The University of Manchester Research

The CARMA-NRO Orion Survey: Core Emergence andKinematics in the Orion A CloudDOI:10.3847/1538-4357/ab311e

Document VersionAccepted author manuscript

Link to publication record in Manchester Research Explorer

Citation for published version (APA):Smith, R., & et al. (2019). The CARMA-NRO Orion Survey: Core Emergence and Kinematics in the Orion A Cloud.The Astrophysical Journal. https://doi.org/10.3847/1538-4357/ab311e

Published in:The Astrophysical Journal

Citing this paperPlease note that where the full-text provided on Manchester Research Explorer is the Author Accepted Manuscriptor Proof version this may differ from the final Published version. If citing, it is advised that you check and use thepublisher's definitive version.

General rightsCopyright and moral rights for the publications made accessible in the Research Explorer are retained by theauthors and/or other copyright owners and it is a condition of accessing publications that users recognise andabide by the legal requirements associated with these rights.

Takedown policyIf you believe that this document breaches copyright please refer to the University of Manchester’s TakedownProcedures [http://man.ac.uk/04Y6Bo] or contact uml.scholarlycommunications@manchester.ac.uk providingrelevant details, so we can investigate your claim.

Download date:12. Oct. 2020

Draft version July 9, 2019Typeset using LATEX preprint2 style in AASTeX61

THE CARMA-NRO ORION SURVEY: CORE EMERGENCE AND KINEMATICS IN THEORION A CLOUD

Shuo Kong,1 Hector G. Arce,1 Anneila I. Sargent,2 Steve Mairs,3 Ralf S. Klessen,4, 5

John Bally,6 Paolo Padoan,7, 8 Rowan J. Smith,9 Marıa Jose Maureira,10

John M. Carpenter,11 Adam Ginsburg,12 Amelia M. Stutz,13, 14 Paul Goldsmith,15

Stefan Meingast,16 Peregrine McGehee,17 Alvaro Sanchez-Monge,18 Sumeyye Suri,18

Jaime E. Pineda,19 Joao Alves,16, 20 Jesse R. Feddersen,10 Jens Kauffmann,21 andPeter Schilke18

1Dept. of Astronomy, Yale University, New Haven, Connecticut 06511, USA2Cahill Center for Astronomy and Astrophysics, California Institute of Technology, 249-17, Pasadena, CA 91125, USA3East Asian Observatory, 660 N. A’ohoku Place, Hilo, Hawaii, 96720, s.mairs@eaobservatory.org4Universitat Heidelberg, Zentrum fur Astronomie, Albert-Ueberle-Str. 2, 69120 Heidelberg, Germany5Universitat Heidelberg, Interdisziplinares Zentrum fur Wissenschaftliches Rechnen, INF 205, 69120 Heidelberg,Germany

6Department of Astrophysical and Planetary Sciences, University of Colorado, Boulder, Colorado, USA7Institut de Ciencies del Cosmos, Universitat de Barcelona, IEEC-UB, Martı i Franques 1, E08028 Barcelona, Spain;ppadoan@icc.ub.edu

8ICREA, Pg. Lluıs Companys 23, 08010 Barcelona, Spain9Jodrell Bank Centre for Astrophysics, School of Physics and Astronomy, University of Manchester, Oxford Road,Manchester M13 9PL, UK

10Department of Astronomy, Yale University, New Haven, Connecticut 06511, USA11Joint ALMA Observatory, Alonso de Cordova 3107 Vitacura, Santiago, Chile12National Radio Astronomy Observatory, 1003 Lopezville road, Socorro NM, 8780113Departmento de Astronomıa, Facultad de Ciencias Fısicas y Matematicas, Universidad de Concepcion, Concepcion,Chile

14Max-Planck-Institute for Astronomy, Konigstuhl 17, 69117 Heidelberg, Germany15Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, USA16Department of Astrophysics, University of Vienna, Turkenschanzstrasse 17, 1180 Wien, Austria17Department of Earth and Space Sciences, College of the Canyons, Santa Clarita, CA 9135518I. Physikalisches Institut, Universitat zu Koln, Zulpicher Str. 77, D-50937 Koln, Germany19Max-Planck-Institut fur extraterrestrische Physik, Giessenbachstrasse 1, 85748 Garching, Germany20Radcliffe Institute for Advanced Study, Harvard University, 10 Garden Street, Cambridge, MA 02138, USA21Haystack Observatory, Massachusetts Institute of Technology, 99 Millstone Road, Westford, MA 01886, USA

ABSTRACT

We have investigated the formation and kinematics of sub-mm continuum cores in the Orion Amolecular cloud. A comparison between sub-mm continuum and near infrared extinction shows acontinuum core detection threshold of AV ∼ 5-10 mag. The threshold is similar to the star formationextinction threshold of AV ∼ 7 mag proposed by recent work, suggesting a universal star formationextinction threshold among clouds within 500 pc to the Sun. A comparison between the Orion Acloud and a massive infrared dark cloud G28.37+0.07 indicates that Orion A produces more densegas within the extinction range 15 mag . AV . 60 mag. Using data from the CARMA-NRO OrionSurvey, we find that dense cores in the integral-shaped filament (ISF) show sub-sonic core-to-envelope

2 Kong et al.

velocity dispersion that is significantly less than the local envelope line dispersion, similar to whathas been found in nearby clouds. Dynamical analysis indicates that the cores are bound to the ISF.An oscillatory core-to-envelope motion is detected along the ISF. Its origin is to be further explored.

Core Kinematics 3

1. INTRODUCTION

It is generally accepted that stars form ingravitationally bound dense cores within molec-ular clouds. Understanding the formation andevolution of cores is therefore crucial to thestudy of star formation (SF). For example, ob-servations (e.g., Lada et al. 2010; Heidermanet al. 2010) and theory (e.g., McKee 1989) implythat the onset of the star formation process oc-curs at the transition between the cloud photo-dissociation region (PDR) and self-shielded re-gions, where the cloud extinction is at leastAV ∼ 7 mag. It seems reasonable to expect asimilar extinction threshold for core formation(CF). Other outstanding problems in SF con-cern the efficiency of the overall process (Kenni-cutt & Evans 2012), and the effect of core kine-matics on the dynamics of emerging clusters.

At a distance of 400 pc, the Orion A cloud pro-vides an opportunity to investigate the early for-mation of cores in details. In particular, OrionA enables a study of the effects of feedbackfrom new young massive stars on their birth en-vironment. Here, we combine newly acquired12CO(1-0), 13CO(1-0), and C18O(1-0) molecularline data cubes of extended regions on the OrionA cloud from the CARMA-NRO Orion Survey(Kong et al. 2018c, hereafter K18) with a varietyof complementary surveys at other wavelengthsto address these issues. The new high dynamicrange images recover spatial scales from ∼8′′

(0.015 pc) to ∼2.5◦ (18 pc).As noted above, it would be useful to com-

pare the extinction thresholds between SF andCF. Millimeter continuum observations of theOphiuchus and Perseus molecular clouds showevidence of a CF threshold at AV ∼ 5-9 mag(Johnstone et al. 2004; Young et al. 2006; Enochet al. 2006; Kirk et al. 2006). However, see Clark& Glover (2014) for a critical discussion. TheOrion A cloud can be a unique test of the extinc-tion threshold of CF because, unlike the othernearby clouds mentioned above, it has strong

feedback from young massive stars. It is im-portant to show if the CF threshold varies indrastically different environments.

Above the extinction threshold, the SF effi-ciency is of order 10% (Lada et al. 2010; Hei-derman et al. 2010). This means that SF isstill limited even when the cloud extinction isabove the so-called threshold. This has longbeen known as a challenge in SF (Kennicutt &Evans 2012). In our context, it is interesting tosee whether the low efficiency is already evidentin the formation of cores. Unlike the observa-tions of SF efficiency where one counts youngstellar objects (YSOs), and might miss manyof those which may have left their birthplace,the star-forming cores are still embedded in thehost cloud and maintain the primordial infor-mation of the core/star formation. This allowsa more thorough examination of the possible de-pendence of the SF efficiency on natal environ-ment.

