparably-sized low mass cluster, Knee &
Sandell (2000) identify 10 outflows using rotational CO
mapping of a 0.65 pc2 region, corresponding to a density
of 15 outflows per pc2 ). This suggests that NIRCam is a
powerful instrument for surveying protostellar outflows
Figure 4. Top: Annotated F480M image of the knots com- in nearby star-forming regions.
posing Outflow 3. Bottom: Position angle of each identified
knot (relative to the central position of S68Nc) at their re- 4.2. Position Angle alignment
spective radius along the outflow axis. The horizontal dashed
The measured average position angle for each of the
line is the average position angle.
20 outflows and the 2 disks (assuming the outflow axis
is 90◦ to the disk axis) is tabulated in Table 2 and the
3.5. Outflow Dimensions distribution of these 22 angles is shown in Figure 5. Con-
The width of each outflow is measured from the ter- sidering only the 15 high confidence outflows (class A),
minus or shockfront knots of emission, where the cavity at least 8 are aligned to within ±10◦ . The two disk shad-
should be at its widest, perpendicular to the outflow ows in this region have angular momentum axes that are
position angle until clearly defined walls of the outflow aligned with the outflows adding to the total of 10 of 17
10 Green et al.

225
8
Serpens filament PA NW region SE region
200

Outflow position angle [degrees]

7 Source PA distribution
<PA> = 136.3° +/- 24.6°
6 175
Number of young stars

5 150
4
125
3
100
2
75
1
0 0 50 100 150 200 250 300 350
60 80 100 120 140 160 180 200 220 Distance along Serpens filament [arcsec]
Rotation axis position angle [E of N degrees]
Figure 5. Distribution of measured average position angles Figure 6. PA as a function of position along the filament.
for all 22 sources, clustering around the filament PA = 139◦ . The PAs of the NW region are more correlated than the
The black curve is a Gaussian fit to the distribution with pa- SW region. The line is the best linear fit after removing
rameters (mean and standard deviation) given in the legend. three outliers with the highest and lowest measured PA. The
shaded region shows the 99% confidence level of the fit.
high confidence orientations falling within a ±10◦ span.
Further, 14 of 17 objects have PAs falling within a ±30◦ metrical lobes are not a direct indicator of an edge-on
span. inclination system, because of the local extinction or
We used a simple Monte Carlo analysis to test the null distortion through interaction with the local cloud ma-
hypothesis that the catalogued outflow orientations are terial, it is likely that strongly inclined outflows would
randomly distributed. For all of the calculations here, not show such symmetry across the sample. For exam-
we assumed a uniform distribution of outflow PAs be- ple, Habel et al. (2021) consider this criteria in identify-
tween 0 and 180◦ . To determine the likelihood of the ing bipolar outflows with more edge-on systems. The
PA distribution arising randomly from a uniform distri- close symmetry is at least consistent with relatively
bution, we used the numpy random number generator to edge-on, and therefore relatively similar inclination an-
produce 105 instances and note the number of occur- gles. Considering this inclination constraint along with
rences with at least the observed PA clustering. The the tight clustering of position angles, this supports
odds of 10 of 17 uniformly distributed sources falling in the idea that these outflows are similarly oriented in
a single 20◦ bin is ∼ 0.002%, and the odds of 14 of 17 3-dimensional space. However, as many of these out-
high confidence sources being aligned in a 60◦ bin is only flows extend considerably beyond common protostellar
slightly higher at about 0.005%. envelope scales, or have asymmetric structures close to
Figure 6 shows the distribution of outflow position the driving source, bipolar symmetry at large distances
angles as a function of driving source position along the is suggestive rather than conclusive.
axis of the Serpens filament. The axis is estimated to
be at PA of 139◦ , along the line connecting the centers
of mass from the SW to the NE regions from the FIR 4.3. Outflow orientation vs. dust polarization vectors
imagery of the Serpens region / Aquila Rift (Gong et al. To compare the filament and individual outflow ori-
2021). This parameter is used as a measure of location entations with the larger scale magnetic field, we com-
along the filament; north-west to south-east. There is pared our results with archival datasets. First we com-
a strong correlation with the north-western part of out- pared Figure 2 from Kwon et al. (2022) - their map of
flows clustering in position angle around a mean of 136◦ . the inferred magnetic field vectors - with our NIRCam
mosaic. It was immediately apparent that the magnetic
4.2.1. Are the outflows at similar inclination angles?
field lines were roughly perpendicular to the outflow di-
Outflows 1-4. 7-9, 11 and 12 are all bipolar, with rection in the NW region, but are less organized and
their lobe length ratios between 1.02 and 1.29 (ie. 2 - systematic in the rest of the field, except along the iden-
29% deviation from perfect symmetry). Although sym- tified filaments from the Kwon et al. (2022) analysis.
 Alignment of Outflows in Serpens Main 11

