Star formation in 'the Brick': ALMA reveals an active proto-cluster in the Galactic centre cloud G0.253+0.016
Daniel L. Walker, Steven N. Longmore, John Bally, Adam Ginsburg, J. M. Diederik Kruijssen, Qizhou Zhang, Jonathan D. Henshaw, Xing Lu, João Alves, Ashley T. Barnes, Cara Battersby, Henrik Beuther, Yanett A. Contreras, Laura Gómez, Luis C. Ho, James M. Jackson, Jens Kauffmann, Elisabeth A. C. Mills, Thushara Pillai
MMNRAS , 1–23 (2021) Preprint 9 February 2021 Compiled using MNRAS L A TEX style file v3.0
Star formation in ‘the Brick’: ALMA reveals an activeproto-cluster in the Galactic centre cloud G0.253 + Daniel L. Walker , , (cid:63) , Steven N. Longmore , John Bally ,Adam Ginsburg , J. M. Diederik Kruijssen , Qizhou Zhang ,Jonathan D. Henshaw , Xing Lu , Jo˜ao Alves , , Ashley T. Barnes ,Cara Battersby , Henrik Beuther , Yanett A. Contreras , Laura G´omez ,Luis C. Ho , , James M. Jackson , Jens Kauffmann , Elisabeth A. C. Mills ,and Thushara Pillai ∗ Author affiliations are listed at the end of the paper
ABSTRACT
G0.253+0.016, aka ‘the Brick’, is one of the most massive ( > M (cid:12) ) and dense ( > cm − ) molecular clouds in the Milky Way’s Central Molecular Zone. Previousobservations have detected tentative signs of active star formation, most notably awater maser that is associated with a dust continuum source. We present ALMABand 6 observations with an angular resolution of 0.13 (cid:48)(cid:48) (1000 AU) towards this ‘masercore’, and report unambiguous evidence of active star formation within G0.253+0.016.We detect a population of eighteen continuum sources (median mass ∼ (cid:12) ), nineof which are driving bi-polar molecular outflows as seen via SiO (5-4) emission. Atthe location of the water maser, we find evidence for a protostellar binary/multiplewith multi-directional outflow emission. Despite the high density of G0.253+0.016, wefind no evidence for high-mass protostars in our ALMA field. The observed sourcesare instead consistent with a cluster of low-to-intermediate-mass protostars. However,the measured outflow properties are consistent with those expected for intermediate-to-high-mass star formation. We conclude that the sources are young and rapidlyaccreting, and may potentially form intermediate and high-mass stars in the future.The masses and projected spatial distribution of the cores are generally consistentwith thermal fragmentation, suggesting that the large-scale turbulence and strongmagnetic field in the cloud do not dominate on these scales, and that star formationon the scale of individual protostars is similar to that in Galactic disc environments. Key words:
Stars: formation – ISM: clouds – Galaxy: centre
The Milky Way’s Central Molecular Zone (CMZ, innerfew hundred parsecs) contains a substantial reservoir ( > M (cid:12) ) of dense ( > cm − ) molecular gas (Morris &Serabyn 1996). Despite this, the star formation rate (SFR)in the CMZ is at least an order of magnitude lower than pre-dicted by star formation relations that have been calibratedin nearby galactic disc environments (Longmore et al. 2013).This relative dearth of star formation is observed both onglobal scales and on the scales of individual molecular clouds (cid:63) E-mail: [email protected] in the CMZ (e.g. Barnes et al. 2017; Kauffmann et al. 2017b;Lu et al. 2019b). This deviation from the expected star for-mation rate is important, as it suggests that the criteriarequired for stars to form varies as a function of environ-ment. If this is true, then it is crucial that this variation isunderstood and characterised, such that star formation rela-tions can be accurately applied to the varying environmentalconditions found throughout the Universe.While the CMZ appears to be under-producing starsas a whole relative to the amount of dense gas it contains,one molecular cloud in particular has been the focus of sig-nificant research efforts in this context. G0.253+0.016 (alsoknown as ‘the Brick’) stands out as an extreme infra-red © a r X i v : . [ a s t r o - ph . GA ] F e b D. L. Walker et al. dark cloud against the intense mid-IR background (see Fig-ure 1). The cloud contains > M (cid:12) of material within amean radius of only a few parsecs (2-3 pc, e.g. Immer et al.2012; Longmore et al. 2012; Walker et al. 2015). Yet despitethis substantial reservoir of dense material, no evidence ofembedded star formation has been observed in the cloudother than a water maser that coincides with a compact mil-limetre continuum source (e.g. Lis et al. 1994; Immer et al.2012; Kauffmann et al. 2013; Johnston et al. 2014; Rath-borne et al. 2014b; Mills et al. 2015; Lu et al. 2019b). Here-after, we refer to this source as the ‘maser core’ for brevity.We note that there are at least two more water masers inG0.253+0.016 (see Figure 1), however, no counterparts havebeen detected in the dust continuum (Lu et al. 2019b).Deep radio continuum observations and further searchesfor maser emission do not reveal any additional signaturesof embedded star formation towards this source (e.g. Immeret al. 2012; Rodr´ıguez & Zapata 2013; Mills et al. 2015;Lu et al. 