Core kinematics bring another type of insightsto our understanding of CF. In particular, core-to-core kinematics, potentially shaped by thehost cloud, may have a notable impact on thedynamics of the forthcoming star cluster. Withour newly acquired molecular line data from theCARMA-NRO Orion Survey (K18), we are ableto investigate the core kinematics in great de-tails. Throughout the paper, we follow K18 anduse a distance to Orion A of 400 pc (see Kounkelet al. 2018; Brown et al. 2018, for the latest dis-cussions). In the following, we introduce ourdata collection in §2, and our results and analy-sis in §3. Finally, we present the discussion andconclusions in §4 and §5, respectively.

2. DATA COLLECTION

2.1. CARMA-NRO Orion Data

In this paper, we use the molecular line datafrom the CARMA-NRO Orion Survey (see K18for more details). The survey produced cubesfor 12CO(1-0), 13CO(1-0), and C18O(1-0). We

4 Kong et al.

combined single-dish data from the Nobeyama45m telescope (NRO45) and interferometer datafrom the Combined Array for Research in Mil-limeter Astronomy (CARMA), producing spec-tral maps that recover spatial scales from ∼8′′

(0.015 pc) to ∼2.5◦ (18 pc). The velocity reso-lution is 0.25 km s−1 for 12CO(1-0); 0.11 km s−1

for 13CO(1-0) and C18O(1-0).

2.2. Near-infrared Extinction Map

We use the near infrared (NIR) extinctionmap of Orion A recently published by Meingastet al. (2018, hereafter Meingast18) to probe thecloud’s dust column density distribution at aresolution of 1′(or 2400 au), and a pixel scaleof 30′′ (see Figure 1). For our study, we maskthe vicinity of the Orion KL region to avoid un-derestimated extinctions. We convert the mapfrom AK to AV by multiplying by a factor of8.93 (Rieke & Lebofsky 1985). In the following,the extinction map is masked to have the samecoverage as the continuum image.

2.3. Sub-mm Continuum Image and CoreCatalog

We obtain the 850 µm continuum image fromLane et al. (2016, hereafter Lane16) for ourstudy (publicly available at https://doi.org/10.11570/16.0008). The image was acquiredas part of the JCMT Gould Belt Survey (Ward-Thompson et al. 2007; Salji et al. 2015; Mairset al. 2016). The 850 µm map has a resolution of14.6′′ (∼0.03 pc). Below we investigate the re-lation between continuum detection and cloudcolumn density traced by the Meingast18 ex-tinction map. First, we compare the continuumcore catalog defined by Lane16 (the Getsourcescatalog) with the Meingast18 extinction map1.Second, we carry out a pixel-by-pixel compari-son between the Lane16 sub-mm continuum im-age and the Meingast18 extinction map without

1 One caveat is that the cores were defined with 14.6′′

resolution while the extinction map resolution is 1′.

defining “cores”. Before the second comparison,we convolve the dust continuum image with aGaussian kernel to have a final spatial resolutionof 1′ and regrid it to the Meingast18 extinctionmap. The smoothed image has an rms noise ofσ850 = 10 mJy per 1′ beam.

Lane16 used an automask data reductiontechnique for the continuum image. This isa reduction strategy wherein astronomical flux(as opposed to atmospheric flux or noise) isidentified automatically in the map-making pro-cedure based on pixel SNR and the results ofiteratively calculated noise models. For furtherdetails on the data reduction, see Chapin et al.(2013), Mairs et al. (2015), and Lane16. As aresult, sources with sizes between 2.5′ and 7.5′

(0.3 pc and 0.9 pc at a distance of 400 pc) donot have robust flux measurements. Sourcessmaller than 2.5′ likely have robust flux mea-surements. Based on the work in Mairs et al.(2015), Lane16 indicated that strong, compactsources are well recovered, while faint, diffusesources suffer more from flux loss. Since the goalis to target cores that are more centrally concen-trated, i.e., more likely to develop protostars,we consider that the sub-mm continuum imagesatisfactorily represents the emerging cores inOrion A.

2.4. GAS Ammonia Core Catalog

We have obtained the NH3 data from theGAS survey (Friesen et al. 2017; Kirk et al.2017, hereafter Kirk17), which is publicly avail-able at https://dataverse.harvard.edu/

dataverse/GAS_DR1. Kirk17 selected a sam-ple of cores from the larger sample of Lane16based on the detection of NH3 (observed withthe Green Bank Telescope in the GAS survey,see Friesen et al. 2017) and studied their phys-ical properties. In total, 237 continuum cores,most of them in the ISF, were included in theirstudy. Kirk17 performed hyperfine line fittingto the NH3 inversion lines and estimated theNH3 core velocity vNH3 and the kinetic tem-

Core Kinematics 5

208°209°210°211°212°213°Galactic Longitude

-20.0°

-19.5°

-19.0°

Gala

ctic

Latit

ude

2 pc0

1

2

345678

Figure 1. Greyscale: Meingast18 K-band extinction map, following the same coverage as the Lane16 850µm map. Yellow contours: Lane16 850 µm continuum image smoothed to 1′ resolution. The contour levelis Σmm = 0.044 g cm−2 (see §3.2). Blue contours: Meingast18 K-band extinction equivalent to AV = 20mag. Cyan contours: Lane16 Getsources cores. White contour: Lane16 map coverage. Green polygon: Thedefinition of ISF in Figure 10 (see §3.4). The red dashed ellipse shows the region that is masked for thedetection probability study. The 1′ beam size is shown at the lower-left corner. The vertical dashed linesplits the cloud into two regions with different level of feedback from massive stars. See §3.2.

perature based on the ammonia emission. TheGAS survey resolution is 32′′, larger than someof the sub-mm cores in Lane16. Kirk17 (seetheir §3.2) have argued that the NH3 traces thedense cores reasonably well. In this paper, weadopt the assumption that the dense core line-of-sight velocity is traced by the Kirk17 NH3

line fitting result.

3. RESULTS AND ANALYSES

3.1. Sub-mm continuum detection

Figure 2(a) shows the continuum detectionprobability in Orion A. First, we investigatehow likely the Lane16 cores are detected at agiven AV . We make AV bins every 2 mag. Ineach bin, the detection probability of the coresPc,850 µm is defined as the number of pixels thatcontain cores (Nc) divided by the number of to-tal pixels in the extinction bin Ntot, i.e.,

Pc,850 µm = Nc/Ntot. (1)

The error is estimated as

σP = [Pc,850 µm(1− Pc,850 µm)/Ntot]0.5. (2)

This estimation gives zero error when Pc,850 µm

is 0 or 1 (see discussions in Kong et al. 2018a).Figure 2(b) zooms in to 1 mag < AV < 15 mag,with a bin size of 1 mag.