Figure 7. Left: Overlay of HAWC+ Band E polarization vectors (white arrows) on the full 6.6 × 4.3 arcmin NIRCam F480M
image from this work. The SOFIA vectors match well with SCUBA maps from Kwon et al. (2022). Right: Zoom on the NW
region filament, where the most aligned outflows are located. This inset region spans approximately 1.7 × 1.9 arcmin (N-S ×
E-W extent, respectively).

To improve the resolution and better resolve individ- 4.4. Comparison to outflow alignments in other regions
ual driving sources/cores, we re-reduced and interpo- Outflow surveys in other star-forming regions of-
lated the HAWC+ Band E dust polarization vectors to ten find no preferred outflow orientation, but typically
the positions of each of the 22 center positions, shown on much larger scales than Serpens Main (5-10 pc;
in Table 3, and displayed them in Figure 7 overlaid on Stephens et al. 2017; Baug et al. 2020; Reiter et al.
the F480M NIRCam image. For display purposes, we 2022). The sensitivity and spatial resolution of NIRCam
scaled the lengths of the polarization vectors for easier to detect a statistically significant number of outflows on
visual comparison with outflow orientations. It is ap- scales smaller than ∼1 pc may explain, in part, why we
parent that the Band E vectors closely track most of detect the alignment in Serpens Main. Indeed, there is
the outflows, and a comparison of the position angles in existing evidence of relative alignment between outflow
Table 2 confirms this. Of the 12 outflows in the NW axes on such scales for the youngest clusters in filaments
region (see the rightmost panel of Fig 7 for a zoomed-in (e.g., Davis et al. 2007; Kong et al. 2019). Thus, we
view), all but 2 are within 25 degrees of alignment with suggest that our NIRCam image indicates that align-
their respective magnetic field polarization vectors. One ment has a coherence scale of ≲ 1 pc, and that align-
of those 2 (outflow 9) has no identified driving source. ment is rapidly degraded with time due to precession
The other, outflow 6, is the only significant outlier in and binary interactions. Misalignment processes are
this region. The alignment with the magnetic field in the predicted to occur on timescales of 105 - 106 yr (Lai
SE region is less correlated. Of outflows 13-20, only 2 2014). If these effects randomized spins on timescales
(outflows 18 and 20) are closely aligned with the nearby much shorter than that, the observed alignment would
polarization vectors. The polarization vectors do not not be possible. While Misugi et al. (2023) predicts
align with the two disk shadows either. Conversely, the that the core rotation axes (not necessarily individual
polarization fraction, indicated by the vector length, is outflows) are perpendicular to the filament, and Kong
larger in the SE region, and around the disk shadows, et al. (2019) find an example of this, this is inconsistent
than it is in the NW region. with the dominant orientation of the Serpens outflows,
12 Green et al.

which instead appear to be aligned with the filament this paper. Using the same outflow tracers would pro-
axis (Figure 5). However, the large-scale orientation of vide a more direct comparison of the outflow orientation
the dynamical filament in the Serpens Main region may of these regions.
be different than the simplified axis defined by the vec- Weaker fields may also lead to less outflow alignment
tor between the NW to SE clusters. The filament seen in a given region. Xu et al. (2022) propose that weaker
in the extinction map in Fiorellino et al. (2021) (their field strengths may contribute to the lack of outflow
Figure 15) presents an arc, rather than a linear struc- alignment in Perseus (Stephens et al. 2017). If true,
ture, suggesting a more complex arrangement in which this predicts a stronger magnetic field in Serpens Main
the orientation of the filament potentially changed since (≈ 60 − 300µG; Kwon et al. 2022). However, a more di-
the initial fragmentation of the cluster. Thus, while we rect comparison of the degree of outflow alignment with
note the discrepancy in the protostellar alignment with the local magnetic field is required to test this hypothe-
the apparent filament orientation compared to the theo- sis. Nevertheless, our results are consistent with several
retical expectation (parallel rather than perpendicular), other studies that find a higher degree of outflow align-
this is not necessarily strong evidence against the the- ment in the youngest, darkest regions of the cloud (e.g.
oretical prediction. Further dynamical modeling is re- Davis et al. 2007; Makin & Froebrich 2018).
quired to explain the apparent parallel alignment of the
5. SUMMARY
Serpens outflow axes with the elongation of the local
Serpens cloud. We observed the Serpens Main star forming region
How are alignments related to the magnetic field? Re- with JWST-NIRCam, at 1.4, 2.1, 3.6, and 4.8 µm. In
cently, Xu et al. (2022) showed that outflow orientations addition to new views of the star forming complex, the
are not random compared to the large-scale magnetic images were sensitive to protostellar outflows.
field. We see close alignment in magnetic field orienta- We identified 20 outflows by their bow shock mor-
tion and the outflows in the NW region presented here, phology and ancillary data on driving sources, devel-
but not in the SE region. These alignments suggest oping a catalog of outflows including knot locations,
that the large-scale magnetic fields that help funnel ma- radii, length, and position angle. 15 of the 20 out-
terial onto filaments also determine the initial orienta- flows fall into our highest confidence detection bins, with
tion of the outflow axes. Observations suggest less out- identified driving sources, most noted in previous sur-
flow alignment over time as stellar feedback disrupts the veys. We examined dust polarization images taken by
magnetic field alignment and anisotropic accretion alters SOFIA/HAWC+ to provide magnetic field alignment
the outflow axes of the embedded protostars. An alter- and context, considering published ancillary measure-
nate, larger scale effect could be cloud-cloud collisions. ments from JCMT-SCUBA, ALMA, and Spitzer-IRAC.
Duarte-Cabral et al. (2010) also identified the NE region We analyzed the outflows and summarize our results
as containing more uniform conditions for young stars, below:
but they argued instead the SE region was “perturbed”
• NIRCam/F480M is particularly well-suited to
by a cloud-cloud collision in progress, while the NW re-
detect outflows because it contains molecular,
gion was “homogeneous”. We argue here that the align-
atomic, and ionic tracers that all emit strongly in
ment of spin axes is further evidence for a lack of per-
protostellar outflows/jets. The result is a mixed
turbation of the NW clump. Alignments are more pro-
morphological catalog with a high detection rate.
nounced in young regions (e.g., Kong et al. 2019) while
there is less evidence for a preferential outflow direction • 12 outflows were identified in the northwestern
in older regions and those significantly affected by stel- filament/region, while 8 outflows were identified
lar feedback (e.g., Feddersen et al. 2020). This suggests in the southeastern filament/region. Additionally,
that outflow alignment may be common in young re- two prominent disk shadows were confirmed in the
gions but quickly disrupted. The disruption time likely central region.
depends on the strength of the magnetic field and the
• The axes of the 12 outflows in the NW region are
density of the region.
inconsistent with random orientations and align
Serpens Main is similar to Ophiuchus in age, mass,
with the filament direction from NW to SE. Ad-
and average density (Evans et al. 2009). However, Xu
ditionally, the position angle of jets/outflows from
et al. (2022) find a larger range of CO outflow position
the 2 identified disk shadows align with the fil-
angles in Ophiucus than we find in Serpens Main. Mil-
ament axis. We estimate <0.005% probability of
limeter CO emission tends to trace less collimated out-
the the observed alignments if sampled from a uni-
flow components than the infrared emission presented in
form distribution in position angle.
 Alignment of Outflows in Serpens Main 13

• The position angles of the outflows align with

SOFIA/HAWC+ 214 µm dust polarization vec-
tors measured locally around each driving source.
However, the disk shadows do not align with their
local magnetic fields. This broad alignment does
not apply in the SE region. Few of the 8 identified
outflows in this region align with the filament axis,
or with the dust polarization vector.

• The density of outflows detected in this catalog (∼

66 outflows per pc2 ) is higher than other low mass
star forming regions (e.g., NGC 1333), and ten
times greater than observed by JWST/NIRCam
in Carina (NGC 3324).

The alignment of outflows with the filament axis in

part of Serpens Main, but not in the rest of the region,
is suggestive of an evolutionary process. It appears that
star formation proceeded along a magnetically confined
filament that set the initial spin for most of protostars.
We hypothesize that in the NW region, which may be
younger, the alignment is preserved, whereas the spin
axes have had time to precess or dissociate through dy-
namic interactions in the SE region. The disk shadows,
which may be more evolved sources, appear to have re-
tained their spin axis relative to the original field lines,
but the magnetic field itself has shifted, or the material
from the early formation period has notably dispersed
(evident by their scattered light emission in F140M) af-
ter their development phase.
Above all, this work shows that even a single pair of
JWST/NIRCam images in medium bands can provide
considerable insight into the history of star-forming re-
gions. We anticipate more detailed studies of star form-
ing filaments with JWST in the future.
14 Green et al.