Figure 2(a) shows that the detection proba-bility of the Lane16 continuum cores remains 0for AV ≤ 5 mag. At AV > 5 mag, the prob-ability increases monotonically until AV ∼ 20mag, after which it has a shallower slope. Inpanel (b), the core detection (histogram)shows an increase after AV ∼ 7 mag. Fol-lowing Mairs et al. (2016, figure 7), we make acumulative distribution function (CDF) for thecloud and core extinction, shown in the panel(c). The core CDF remains approximately flatuntil a turnover at AV ∼ 9 mag. Based onthese, we assign a value to the core detectionthreshold in the range AV ∼ 5 − 10 mag. Thisis consistent with findings in Ophiuchus (9 mag,Young et al. 2006) and Perseus (5 mag, Enochet al. 2006). As a massive star-forming cloud,Orion A does not show a significantly differ-ent extinction threshold of core formation com-pared to other lower-mass clouds. Moreover,these results are also consistent with Goldsmith

6 Kong et al.

0 10 20 30 40 50 60

0.0

0.2

0.4

0.6

0.8

1.0su

b-m

md

etec

tion

pro

bab

ility

(a)pixel SNR ≥ 3pixel SNR ≥ 5pixel SNR ≥ 10core

2 4 6 8 10 12 14−0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

sub-m

md

etec

tion

pro

bab

ility

(b)pixel SNR ≥ 3pixel SNR ≥ 5pixel SNR ≥ 10core

2 4 6 8 10 12 14cloud AV (mag)

0.0

0.2

0.4

0.6

0.8

1.0

frac

tion

alcu

mu

lati

vem

ass

(c)

cloud

cores

Figure 2. Continuum detection probability forOrion A. (a): The black histogram shows the coredetection probability as a function of AV . Thecolor points are pixel detection probability func-tions. The continuum detection thresholds are in-dicated in the top-left legends. The vertical dashedline indicates the potential break. The bin size is2 mag in AV . (b): Same as panel (a), but zoomedin for 1 < AV < 15 mag. The bin size is 1 mag,compared to 2 mag for panel (a). (c): Cumulativedistribution of extinction for the cloud (blue) andthe cores (red). For each AV , the distribution isthe fraction of pixels/cores with extinction greaterthan AV .

et al. (2008); Lada et al. (2010); Heidermanet al. (2010) who investigated the SF extinctionthreshold using YSOs. The similarity betweenthe core formation threshold and the SF thresh-old is not unexpected since stars form in densecores. These comparisons potentially suggest auniversal SF extinction threshold of AV = 5−10mag in the clouds within 500 pc of the Sun.

It is also important to study continuum detec-tions that are not identified as cores. They maybe the precursors of cores as they become cen-trally peaked during fragmentation. We showin Figure 2(a)(b) the pixel-based 850 µm de-tection probability function (DPF). The pixeldetection probability is the ratio between thenumber of extinction pixels that contain 850 µmdust emission detection and the number of totalpixels within each extinction bin. Three contin-uum detection thresholds are applied, namely,SNR≥3 (red); SNR≥5 (green); SNR≥10 (blue).The “detected” pixels defined this way may in-clude noisy spikes in the image. Using a higherthreshold can reduce a false detection at the ex-pense of missing real emission.

As expected, the pixel-based continuum de-tection probability is larger than the core-basedprobability, since the cores are defined with asubset of detected pixels (Lane16). In fact, a no-table number of the detected continuum pixelsare not included in the cores defined by Lane16Getsources catalog. As described in their ap-pendix A.1.1, Lane16 set a threshold SNR = 7for core detection. Further, they smoothed the850 µm image and removed sources that did notappear significant, along with those rejected bythe Getsources method. As a result, the num-ber of “reliable” cores was reduced from 1178 to919 (Lane16 final catalog). Figure 2 of Lane16shows that only a fraction of the emission wasgrouped into cores.

As shown in Figure 2(a)(b), the DPF risesfrom 0 to close to 1 from AV = 0 mag toAV ∼ 50 mag. At a given extinction, only a

Core Kinematics 7

fraction of the pixels are detected in continuumemission, caused by a combination of low tem-perature, spatial filtering effect, and/or opacityvariation. At sub-mm and mm wavelengths, theflux density is approximately proportional tothe column density and the temperature. Theline-of-sight column temperature may have amajor impact on the detection above a giventhreshold. The cores are likely heated by con-traction and/or an embedded protostar (see alsodiscussions in Kong et al. 2018a). Spatial filter-ing of large spatial structures helps in pickingout centrally peaked dense cores. Meanwhile,the extinction traces the total column density.Therefore, we argue that the continuum detec-tion pinpoints the forming dense cores embed-ded in the extinction column density.

The DPF thus shows that only a fractionof the cloud is forming stars at a given time,consistent with the theoretical expectation in,e.g., Krumholz & McKee (2005). An inter-esting feature in Figure 2(a) is the apparentbreak at AV ∼ 20 mag (vertical dashed line).For the DPF with SNR≥10, a linear re-gression for 10 . AV . 20 mag gives aslope of 0.024(±0.001), while the slope for20 . AV . 30 mag is 0.012(±0.001). Inter-estingly, AV = 20 mag is also where the coredetection slope becomes shallower. For refer-ence, Figure 1 shows the regions with AV = 20mag with blue contours. It remains to be seenif the 20 mag break is detected in other clouds.

The total mass of the Lane16 cores is ∼ 1400M�. We follow Lane16 equation (1) for the masscalculation. This was also adopted by Kirk17.Note this number has been scaled down by a fac-tor of 0.8 compared to the Lane16 and Kirk17core mass because we use 400 pc as the clouddistance while they used 450 pc. The total cloudmass converted from the Meingast18 extinctionmap is ∼ 36,000 M� within the Lane16 mappedarea. The Lane16 core mass fraction is therefore∼ 4.0% of the mass estimated from the extinc-

tion map. If we include all Lane16 sub-mm con-tinuum flux above a SNR of 5, the mass fractionincreases to ∼ 14%2.

3.2. Comparison with IRDC G28.37+0.07

To investigate the dense gas emergence in dif-ferent environments, we compare the Orion ADPF with the DPF from the infrared dark cloud(IRDC) G28.37+0.07 (hereafter G28). Konget al. (2018a) studied the DPF in IRDC G28with ALMA 1.3 mm continuum data. Althoughthe DPF in Orion A is derived with 850 µmcontinuum, the comparison is still meaningfulonce we apply the same physical threshold. Forcomparison, the physical resolution is 0.05 pc inthe IRDC G28 study (Kong et al. 2018a). Themaximum detectable scale was 0.5 pc. In IRDCG28, assuming a temperature of 15 K and vol-ume density of 105 cm−3, the Jeans length is∼ 0.1 pc. We also smooth the G28 data to aresolution of 0.12 pc, in order to match the lin-ear resolution of the IRDC data with that ofthe Orion A study. The fiducial 1.3 mm contin-uum detection threshold in mass surface densitywas set to Σmm = 0.044 g cm−2 in Kong et al.(2018a). To match this number in Orion A,we compute the mass surface density followingequation (1) in Lane16 and adopting the sameassumptions of dust temperature and opacity.The threshold is shown as the yellow contoursin Figure 1.

Figure 3(a) shows the comparison of DPFs be-tween Orion A and IRDC G28. We use thesame bin size of 0.02 g cm−2 for both clouds,following Kong et al. (2018a). Note that thisnumber is converted from extinction (hereafterdenoted Σex), which is different from the sub-mm continuum detection threshold (also in gcm−2, but from emission). The highest extinc-

2 The 15 K assumption for the dust temperaturearound Orion KL is probably an underestimate, whichresults in an overestimate of the mass fraction. Exclud-ing the KL region, we find a mass fraction of ∼ 4.0%.

8 Kong et al.

0.0 0.2 0.4 0.6 0.8Σ (g cm−2)

0.0

0.2

0.4

0.6

0.8

1.0

cont

inu

um

det

ecti

onp

rob

abili

ty

(a)

Orion A 15 KOrion A 20 KIRDC G28G28 smoothed

0 42 85 128 171cloud AV (mag)

0.00 0.05 0.10 0.15 0.20 0.25 0.30Σ (g cm−2)

0.0

0.2

0.4

0.6

0.8

1.0

cont

inu

um

det

ecti

onp

rob

abili

ty

(b) l>210◦l<210◦

0 10 21 32 42 53 64cloud AV (mag)

Figure 3. (a): Continuum pixel detection prob-ability for Orion A (blue) and IRDC G28 (black).The red points are from smoothed G28 data thatmatches the physical resolution of the Orion Adata. The cyan points are the same as the blueones except that we use 20 K to calculate the de-tection flux threshold. See §3.2. (b): Continuumpixel detection probability for two sub-regions inOrion A (divided at l=210◦, indicated by the ver-tical dashed line in Figure 1).

tion traced by the Meingast18 map is about 0.3g cm−2. Recently, Stutz (2018) used Herschelfar infrared and APEX 870 µm data to derivea column density map in Orion A. The high-

est surface density reaches ∼ 3 g cm−2 in theOrion KL region. The near infrared extinctiondoes not work well in this region. However, thenumber of pixels in this region only accounts for∼0.1% of the Lane16 coverage. We ignore thesepixels by removing them from the comparison(red dashed ellipse in Figure 1).