Facilities: JWST(NIRCam), SOFIA(HAWC+),

The authors would like to thank Nicole Arulanantham, JCMT(SCUBA)
Sylvia Baggett, Neal Evans, Will Fischer, Nicole Kar-
nath, Tom Megeath, Stella Offner, Amanda Pagul, and Software: Astropy (Astropy Collaboration et al.
Adam Rubinstein for helpful discussions and insights. 2013, 2018, 2022)
We also thank Alyssa Pagan for producing a beautiful Software: DS9 (Joye & Mandel 2003)
composite NIRCam image of Serpens Main. We thank
the JWST Program 1611 team for the use of their pro-
Software: IDL
prietary pre-imaging data. We thank Lapo Fanciullo
and their team for taking the SOFIA-HAWC+ data, and
the SOFIA data pipeline developers for enabling us to
quickly reduce the archival data. We thank the anony-
mous referee for a thorough and considered report that
significantly improved the final manuscript. A portion
of this research was carried out at the Jet Propulsion
Laboratory, California Institute of Technology, under a
contract with the National Aeronautics and Space Ad-
ministration (80NM0018D0004). This research used the
NASA ADS and Simbad (CDS) databases. This work is
based on observations made with the NASA/ESA/CSA
James Webb Space Telescope. The data presented in
this article were obtained from the Mikulski Archive
for Space Telescopes (MAST) at the Space Telescope
Science Institute, which is operated by the Associa-
tion of Universities for Research in Astronomy, Inc., un-
der NASA contract NAS 5-03127 for JWST. The spe-
cific observations analyzed can be accessed via DOI:
10.17909/pv1h-ta47. This research also uses data from
IRSA and the included SOFIA science archive.

APPENDIX

A. NOTES ON INDIVIDUAL OUTFLOWS

A.1. Outflow 1 (SH 2-68N)
This outflow likely corresponds to the molecular (CO) outflow associated with SH 2-68N (S68N or J182948.1+011644;
Aso et al. 2019; Dunham et al. 2015), which in turn is part of the SMM 9 region (Tychoniec et al. 2019). The long
wavelength emission has a PA of ∼135◦ .

A.2. Outflow 2 (S7)

This is likely part of the blue lobe of the larger SMM 1 outflow, also known as S7 (Herbst et al. 1997; Caratti o
Garatti et al. 2006). Like several other sources in our sample, the IR morphology resembles the optical emission from
an HH object with its umbrella-shaped bow shocks. This may be because the rotational H2 emission dominates this
source. While this is consistent with the lack of a clear driving source near the feature, we cannot rule out that this is
a separate outflow coincident with the SMM 1 outflow based on the NIRCam image alone. Because of the ambiguity
in interpretation, we classify this in the middle confidence bin.

A.3. Outflow 3 (S68Nc)

Outflow 3 (Figure 4) is identified as S9 in Herbst et al. (1997), and likely associated with the class 0 star S68Nc (Aso
et al. 2019). It is among the brightest outflows seen in the NIRCam field. The western extent of the outflow appears
to be an isolated bow shock with the rest of the western lobe hidden by extinction. The eastern lobe is a prominent
line of bow shocks, and then a significant break to a final bow shock on the eastern edge. Including all of these shocks
 Alignment of Outflows in Serpens Main 15

increases the uncertainty of the PA, but the consistent arc suggests a slow precession. If we neglect the final Shock2E
location, we find a mostly symmetric outflow with a steady precession rate of 5◦ over 14′′ of flow. Assuming a flow
velocity of 100 km/s, that translates to a precession rate of ∼ 1◦ per 57 yr. This is comparable to the rate of some
other known precessing protostellar systems traced via outflow ejecta (e.g. Cunningham et al. 2009).

A.4. Outflow 4 (OO Ser)

This outflow is associated with the FUor candidate OO Ser (Hodapp et al. 1996). It has a broad hourglass shape
and relatively short extent in both directions, leading to a somewhat larger uncertainty in PA.

A.5. Outflow 5 (V370 Ser)

This chain of knots points back to EC37/V370 Ser. While Hodapp et al. (2012) was not able to measure a PA from
H2 emission near the source, the presence of this remote bow shock suggests a more edge-on orientation of the EC 37
system. Hodapp et al. (2012) indicated that the knots to the west (MHO 2218) were likely associated with the nearer
EC 37 system, and these knots (MHO 3245) are associated with OO Ser (based on the catalog from Davis et al. 2010),
the NIR bow shock directions suggest that OO Ser is ejecting the material in outflow 4.