As shown in Figure 3(a), the Orion A cloudshows more continuum detection of dense gaswith Σmm ≥ 0.044 g cm−2 than IRDC G28.The difference appears within 0.08 . Σex .0.28 g cm−2. For this we followed Lane16 andKirk17 and assumes dust temperature of 15 K.Kong et al. (2018a) assumed 20 K for their dustcontinuum emission because they argued thatthe cores were mostly protostellar. For consis-tency, we also calculate a new DPF assuminga dust temperature of 20 K in Orion A (cyandots in Figure 3(a)). A higher dust tempera-ture requires a higher continuum flux to reachthe same mass surface density threshold of 0.044g cm−2, therefore the cyan detection probabil-ity becomes lower. Nonetheless, the Orion ADPF is still higher than the G28 DPF within0.08 . Σex . 0.28 g cm−2. A two-sample KStest gives a KS statistic of 0.8 with a p-valueof 0.0012, meaning the null hypothesis that thetwo samples are drawn from the same distribu-tion can be rejected at a significance level below1% (confidence over 99%).

The main source of uncertainty is likely thetemperature assumption. For a given mass, ahigher dust temperature indicates a higher con-tinuum flux. Therefore, we need to set a higherflux threshold for the same mass threshold. Thisgives rise to lower detection probability, whichis illustrated by the blue and cyan points in Fig-ure 3(a). Lane16 pointed out that the high dusttemperature from Herschel results (Lombardiet al. 2014) likely trace the dust in lower densityregions along the line-of-sight and thus underes-timate the core mass. In Figure 4, we show theKirk17 NH3 kinetic temperature map. Again,

Core Kinematics 9

208.5°209.0°209.5°210.0°Galactic Longitude

-19.8°

-19.6°

-19.4°

-19.2°

Gala

ctic

Latit

ude

2 pc5

10

15

20

25

30

Figure 4. Kirk17 NH3 gas kinetic temperature map. The color bar ranges linearly from 5 K to 30 K. Theyellow contours and the red ellipse are the same as in Figure 1.

the yellow contours show the regions with a con-tinuum surface density equal to or greater than0.044 g cm−2. The regions enclosed by the yel-low contours in Figure 3 are dominated by tem-peratures below 20 K.

The effect of spatial filtering may be impor-tant when comparing the DPF in both clouds.Both the ALMA data and the JCMT data fil-ter out emission from structures at large spa-tial scales. Therefore, the continuum emissionin both maps detect relatively compact struc-tures, that presumably trace the dense gas inti-mately involved in the core formation process.In the JCMT 850 µm image, structures largerthan 2.5′ (0.3 pc) are not robustly recovered. Inthe ALMA 1.3 mm image, the maximum sensi-ble scale is 12′′ (0.3 pc, which is 0.6 times themaximum scale that corresponds to the short-est baseline, see Kong et al. 2018a). Therefore,we argue that both DPFs are tracing similardense structures, and Orion A is more capableof forming dense gas/cores than the IRDC G28.

One significant difference between G28 andOrion A is that the former is likely forming thevery first generation of stars in the mapped re-

gion. In particular, the area studied by Konget al. (2018a) is mostly dark up to 70 µm. Asopposed to Orion A, there is no feedback (radia-tion, spherical wind/bubble) from massive starsin IRDC G28, although massive stars may even-tually form there (Zhang et al. 2015; Kong et al.2018b).

To test if the high DPF in Orion A is related tofeedback, we split the Orion A map in two, indi-cated by the vertical dashed line along l=210◦ inFigure 1. The region on the east side (l>210◦) isless influenced by feedback from the Trapeziumcluster compared to the region on the west side(l<210◦, Bally 2008). As shown in Figure 3(b),the high sub-mm continuum detection probabil-ity in Orion A is dominated by the l<210◦sub-region, suggesting that strong feedback may in-deed result in higher values in the DPF for thesame cloud extinction (see §4.1).

3.3. Core-to-envelope velocity

We investigate the relative motion betweenthe dense cores traced by NH3 (§2.4) and theambient cloud traced by CO isotopologues(§2.1). Similar studies include Walsh et al.

10 Kong et al.

(2004) and Kirk et al. (2010). We first con-volve the CARMA-NRO cubes to 32′′ angularresolution to match the GAS data. Then, foreach core we extract a spectrum averaged overthe 32′′ circular area centered on the Kirk17core position. We fit a Gaussian to the spectrain order to derive the bulk gas velocity com-ponent (1D) along the line of sight, with theoptically thin assumption. In fact, wehave shown in K18 that more than 99%of the 13CO pixels have τ <1 (see their§5.2). The maximum core optical depthfor the 13CO(1-0) line is 1.12 and 98% ofthe cores have optical depth < 1.

0

20

40

60

80

Tm

b(K

)

12CO(1-0) (a)

core5

0

10

20

30

40

Tm

b(K

)

13CO(1-0) (b)

0 2 4 6 8 10 12 14 16vLSR (km s−1)

0.0

2.5

5.0

7.5

Tm

b(K

)

C18O(1-0) (c)

Figure 5. 12CO(1-0) (top), 13CO(1-0) (mid-dle), and C18O(1-0) (bottom) spectra for core 5 inKirk17. The baselines are shown as the black hori-zontal dashed lines. The blue vertical dashed linesindicate vNH3 . The green curves show the Gaussianfit to the spectra (only for 13CO and C18O), withthe green vertical dashed lines showing the centroidvelocities vgauss.

Figure 5 shows the spectra of core number 5in the Kirk17 sample. We show the spectra of12CO(1-0), 13CO(1-0), and C18O(1-0) in black.In order to perform the Gaussian fit, we visually

0

20

40

60

Tm

b(K

)

12CO(1-0) (a)

core41

0

10

20

30

Tm

b(K

)

13CO(1-0) (b)

0 2 4 6 8 10 12 14 16vLSR (km s−1)

0

2

4

Tm

b(K

)

C18O(1-0) (c)

Figure 6. Same as Figure 5, but for core 41. Thegreen dashed line shows the centroid velocity vgaussof the Gaussian component that is closest to vNH3 .The green dotted curves show the fitting to eachGaussian component.

identify the number of Gaussian components foreach spectrum and provide the initial guess foramplitude, centroid velocity, and velocity dis-persion for each spectrum. Then, we use thescipy.optimize.curve fit function from thepython package Scipy to perform single-/multi-Gaussian fitting to 13CO and C18O spectra. Ifthere is only one component, we take the cen-troid velocity vgauss as the envelope velocity. Ifmore than one velocity component is identified,we take the one that is nearest to vNH3 as theenvelope velocity3, similar to the study by Kirk

3 This could introduce a bias toward a small differ-ence between core and envelope velocity. However, weargue that this is unlikely. Assuming the intensity ratioNH3/13CO is proportional to the abundance of NH3, alarge difference between the velocity centroids of NH3

and 13CO would result in relatively high NH3 abun-dance since the peak of the NH3 component would becoincident with the 13CO line wing. Moreover, we selectcores with only one CO velocity component and derivethe core-to-envelope velocity dispersion, σce, and obtain

Core Kinematics 11

et al. (2007, see their discussions in section 5and figure 6).

Core 5 (Figure 5) shows one Gaussian com-ponent in 13CO and C18O spectra. However, asignificant fraction of the cores show complexvelocity features. In Figure 6, we show anotherexample, core 41. This core has roughly threevelocity components in the 12CO(1-0) line andtwo components in 13CO(1-0) and C18O(1-0).vNH3 of this core does not corresponds to any ofthe carbon monoxide line components. Basedon our visual identification, 100 out of the 237cores from Kirk17 show multiple velocity com-ponents in either the 13CO or the C18O spec-trum (or both). 12CO spectra typically showmore complex profiles.

To estimate possible systematic difference be-tween the dense core velocity vNH3 and the enve-lope gas velocity, we calculate for each core the1D core-to-envelope velocity vce ≡ vNH3 - vgaussusing the 13CO(1-0) and C18O(1-0) line Gaus-sian fitting results. In the top panels of Fig-ure 7, we show the distributions of vce(

13CO)and vce(C

18O) (black bins), respectively. Thebin size is 0.1 km s−1. We then fit a Gaus-sian to the vce distribution. We can see the1D core-to-envelope Gaussian dispersion σce is0.20±0.01 km s−1 for C18O and 0.30±0.01 kms−1 for 13CO. We also apply the maximum like-lihood method (using scipy.stats.norm.fit)to estimate σce, which gives σce(

13CO) = 0.33km s−1 and σce(C

18O) = 0.25 km s−1 (removingthe outlier beyond 2 km s−1; if not, σce = 0.30km s−1).

We investigate the effect of varying envelopesizes by extracting envelope spectra from anumber of circular regions centered on the corebut with different radii. For this purpose we

σce(13CO) = 0.28 km s−1 and σce(C

18O) = 0.17 kms−1. This is very similar to the values obtained usingthe entire sample of cores, hence we believe that includ-ing multi-component spectra does not produce any biasin our results.

use the original CARMA-NRO Orion Surveydata cubes (i.e., before smoothing to 32′′ reso-lution). At each core location, we extract 13COand C18O spectra from circular regions with di-ameters 8′′, 16′′, 32′′, 64′′, and 128′′, i.e., factorsof 2 increase from the beam size (8′′ correspondsto a physical scale of 0.015 pc). The respectivefitting results of the core-to-envelope dispersionfor 13CO are 0.29 km s−1, 0.30 km s−1, 0.30 kms−1, 0.34 km s−1, 0.39 km s−1; for C18O theyare 0.26 km s−1, 0.25 km s−1, 0.25 km s−1, 0.26km s−1, 0.32 km s−1. As we compare the corevelocity to that of the envelopes averaged overa larger area, the velocity dispersion becomesslightly higher.

In the top panels of Figure 7 we indicate thesound speed, calculated using a gas tempera-ture of 18 K (following appendix A1 in Orkiszet al. 2017), by the vertical dotted lines. Figure8 shows the distribution of core gas kinetic tem-peratures derived from NH3 data (Kirk17). Themean temperatures are 18.3 and 17.8 K, respec-tively, while the minimum and maximum tem-perature are 11.3 and 29.3 K, respectively Weadopt 18 K as the representative temperaturefor the sound speed calculation (which resultsin a value of 0.31 km s−1). For C18O, the soundspeed is at 1.55σce, which from the Gaussian fit-ting to vce(C

18O) would imply that about 88%of the cores have relative motion slower than thelocal sound speed (note that from the data, theactual fraction of cores with vce(C

18O) < 0.31km s−1 is 78%.) For 13CO, the sound speed is at1.0σce. Based on the Gaussian fitting, this im-plies that 68% of the cores have relative motionslower than the local sound speed (the actualfraction of cores with vce(

13CO) < 0.31 km s−1

is 66%). Note that the velocity dispersion andsound speed discussed here are both 1D quan-tities.

Figure 9 shows the distribution of line ve-locity dispersions for 13CO(1-0) and C18O(1-0) (black bins). Again, we indicate the sound

12 Kong et al.

10

20

30

40

50

num

ber

13CO(1-0)allfit σ = 0.30± 0.01YSOfit σ = 0.27± 0.02no YSOfit σ = 0.31± 0.02

−3.0 −2.0 −1.0 0.0 1.0 2.0 3.0vce (km s−1)

42

21

0

−21

−42

−63

Off

sets

(arc

min

)

10

20

30

40

50

num

ber

C18O(1-0)allfit σ = 0.20± 0.01YSOfit σ = 0.22± 0.01no YSOfit σ = 0.19± 0.01

−3.0 −2.0 −1.0 0.0 1.0 2.0 3.0vce (km s−1)

42

21

0

−21

−42

−63

Off

sets

(arc

min

)

Figure 7. Left: Top panel: Histograms of core-to-envelope velocity differences vce(13CO). The black

histogram shows the distribution for the entire Kirk17 core sample. The black dashed curve shows theGaussian fit to the black histogram. The blue histogram and dashed curve correspond to the cores withYSO; the yellow without YSO. The dispersion values obtained from Gaussian fits to the histograms areshown in the upper left corner. The vertical dotted lines show the sound speed. Bottom panel: The graybackground shows the PV-diagram along the orange curve in Figure 10. The zero offset is the big red dot onthe orange curve between OMC-1 and ONC. Positive offset is toward the north (OMC-2/3 direction). Notethat 21 arcmin corresponds to ∼2.5 pc at 400 pc. At each offset, the PV-diagram zero velocity is definedat the intensity peak. The gray contour denotes the half maximum at each offset. The yellow dots denotethe starless cores and the blue dots denote the cores with YSO. The offset (y-axis) is along the ISF spine(orange curve). The empty red circles are offset bins every 6′. Their vce values are averaged within the bins.Right: Same as left, but for vce(C

18O).

Core Kinematics 13

10 20 30 40Tk,NH3

(K)

0

10

20

30

40

50nu

mb

er

Figure 8. Histogram of core kinetic temperaturederived from NH3 inversion lines. The data aretaken from Kirk17. The vertical dashed line showsthe median temperature (18 K) used to computethe sound speed.

speed as the vertical dotted line. The figureshows that the majority of cores are located inregions where the carbon monoxide gas line dis-persion is supersonic. Therefore, the core-to-envelope velocity dispersion σce is significantlysmaller than the local line width. In the bot-tom panels of Figure 7, we show the core-to-envelope velocities on top of the PV-diagramalong the ISF ridge defined in K18 and Figure10. As can be seen, for the vast majority ofcores, the core-to-envelope velocities vce(

13CO)and vce(C

18O) are predominantly smaller thanthe velocity FWHM at the position of the core(shown as gray contours in Figure 7), confirmingthe results from the Gaussian fittings. Througha study of the kinematics in the Orion NebulaCluster (ONC), Tobin et al. (2009) showed thatthe stars and 13CO gas are kinematically tied,consistent with our results for dense cores.

We made separate histograms of σce for coreswith and without YSOs in order to investigatewhether there is a difference between pre- andprotostellar cores as a result of evolution. TheYSO-core assignment is given by Lane16, wherethey checked against catalogs from Megeathet al. (2012, Spitzer observation) and Stutz et al.

10

20

30

40

50

num

ber

13CO(1-0) Gauss2nd moment

0.0 0.5 1.0 1.5 2.0σ (km s−1)

0

10

20

30

40

50

num

ber

C18O(1-0) Gauss2nd moment

Figure 9. Histograms showing the distributionof the line dispersion in the 13CO (top panel) andC18O (bottom panel) spectra from cores. The blackhistograms show the distribution of dispersion val-ues obtained from gaussian fits to the spectra, whilethe blue histograms show the distribution of valuesobtained from the second moment of the spectra.The vertical dotted line shows the sound speed cal-culated at 18 K.

(2013, Herschel observation). In Figure 7, weshow vce distribution for protostellar cores asblue histograms; starless cores as yellow his-tograms. The vce dispersions between the twopopulations are different by 10%. The vce cen-troid for the starless cores is 0.03±0.02km s−1 in the 13CO fitting; for the pro-tostellar cores 0.13±0.02 km s−1. In theC18O fitting, the vce centroid for the star-less cores is 0.04±0.01 km s−1; for the pro-tostellar cores 0.08±0.01 km s−1. The coreswith YSOs tend to have an excess of positivecore-to-envelope velocities. However, note thelimited number of cores with associated YSOs(see the blue dots in Figure 7).