A.6. Outflow 6 (S68Nb2)

This outflow is associated with the infrared-bright class 0/I source Serpens 7/S68Nb2 (Gutermuth et al. 2009; Aso
et al. 2019).

A.7. Outflows 7 and 8

Outflows 6 and 7 are likely associated with the SMM 1a and 1b binary, respectively, as their location and PAs match
well with the CO outflows in Tychoniec et al. (2019). An alternate interpretation for outflow 7 has the driving source
as the red protostar south of Serpens SMM1, known as EC 40 or SSTc2d J182949.6+011456 (Gutermuth et al. 2009).
In this work, we assume the latter scenario, because of the ALMA-derived kinematics of the dual outflow from the
SMM1 binary (Tychoniec et al. 2019).

A.8. Outflow 9
This has in the past been associated with the SMM 1 outflow (Caratti o Garatti et al. 2006). However, we identify
bipolar shapes that appear as bow shocks, which could be contrary to this interpretation. The orientation points back
to SSTc2d J182950.5+011417 (Harvey et al. 2007), although this would be newly identified as a driving source.

A.9. Outflow 10 (EC 53)

We interpret the driving source as the episodic flaring young protostar EC 53 (Baek et al. 2020), driving a long chain
of knots. The distance between the southernmost knots would suggest that mass loss episodes are ∼ 1000 yr apart,
which is not consistent with the burst phase of EC 53 (∼ 1.5 yr), although it is possible that the individual knots are
each the result of a series of bursts. There is some evidence for precession as well.

A.10. Outflow 11
This outflow may be driven by Serpens 20. The center point is coincident with J182952.22+011547.4 (Gutermuth
et al. 2009), a young stellar object identified in the Spitzer catalog.

A.11. Outflow 12
It is unclear where in this morphologically complex flow the origin/driving source lies. There are a few options
of nearby sources, including EC 55 (Eiroa & Casali 1992), which lies at the western terminus of the outflow as we
characterize it in this catalog. For our purposes, we identify a knot of emission in the center that we ascribe to a previ-
ously unknown candidate driving source. While the well-characterized shape of the flow and clear directionality lends
confidence in our identification of the outflow, a future proper motion observation would be required for confirmation.

A.12. Outflow 13
Outflow 13 is only detected via a single bow shock and no driving source is identified. The bow shock does not
appear in F140M, which supports the shock interpretation, rather than scattered light off a pillar. The closest YSO is
J18295354+0113051, a 2MASS source (Cutri et al. 2003) and detected with Gaia (Herczeg et al. 2019).
16 Green et al.

A.13. Outflow 14 (Serpens 56)

This outflow is associated with the nearby bright young star Serpens 56 (Gutermuth et al. 2009).

A.14. Outflow 15
Outflow 15 is only detected via a single bow shock and no driving source is identified. The bow shock does appear in
F210M (and not F140M), which supports the shock interpretation. The closest YSO is J18295914+0114411, a 2MASS
source (Cutri et al. 2003).

A.15. Outflow 16 (SMM3)

This outflow is associated with the SMM3 protostar, which is not itself visible in any of the NIRCam bands, although
it is well-detected by ALMA at 230 GHz, and by SCUBA at 450 µm (Davis et al. 1999). The Class II YSO CK 8
is located (in projection) along the outflow, but does not appear to be interacting with it. The northern bow shock
is very bright, and is partially saturated in F210M. There is a point source in the bow shock visible at F360M and
F480M, but it is not clear if this is an unrelated embedded source.

A.16. Outflow 17
Serpens 9, a radio (VLA) source and protostar to the east of the main cluster (Bontemps et al. 1996), is well-aligned
with the unipolar outflow, and we identify it as the driving source candidate.

A.17. Outflow 18
This outflow falls into our lowest confidence bin because of a non-visible driving source, and the somewhat disorga-
nized shape of the knots to which we ascribe it, but does appear to be a symmetric bow shock around a submm source
SMM11 (Aso et al. 2017).

A.18. Outflow 19
This outflow candidate falls into our lowest confidence bin. First, although we identify a potential driving source
(Ser-emb 4E; Enoch et al. 2011) based on the orientation of the bow shocks, there is no obvious nebulosity link between
it and the outflow. Second, the outflow does not appear in the F210M band at all, suggesting it could have a different
origin than shocked emission. Third, the tip resembles a cloud pillar, and sits near the highest extinction region in the
southern region.

A.18.1. Outflow 20
No apparent driving source is identified, but this object was previously noted as HH 459 (Ziener & Eislöffel 1999).
A candidate driving source is 2MASS J18300491+0114393.