14 Kong et al.

3.4. Filament and core dynamics

We re-analyze the filament virial status in thecontext of the model by Fiege & Pudritz (2000).In this model, the virial velocity for an unmag-netized filamentary cloud is related to the massper unit length ml by:

σvir =

√Gml

2, (3)

where G is the gravitational constant.Fiege & Pudritz (2000) used the Orion A 13CO

data with a resolution of 1.7′ from Bally et al.(1987) to derive a value ofml = 355 M� pc−1 forthe ISF. We use the extinction map (Meingastet al. 2018) to estimate the ISF mass. Figure10 shows our definition of the filamentary re-gion. We define the entire ISF filament regionto roughly cover the Kirk17 core sample (greendashed polygon). The total mass of the filamentis ∼3835 M� and the total length is ∼12 pc.The length is derived from a visually-definedcurve that follows the ridge line of the ISF (orig-inally defined in figure 5 of K18, but extendedto the eastern end of the filament). The recentGAIA data show that the ISF is likely parallelto the plane of the sky (Großschedl et al. 2018),although with a few pc variation (Stutz et al.2018). Hence, we do not correct the length ofthe filament due to inclination effects. The ISFml we derive is 320 M� pc−1, a factor of 0.90smaller than Fiege & Pudritz (2000). This fallsin the ml range (125-800 M� pc−1) estimated byStutz & Gould (2016, see their figure 5). Ourestimated virial velocity for an unmagnetizedfilament is σvir ∼0.83 km s−1.

Recall from §3.3 that the core-to-envelope ve-locity dispersion σce for 13CO and C18O is 0.30and 0.20 km s−1, respectively. Applying a fac-tor of

√3, the 3D dispersion for 13CO is 0.52 km

s−1 and for C18O is 0.35 km s−1. Both numbersare smaller than the σvir in the unmagnetizedscenario, indicating that the cores are bound tothe ISF filament. In the presence of magnetic

fields, σvir depends on both the poloidal field inthe filament and the helical field that wraps thefilament. More robust measurements of the fieldstrength are needed (note the recent progress byPattle et al. 2017; Tahani et al. 2018).

Note that the ISF has an overall velocity gra-dient and the core kinematics follow the trend.To illustrate this, we make PV-diagrams alongthe ISF for 13CO and C18O, shown in Figure11. The PV-cut follows the orange curve fromFigure 10. The offsets in Figure 11 are labeledas the red numbers in Figure 10. We also over-plot the cores on the PV-diagram. As one cansee, the ISF shows a wave-like velocity structure(see K18 figures 20-22; also see Stutz & Gould2016) and the cores closely follow the ISF kine-matics. The overall velocity distribution of theISF results in a super-virial core-to-core veloc-ity dispersion of 2.92 km s−1, which is differentfrom what was seen in Perseus (sub-virial core-to-core dispersion, Kirk et al. 2010; Foster et al.2015). While the final fraction of gas in Orion Aand Perseus that is deposited into stars (SF ef-ficiency) is unclear, the forthcoming star clusterin Orion A is less likely to be bound comparedto the cluster in Perseus. This emphasizes theimportance of the initial condition for star clus-ter formation.

4. DISCUSSION

4.1. Comparing core formation capabilitybetween Orion A and IRDC G28

Within the range 0.08 . Σex . 0.28 g cm−2,Orion A appears to be more capable of formingdense cores than G28 (Figure 3). In §3.2 wesuggested that this could be caused by SF feed-back in Orion A, although we did not indicatethe main feedback mechanism that could be re-sponsible for this. Outflows are unlikely thesource for the difference between Orion A andG28, as both clouds have widespread outflows(Davis et al. 2009; Bally et al. 2017; Kong et al.2019). The main difference likely comes from

Core Kinematics 15

208.5◦209.0◦209.5◦210.0◦

Galactic Longitude

−19.8◦

−19.6◦

−19.4◦

−19.2◦

Gal

acti

cL

atit

ud

e

0

21

42

-21

-42

-63

ONC

OMC1

OMC2

OMC3

OMC4

NGC1980L1641-N

2 pc0

1

2

3

4

5

6

7

8

AK

(mag

)

Figure 10. Cores on the Integral-Shaped Filament in Orion A. The gray-scale background image showsthe extinction map from Meingast et al. (2018). The blue circles mark the position of the Kirk17 cores. Thegreen dashed contour shows the filament defined for estimating ml in §3.4. The orange curve defines thePV cut following figure 5 of Kong et al. (2018c) but extending further to cover the entire core sample in theeast. The red dots along the curve indicate 3′ segments. The red numbers show the offsets in Figure 11.

the presence of massive star feedback in Orion A(the region of G28 studied in Kong et al. (2019)is not impacted by massive stars). In Figure4 we have shown that the dense gas in OrionA is not significantly heated, thus the heatingfrom ionizing photons is not likely dominatingthe difference between these two clouds. Wesuggest that pressure from the expanding HIIregion (Pabst et al. 2019), which possibly com-presses the cloud and generates more dense gas,may be responsible for the higher incidence ofdense cores in Orion A. Feddersen et al. (2018)reported expanding shells and bubbles in OrionA, which may also be important in generatingdense gas. Our results emphasize the importantrole of stellar feedback in SF.

4.2. ISF core group dynamics

The sub-sonic core-to-envelope kinematic fea-ture we detect in Orion A is seen in other nearbystar-forming clouds (i.e., Taurus, Perseus, Ophi-

uchus, see Walsh et al. 2004; Kirk et al. 2007,2010; Foster et al. 2015), although the OrionA cloud is quite different from these clouds interms of active massive SF and destructive feed-back. These suggest that dense molecular coresremain tied to their host clouds during their for-mation regardless of the environment.

The kinematics of cores can be useful in test-ing SF theories. For instance, in the competi-tive accretion model (Bonnell et al. 2001; Smithet al. 2009), protostars and the host cloud glob-ally collapse toward the cloud’s gravitationalpotential minimum. As shown by Bonnell et al.(2001), protostars compete for the gas throughtidal accretion and the relative velocity betweena protostar and the surrounding gas is sub-sonicthroughout the global collapse. However, it is tobe clarified how the cores seen in observationsare linked to the structures seen in the simu-lations. For instance, detailed radiative trans-fer modeling may be necessary to show if the

16 Kong et al.

−63−42−2102142Offsets (arcmin)

2

4

6

8

10

12

14

16

Vel

ocit

y(k

ms−

1 )

PV − diagram 13CO

0

5

10

15

20

25

30

K

−63−42−2102142Offsets (arcmin)

2

4

6

8

10

12

14

16

Vel

ocit

y(k

ms−

1 )

PV − diagram C18O

0

1

2

3

4

K

Figure 11. Position-velocity (PV) diagrams along the ISF using the 13CO (top) and C18O (bottom) data.The PV cut is defined as the orange curve along the ISF spine with a width of 3′(Figure 10). The zero-offsetposition is the large red dot between OMC1 and ONC in Figure 10. The positive offset is toward the westdirection (OMC-2/3). Note that 21 arcmin corresponds to ∼2.5 pc at 400 pc. The blue dots show theKirk17 cores. The velocity is different from Figure 7 bottom panels where the zero velocity isthe gas intensity peak.

tidal sphere in simulations corresponds to thedense cores in observation. The ISF is not eas-ily explained by a global collapse. On the otherhand, the fact that dense cores have low core-to-envelope velocity dispersion is a fundamen-tal prediction of the “colliding flow” model inPadoan et al. (2001). Cores do not move freelyin the envelope because they form in the post-shock gas at the intersection of colliding flows(i.e., in regions where the flow velocity stag-nates, as it is dissipated in shocks).

4.3. The dynamical status of the ISF

The work by Tobin et al. (2009) have shownthe similarity in the kinematics between thestars and 13CO gas in the ISF. Subsequently,Hacar et al. (2016); Stutz & Gould (2016); Get-man et al. (2019) explored the kinematic linkbetween YSOs and the cloud gas in Orion A.Together with our findings, these results sug-gest that SF is ongoing in the ISF and that thelow velocity dispersion in the protostars rela-tive to their local gas probably originates fromthe low core-to-envelope velocity reported here.However, the explanation of the origin and the

Core Kinematics 17

future of the ISF remains unclear, and varioushypotheses have been proposed.

Tobin et al. (2009) argued that the ISF cloudand stars are undergoing a global collapse to-ward the ONC. The basis for their argumentwas the observed velocity gradient along thedeclination direction. The gas kinematics andthe radial velocity of the young stars showa change toward red-shifted velocities in theOMC-2/3 regions north of the ONC (their fig-ure 3), while the southern ISF shows an approx-imately flat velocity distribution. Tobin et al.(2009) thus argued that the ISF is collapsingtoward the ONC, with the OMC-2/3 regions lo-cated on the near side but having red-shiftedvelocities.

With a higher resolution (30′′) N2H+ data,

Hacar et al. (2017) revisited the topic. Theyalso argued that the ISF is undergoing gravita-tional collapse toward the ONC. However, theirpicture supports a collapse from the far side.The basis for their argument is the blue-shiftedvelocity gradients in both the north and southof the cluster (see their Figure 1c) combinedwith the inference that the ONC must be onthe near side of the ISF. Note that the collapsescale in Hacar et al. (2017) is smaller than thatdiscussed in Tobin et al. (2009).

Recent analyses with APOGEE and Gaia sur-veys have shown that the ONC is slowly ex-panding (Kounkel et al. 2018; Kuhn et al. 2019),raising the question that whether gas is fallinginto ONC. However, it is possible if the ISF col-lapsing direction is toward OMC-1 (the dens-est gas region in the ISF), as Getman et al.(2019) recently have suggested that the starsin the OMC-1 region are undergoing gravita-tional contraction. On the other hand, Getmanet al. (2019) showed that OMC-4/5 are not mov-ing toward OMC-1. This raises the questionwhether the ONC/OMC-1 region is the gravi-tational focus. Meanwhile, it is possible thatthe global collapse is strongly impacted by the

intense feedback in the area, resulting in a verycomplex environment.

Alternatively, Stutz & Gould (2016) and Stutz(2018) proposed the “slingshot mechanism” forthe formation of stars in ISF of Orion A. Theyemphasized the wave-like geometry of the ISFand postulated that the filament has an oscil-latory motion. In this scenario, cores movewith the filament and the newly-formed starsare ejected from the oscillating filament oncethey are massive enough to decouple from thedense gas. The model addressed the wave-likespatial and kinematic morphology of the ISFand provided another explanation of the highvelocity of newly-formed stars. Recently, Stutzet al. (2018) used Gaia data to suggest that theISF filament is a standing wave.

With our newly acquired CARMA-NROOrion gas data, but making the PV-diagramalong the ISF (instead of declination, see Fig-ure 10), we detect more complicated kinematicfeatures (Figure 11). The cloud shows a “wave-like” feature (more so in the north with positiveoffsets in Figure 11 than in the south). Suchkinematics present a challenge in terms of sim-ply explaining the ISF with a global collapse to-ward ONC. Moreover, the northern part showsan oscillatory feature at scales of ∼ 3′. All theseresults suggest rather complex cloud kinematicsand more information is needed to understandthe origin and evolution of the ISF.

We notice something interesting from thecore-to-envelope PV-diagrams in Figure 7. Inthe bottom panels, we bin the core-to-envelopevelocities along the ISF. Each empty red circlerepresents the average core-to-envelope veloc-ity within a 6′ bin. Along the ISF, the dis-tribution of the red circles is again oscillatory.Note that this is not the wave-like structure ofthe filament gas. As shown in the figures, thecores are red-shifted relative to the gas in theOMC-2/3 regions (positive offsets), while in thesouthern region (negative offsets) the cores are

18 Kong et al.

blue-shifted relative to the gas. The origin ofthe core motion is unclear, and this is the firsttime they are detected. It is not clear why oneshould expect an oscillatory motion in veloci-ties between the core and the gas, but it doesremind us of the slingshot mechanism becauseof the “wave”. The expanding bubble from theTrapezium cluster may push away low-densityenvelopes (Pabst et al. 2019), resulting in anapparent shift of the core-to-envelope velocity.However, this scenario would require the impactof the feedback at very high extinctions and faraway from the location of massive stars alongthe entire ISF (see discussions in Großschedlet al. 2018), as the cores are deeply embedded.

5. CONCLUSIONS

In this paper, we studied sub-mm continuumcore formation and kinematics in the Orion Amolecular cloud. Various sources of publiclyavailable and proprietary data were combinedwith our CARMA-NRO Orion Survey data forthe study. Below we list our major conclusions.

1. We have compared 850 µm continuumcore detections (Lane et al. 2016) withNIR extinction (Meingast et al. 2018).We find a detection threshold of AV ≈5-10 mag for the cores within the decli-nation range of (-9.5◦, -4.5◦). Togetherwith previous findings in Ophiuchus andPerseus, we suggest a universal core for-mation extinction threshold of AV ∼ 5-10 mag for clouds within 500 pc from theSun. This threshold also appears to beconsistent with the star formation extinc-tion threshold suggested by Goldsmithet al. (2008); Lada et al. (2010); Heider-man et al. (2010). The total 850 µm coremass is about 4% of the mass of the entireOrion A cloud.

2. A pixel-by-pixel comparison between thesub-mm continuum and the NIR extinc-tion shows a steady increase in the contin-

uum detection probability for AV between∼ 5 and 20 mag, and a potential turningpoint at AV ∼20 mag where the slope inthe relation decreases. This turning pointis also seen in continuum core detection.

3. We have compared the pixel-based con-tinuum detection in Orion A with the in-frared dark cloud G28.37+0.07. Withinthe cloud extinction range of 15 mag .AV . 60 mag, Orion A shows more dustcontinuum emission detection, suggestingOrion A has been able to form more densegas (Σmm & 0.044 g cm−2). Our studyshows that the feedback from high-massstars in Orion A, possibly expanding HIIregions, is a potential cause of the dif-ference between these two clouds. Thecontinuum detection probability function(DPF) can be useful in showing the differ-ences in core formation between clouds.

4. We have investigated the dense core kine-matics within the Orion A integral-shapedfilament (ISF). In particular, we comparethe velocities of the dense cores (tracedby NH3) with the velocity of their sur-rounding lower density envelope (tracedby 13CO and C18O). We find that the 1Ddispersions in the distribution of the core-to-envelope velocity are 0.20 and 0.30 kms−1, when using C18O and 13CO to tracethe envelope, respectively. These are com-parable to the sound speed of the densegas and smaller than the local line dis-persion in the 13CO(1-0) and C18O(1-0)spectra. This is consistent with previ-ous findings in Perseus, suggesting thatthe dense molecular cores are kinemati-cally tied to their parent cloud, indepen-dent of the cloud environment. However,compared to Perseus, the ISF core-to-corevelocity dispersion is super-virial, largelydue to the overall velocity gradient in theISF. This difference may impact the dy-

Core Kinematics 19

namics of the forthcoming cluster in thefuture, which emphasizes the importanceof initial cloud conditions for star clusterformation.

5. The cloud gas velocity and the core-to-envelope velocity both show wave-like fea-tures in PV-diagrams, although the twoare conceptually different. Their relationto the overall dynamical picture of the ISFis to be determined, but see Stutz (2018)for recent work addressing the wave-likesignature in the ISF. The kinematic dif-ference between the dense cores and thecloud presented in this work remains tobe theoretically addressed.

Software: The data analysis in this paperuses python packages Astropy (Astropy Collab-oration et al. 2013), SciPy (Jones et al. 2001–),and Numpy (van der Walt et al. 2011). The fig-ures are plotted using python pacakges APLpy(Robitaille & Bressert 2012) and Matplotlib(Hunter 2007).