REFERENCES
Andersson, B. G., Lazarian, A., & Vaillancourt, J. E. 2015, Astropy Collaboration, Price-Whelan, A. M., Lim, P. L.,
ARA&A, 53, 501, et al. 2022, ApJ, 935, 167, doi: 10.3847/1538-4357/ac7c74
doi: 10.1146/annurev-astro-082214-122414 Baek, G., MacFarlane, B. A., Lee, J.-E., et al. 2020, ApJ,
André, P., Men’shchikov, A., Bontemps, S., et al. 2010, 895, 27, doi: 10.3847/1538-4357/ab8ad4
A&A, 518, L102, doi: 10.1051/0004-6361/201014666 Bally, J. 2016, ARA&A, 54, 491,
Aso, Y., Ohashi, N., Aikawa, Y., et al. 2017, ApJL, 850, L2, doi: 10.1146/annurev-astro-081915-023341
doi: 10.3847/2041-8213/aa9701 Bally, J., Langer, W. D., Stark, A. A., & Wilson, R. W.
Aso, Y., Hirano, N., Aikawa, Y., et al. 2019, ApJ, 887, 209, 1987, ApJL, 312, L45, doi: 10.1086/184817
doi: 10.3847/1538-4357/ab5284 Banerjee, R., & Pudritz, R. E. 2006, ApJ, 641, 949,
Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., doi: 10.1086/500496
et al. 2013, A&A, 558, A33, Baug, T., Wang, K., Liu, T., et al. 2020, ApJ, 890, 44,
doi: 10.1051/0004-6361/201322068 doi: 10.3847/1538-4357/ab66b6
Astropy Collaboration, Price-Whelan, A. M., Sipőcz, B. M., Bontemps, S., Andre, P., Terebey, S., & Cabrit, S. 1996,
et al. 2018, AJ, 156, 123, doi: 10.3847/1538-3881/aabc4f A&A, 311, 858
 Alignment of Outflows in Serpens Main 17

Bushouse, H., Eisenhamer, J., Dencheva, N., et al. 2023, Froebrich, D., & Makin, S. V. 2016, MNRAS, 462, 1444,
JWST Calibration Pipeline, 1.11.2, Zenodo, doi: 10.1093/mnras/stw1766
doi: 10.5281/zenodo.8140011 Gardner, J. P., Mather, J. C., Abbott, R., et al. 2023,
Caratti o Garatti, A., Giannini, T., Nisini, B., & PASP, 135, 068001, doi: 10.1088/1538-3873/acd1b5
Lorenzetti, D. 2006, A&A, 449, 1077, Gómez, G. C., Vázquez-Semadeni, E., & Zamora-Avilés, M.
doi: 10.1051/0004-6361:20054313 2018, MNRAS, 480, 2939, doi: 10.1093/mnras/sty2018
Crutcher, R. M. 2012, ARA&A, 50, 29, Gong, Y., Belloche, A., Du, F. J., et al. 2021, A&A, 646,
doi: 10.1146/annurev-astro-081811-125514 A170, doi: 10.1051/0004-6361/202039465
Cunningham, N. J., Moeckel, N., & Bally, J. 2009, ApJ, Gordon, M. S., Lopez-Rodriguez, E., Andersson, B. G.,
692, 943, doi: 10.1088/0004-637X/692/2/943 et al. 2018, arXiv e-prints, arXiv:1811.03100,
Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, doi: 10.48550/arXiv.1811.03100
VizieR Online Data Catalog: 2MASS All-Sky Catalog of Graves, S. F., Richer, J. S., Buckle, J. V., et al. 2010,
Point Sources (Cutri+ 2003), VizieR On-line Data MNRAS, 409, 1412,
Catalog: II/246. Originally published in: doi: 10.1111/j.1365-2966.2010.17140.x
2003yCat.2246....0C Gutermuth, R. A., Megeath, S. T., Myers, P. C., et al.
Davis, C. J., Gell, R., Khanzadyan, T., Smith, M. D., & 2009, ApJS, 184, 18, doi: 10.1088/0067-0049/184/1/18
Jenness, T. 2010, A&A, 511, A24, Gutermuth, R. A., Bourke, T. L., Allen, L. E., et al. 2008,
doi: 10.1051/0004-6361/200913561
ApJL, 673, L151, doi: 10.1086/528710
Davis, C. J., Kumar, M. S. N., Sandell, G., et al. 2007,
Habel, N. M., Megeath, S. T., Booker, J. J., et al. 2021,
MNRAS, 374, 29, doi: 10.1111/j.1365-2966.2006.11163.x
ApJ, 911, 153, doi: 10.3847/1538-4357/abded8
Davis, C. J., Matthews, H. E., Ray, T. P., Dent, W. R. F.,
Hansen, C. E., Klein, R. I., McKee, C. F., & Fisher, R. T.
& Richer, J. S. 1999, MNRAS, 309, 141,
2012, ApJ, 747, 22, doi: 10.1088/0004-637X/747/1/22
doi: 10.1046/j.1365-8711.1999.02836.x
Harvey, P., Merı́n, B., Huard, T. L., et al. 2007, ApJ, 663,
Duarte-Cabral, A., Fuller, G. A., Peretto, N., et al. 2010,
1149, doi: 10.1086/518646
A&A, 519, A27, doi: 10.1051/0004-6361/200913919
Hennebelle, P., & Falgarone, E. 2012, A&A Rv, 20, 55,
Dunham, M. M., Allen, L. E., Evans, Neal J., I., et al.
doi: 10.1007/s00159-012-0055-y
2015, ApJS, 220, 11, doi: 10.1088/0067-0049/220/1/11
Herbst, T. M., Beckwith, S. V. W., & Robberto, M. 1997,
Eiroa, C., & Casali, M. M. 1992, A&A, 262, 468
ApJL, 486, L59, doi: 10.1086/310828
Eiroa, C., Djupvik, A. A., & Casali, M. M. 2008, in
Herczeg, G. J., Kuhn, M. A., Zhou, X., et al. 2019, ApJ,
Handbook of Star Forming Regions, Volume II, ed.
878, 111, doi: 10.3847/1538-4357/ab1d67
B. Reipurth, Vol. 5, 693, doi: 10.48550/arXiv.0809.3652
Hoang, T., Lazarian, A., & Martin, P. G. 2014, ApJ, 790, 6,
Eisloffel, J., & Mundt, R. 1997, AJ, 114, 280,
doi: 10.1088/0004-637X/790/1/6
doi: 10.1086/118473
Enoch, M. L., Corder, S., Duchêne, G., et al. 2011, ApJS, Hodapp, K. W., Chini, R., Watermann, R., & Lemke, R.
195, 21, doi: 10.1088/0067-0049/195/2/21 2012, ApJ, 744, 56, doi: 10.1088/0004-637X/744/1/56
Evans, Neal J., I., Dunham, M. M., Jørgensen, J. K., et al. Hodapp, K.-W., Hora, J. L., Rayner, J. T., Pickles, A. J., &
2009, ApJS, 181, 321, doi: 10.1088/0067-0049/181/2/321 Ladd, E. F. 1996, ApJ, 468, 861, doi: 10.1086/177742
Feddersen, J. R., Arce, H. G., Kong, S., et al. 2020, ApJ, Hull, C. L. H., Plambeck, R. L., Bolatto, A. D., et al. 2013,
896, 11, doi: 10.3847/1538-4357/ab86a9 ApJ, 768, 159, doi: 10.1088/0004-637X/768/2/159
Federman, S., Megeath, S. T., Rubinstein, A. E., et al. Hull, C. L. H., Mocz, P., Burkhart, B., et al. 2017, ApJL,
2023, arXiv e-prints, arXiv:2310.03803, 842, L9, doi: 10.3847/2041-8213/aa71b7
doi: 10.48550/arXiv.2310.03803 Jakobsen, P., Ferruit, P., Alves de Oliveira, C., et al. 2022,
Federrath, C., & Klessen, R. S. 2012, ApJ, 761, 156, A&A, 661, A80, doi: 10.1051/0004-6361/202142663
doi: 10.1088/0004-637X/761/2/156 Joye, W. A., & Mandel, E. 2003, in Astronomical Society of
Fiorellino, E., Elia, D., André, P., et al. 2021, MNRAS, the Pacific Conference Series, Vol. 295, Astronomical
500, 4257, doi: 10.1093/mnras/staa3420 Data Analysis Software and Systems XII, ed. H. E.
Frank, A., Ray, T. P., Cabrit, S., et al. 2014, in Protostars Payne, R. I. Jedrzejewski, & R. N. Hook, 489
and Planets VI, ed. H. Beuther, R. S. Klessen, C. P. Knee, L. B. G., & Sandell, G. 2000, A&A, 361, 671
Dullemond, & T. Henning, 451–474, Kong, S., Arce, H. G., Maureira, M. J., et al. 2019, ApJ,
doi: 10.2458/azu uapress 9780816531240-ch020 874, 104, doi: 10.3847/1538-4357/ab07b9
18 Green et al.