We thank the anonymous referee for construc-tive reports that significantly improve the qual-ity of the paper. We acknowledge the fruitfuldiscussions with James Lane and Helen Kirk.SK and HGA were funded by NSF award AST-

1140063 while conducting this study. RSKthanks for support from the German ScienceFoundation (DFG) via the Collaborative Re-search Center SFB 881 “The Milky Way Sys-tem” (subproject B1 and B2). AS acknowledgesfunding through Fondecyt regular (project code1180350), “Concurso Proyectos Internacionalesde Investigacion” (project code PII20150171),and Chilean Centro de Excelencia en Astrofısicay Tecnologıas Afines (CATA) BASAL grantAFB-170002. The JCMT has historically beenoperated by the Joint Astronomy Centre on be-half of the Science and Technology FacilitiesCouncil of the United Kingdom, the NationalResearch Council of Canada and the Nether-lands Organisation for Scientific Research. Ad-ditional funds for the construction of SCUBA-2were provided by the Canada Foundation for In-novation. The identification number for the pro-gramme under which the SCUBA-2 data used inthis paper is MJLSG313. This research used thefacilities of the Canadian Astronomy Data Cen-tre operated by the National Research Councilof Canada with the support of the CanadianSpace Agency. This research was carried out inpart at the Jet Propulsion Laboratory which isoperated for NASA by the California Instituteof Technology.

Facilities: CARMA, No:45m, JCMT, GBT,VISTA, Herschel

REFERENCES

Astropy Collaboration, Robitaille, T. P., Tollerud,E. J., et al. 2013, A&A, 558, A33

Bally, J. 2008, Overview of the Orion Complex,ed. B. Reipurth, 459

Bally, J., Ginsburg, A., Arce, H., et al. 2017, ApJ,837, 60

Bally, J., Langer, W. D., Stark, A. A., & Wilson,R. W. 1987, ApJL, 312, L45

Bonnell, I. A., Bate, M. R., Clarke, C. J., &Pringle, J. E. 2001, MNRAS, 323, 785

Brown, A. G. A., Vallenari, A., Prusti, T., et al.2018, A&A, 616, A1

Chapin, E. L., Berry, D. S., Gibb, A. G., et al.2013, MNRAS, 430, 2545

Clark, P. C., & Glover, S. C. O. 2014, MNRAS,444, 2396

Davis, C. J., Froebrich, D., Stanke, T., et al. 2009,A&A, 496, 153

Enoch, M. L., Young, K. E., Glenn, J., et al. 2006,ApJ, 638, 293

Feddersen, J. R., Arce, H. G., Kong, S., et al.2018, ApJ, 862, 121

Fiege, J. D., & Pudritz, R. E. 2000, MNRAS, 311,85

20 Kong et al.

Foster, J. B., Cottaar, M., Covey, K. R., et al.2015, ApJ, 799, 136

Friesen, R. K., Pineda, J. E., co-PIs, et al. 2017,ApJ, 843, 63

Getman, K. V., Feigelson, E. D., Kuhn, M. A., &Garmire, G. P. 2019, MNRAS, 1395

Goldsmith, P. F., Heyer, M., Narayanan, G., et al.2008, ApJ, 680, 428

Großschedl, J. E., Alves, J., Meingast, S., et al.2018, A&A, 619, A106

Hacar, A., Alves, J., Forbrich, J., et al. 2016,A&A, 589, A80

Hacar, A., Alves, J., Tafalla, M., & Goicoechea,J. R. 2017, A&A, 602, L2

Heiderman, A., Evans, Neal J., I., Allen, L. E.,Huard, T., & Heyer, M. 2010, ApJ, 723, 1019

Hunter, J. D. 2007, Computing in Science andEngineering, 9, 90

Johnstone, D., Di Francesco, J., & Kirk, H. 2004,ApJ, 611, L45

Jones, E., Oliphant, T., Peterson, P., et al. 2001–,SciPy: Open source scientific tools for Python, ,, [Online; accessed ¡today¿]

Kennicutt, R. C., & Evans, N. J. 2012, ARA&A,50, 531

Kirk, H., Johnstone, D., & Di Francesco, J. 2006,ApJ, 646, 1009

Kirk, H., Johnstone, D., & Tafalla, M. 2007, ApJ,668, 1042

Kirk, H., Pineda, J. E., Johnstone, D., &Goodman, A. 2010, ApJ, 723, 457

Kirk, H., Friesen, R. K., Pineda, J. E., et al. 2017,ApJ, 846, 144

Kong, S., Arce, H. G., Maureira, M. J., et al.2019, ApJ, 874, 104

Kong, S., Tan, J. C., Arce, H. G., et al. 2018a,ApJ, 855, L25

Kong, S., Tan, J. C., Caselli, P., et al. 2018b, ApJ,867, 94

Kong, S., Arce, H. G., Feddersen, J. R., et al.2018c, The Astrophysical Journal SupplementSeries, 236, 25

Kounkel, M., Covey, K., Suarez, G., et al. 2018,AJ, 156, 84

Krumholz, M. R., & McKee, C. F. 2005, ApJ, 630,250

Kuhn, M. A., Hillenbrand, L. A., Sills, A.,Feigelson, E. D., & Getman, K. V. 2019, ApJ,870, 32

Lada, C. J., Lombardi, M., & Alves, J. F. 2010,ApJ, 724, 687

Lane, J., Kirk, H., Johnstone, D., et al. 2016,ApJ, 833, 44

Lombardi, M., Bouy, H., Alves, J., & Lada, C. J.2014, A&A, 566, A45

Mairs, S., Johnstone, D., Kirk, H., et al. 2015,MNRAS, 454, 2557

—. 2016, MNRAS, 461, 4022McKee, C. F. 1989, ApJ, 345, 782Megeath, S. T., Gutermuth, R., Muzerolle, J.,

et al. 2012, AJ, 144, 192Meingast, S., Alves, J., & Lombardi, M. 2018,

A&A, 614, A65Orkisz, J. H., Pety, J., Gerin, M., et al. 2017,

A&A, 599, A99Pabst, C., Higgins, R., Goicoechea, J. R., et al.

2019, Nature, 565, 618Padoan, P., Juvela, M., Goodman, A. A., &

Nordlund, A. 2001, ApJ, 553, 227Pattle, K., Ward-Thompson, D., Berry, D., et al.

2017, ApJ, 846, 122Rieke, G. H., & Lebofsky, M. J. 1985, ApJ, 288,

618Robitaille, T., & Bressert, E. 2012, APLpy:

Astronomical Plotting Library in Python,Astrophysics Source Code Library, , ,ascl:1208.017

Salji, C. J., Richer, J. S., Buckle, J. V., et al.2015, MNRAS, 449, 1769

Smith, R. J., Longmore, S., & Bonnell, I. 2009,MNRAS, 400, 1775

Stutz, A. M. 2018, MNRAS, 473, 4890Stutz, A. M., Gonzalez-Lobos, V. I., & Gould, A.

2018, ArXiv e-prints, arXiv:1807.11496Stutz, A. M., & Gould, A. 2016, A&A, 590, A2Stutz, A. M., Tobin, J. J., Stanke, T., et al. 2013,

ApJ, 767, 36Tahani, M., Plume, R., Brown, J. C., &

Kainulainen, J. 2018, A&A, 614, A100Tobin, J. J., Hartmann, L., Furesz, G., Mateo, M.,

& Megeath, S. T. 2009, ApJ, 697, 1103van der Walt, S., Colbert, S. C., & Varoquaux, G.

2011, Computing in Science & Engineering, 13,22

Walsh, A. J., Myers, P. C., & Burton, M. G. 2004,ApJ, 614, 194

Ward-Thompson, D., Di Francesco, J., Hatchell,J., et al. 2007, Publications of the AstronomicalSociety of the Pacific, 119, 855

Core Kinematics 21

Young, K. E., Enoch, M. L., Evans, Neal J., I.,

et al. 2006, ApJ, 644, 326

Zhang, Q., Wang, K., Lu, X., & Jimenez-Serra, I.2015, ApJ, 804, 141