Konigl, A., & Pudritz, R. E. 2000, in Protostars and Pontoppidan, K. M., Charnley, S. B., Dartois, E., et al.
Planets IV, ed. V. Mannings, A. P. Boss, & S. S. Russell, 2021, A chemical census of volatile ices in protostellar
759, doi: 10.48550/arXiv.astro-ph/9903168 envelopes, JWST Proposal. Cycle 1, ID. #1611
Kusune, T., Nakamura, F., Sugitani, K., et al. 2019, PASJ, Pontoppidan, K. M., Barrientes, J., Blome, C., et al. 2022,
71, S5, doi: 10.1093/pasj/psz040 ApJL, 936, L14, doi: 10.3847/2041-8213/ac8a4e
Kwan, J., & Tademaru, E. 1988, ApJL, 332, L41, Ray, T. P., McCaughrean, M. J., Caratti o Garatti, A.,
doi: 10.1086/185262 et al. 2023, Nature, 622, 48,
Kwon, W., Pattle, K., Sadavoy, S., et al. 2022, ApJ, 926, doi: 10.1038/s41586-023-06551-1
163, doi: 10.3847/1538-4357/ac4bbe Reipurth, B., & Bally, J. 2001, ARA&A, 39, 403,
Lai, D. 2014, MNRAS, 440, 3532, doi: 10.1146/annurev.astro.39.1.403
doi: 10.1093/mnras/stu485 Reipurth, B., Bally, J., & Devine, D. 1997, AJ, 114, 2708,
Lee, K. I., Fernández-López, M., Storm, S., et al. 2014, doi: 10.1086/118681
ApJ, 797, 76, doi: 10.1088/0004-637X/797/2/76 Reiter, M., Morse, J. A., Smith, N., et al. 2022, MNRAS,
Makin, S. V., & Froebrich, D. 2018, ApJS, 234, 8, 517, 5382, doi: 10.1093/mnras/stac2820
doi: 10.3847/1538-4365/aa8862 Rieke, M. J., Kelly, D. M., Misselt, K., et al. 2023, PASP,
Misugi, Y., Inutsuka, S.-i., & Arzoumanian, D. 2023, ApJ, 135, 028001, doi: 10.1088/1538-3873/acac53
943, 76, doi: 10.3847/1538-4357/aca88d
Rubinstein, A. E., Tyagi, H., Nazari, P., et al. 2023, arXiv
Nagai, T., Inutsuka, S.-i., & Miyama, S. M. 1998, ApJ, 506,
e-prints, arXiv:2312.07807,
306, doi: 10.1086/306249
doi: 10.48550/arXiv.2312.07807
Ouyed, R., & Pudritz, R. E. 1997, ApJ, 482, 712,
Smith, R. J., Glover, S. C. O., Klessen, R. S., & Fuller,
doi: 10.1086/304170
G. A. 2016, MNRAS, 455, 3640,
Plunkett, A. L., Arce, H. G., Corder, S. A., et al. 2015,
doi: 10.1093/mnras/stv2559
ApJ, 803, 22, doi: 10.1088/0004-637X/803/1/22
Stephens, I. W., Dunham, M. M., Myers, P. C., et al. 2017,
Pokhrel, R., Megeath, S. T., Gutermuth, R. A., et al. 2023,
ApJ, 846, 16, doi: 10.3847/1538-4357/aa8262
ApJS, 266, 32, doi: 10.3847/1538-4365/acbfac
Tychoniec, L., Hull, C. L. H., Kristensen, L. E., et al. 2019,
Pontoppidan, K. M., Evans, N., Bergner, J., & Yang, Y.-L.
A&A, 632, A101, doi: 10.1051/0004-6361/201935409
2024, Research Notes of the AAS, 8, 68
Winston, E., Megeath, S. T., Wolk, S. J., et al. 2007, ApJ,
Pontoppidan, K. M., Green, J. D., Pauly, T. A., Salyk, C.,
669, 493, doi: 10.1086/521384
& DePasquale, J. 2020, ApJ, 896, 169,
Xu, D., Offner, S. S. R., Gutermuth, R., & Tan, J. C. 2022,
doi: 10.3847/1538-4357/ab91ae
ApJ, 941, 81, doi: 10.3847/1538-4357/aca153
Pontoppidan, K. M., van Dishoeck, E. F., & Dartois, E.
Ziener, R., & Eislöffel, J. 1999, A&A, 347, 565
2004, A&A, 426, 925, doi: 10.1051/0004-6361:20041276