2019a). Potentially embedded star formation hasbeen inferred in G0.253+0.016 due to the presence of warmdust along one edge of the cloud (Marsh et al. 2016). Liset al. (2001) also suggest that the internal luminosity ofthe cloud could correspond to the presence of ∼ four B0zero-age main-sequence stars. Another potential indicationof star formation is the detection of an arc-like structurein the cloud that is close to the maser core in projectedposition (Higuchi et al. 2014; Mills et al. 2015; Henshawet al. 2019). Though the origin of this structure has beendisputed, new results suggest that it could be a feedback-driven shell of material, which may indicate embedded starformation (Henshaw et al. in prep.).These properties make G0.253+0.016 one of the mostmassive and dense molecular clouds known to exist inthe Galaxy in which there are no unambiguous signs ofwidespread star formation. The lack of on-going star forma-tion in G0.253+0.016, coupled with similar evidence in otherCMZ clouds, has been argued to favour an environmentally-dependent critical density threshold for star formation (e.g.Rathborne et al. 2014b; Kruijssen et al. 2014; Walker et al.2018; Ginsburg et al. 2018; Barnes et al. 2019). It has beenproposed that the CMZ undergoes an episodic cycle of starformation, and is currently at a low point due to the highturbulent energy there (Kruijssen et al. 2014; Krumholz& Kruijssen 2015; Krumholz et al. 2017; Armillotta et al.2019).The high turbulent energy is evidenced observationallyas broad line-widths of ∼
10 – 20 km s − on large (parsec)scales (Henshaw et al. 2016). This high turbulence will actto drive up the critical volume density threshold for star for-mation (e.g. Krumholz & McKee 2005; Padoan & Nordlund2011; Federrath & Klessen 2012; Hennebelle & Chabrier2013; Padoan et al. 2014), and may therefore explain the dis-crepancy between the observed current SFR and predictionsbased upon proposed density thresholds (e.g. Lada et al.2010, 2012). Recent results from a high-resolution surveyof the CMZ using the Submillimeter Array, CMZoom , showthat there is an overall lack of compact substructure withinthe the dense CMZ clouds, which is likely due to their in-ability to form such structure in this turbulent environment(Battersby et al. 2020; Hatchfield et al. 2020).Federrath et al. (2016) explored this in G0.253+0.016specifically, and concluded that the turbulence in the cloud is likely dominated by solenoidal turbulence, which is drivenby the strong shear in the CMZ’s deep gravitational poten-tial (Kruijssen et al. 2019) and could suppress the SFR bya factor of several (Dale et al. 2019). The strong ( ∼ mG)magnetic field in G0.253+0.016 has also been discussed asa potential source of support, which could suppress frag-mentation and thus star formation in the cloud (Pillai et al.2015). Given their relative proximity ( ∼ > (cid:12) ) andmassive stellar clusters ( > M (cid:12) ). The fact that the cloudcontains > M (cid:12) within a mean radius of ∼ ∼ M (cid:12) cluster. If the cloud were to ultimately form such a massivecluster, then a statistical argument would also suggest thelikely presence of precursors to high-mass stars due to sig-nificant sampling of the stellar initial mass function (IMF).Indeed, it is known that the CMZ harbours several youngmassive stellar clusters, such as the Arches and Quintuplet,that contain many high-mass stars, and even some extremelymassive stars ( >
100 M (cid:12) , e.g. Figer et al. 1999, 2002). Giventhat G0.253+0.016 is one of the best candidates for repre-senting a quiescent precursor to such clusters, it thereforefollows that it is a good candidate for hosting the initialconditions for massive star formation.While the ‘maser core’ in G0.253+0.016 constitutes thebest evidence for potentially active star formation withinthe cloud, the source has not been found to coincide withany 70 µ m point sources, radio continuum emission, nor anysignificant molecular line emission that would indicate thepresence of hot cores (e.g. Kauffmann et al. 2013; Rath-borne et al. 2014b). In this paper, we present high angularresolution Atacama Large Millimeter/submillimeter Array(ALMA) observations of this ‘maser core’ in G0.253+0.016.These observations reveal the presence of fragmentation,bipolar outflows and internal heating – unambiguous confir-mation of active star formation in G0.253+0.016. Section 2gives an overview of the observations and imaging techniquesused. Section 3 presents the results of the observations: (i)the 1.3 mm dust continuum and the physical properties ofthe detected sources, and (ii) the molecular line emission,specifically from SiO (5-4), CO (2-1), and CH CN J=12-11. Section 4 provides a discussion of the results and theimplications for our understanding of star formation bothin G0.253+0.016 and the CMZ in general.
We obtained high-sensitivity, high-angular-resolution
MNRAS , 1–23 (2021) tar formation in ‘the Brick’ Table 1.
Details of the three observed execution blocks. Listed are the observation dates, nominal array configurations, number of 12 mantennas in the array, full range of antenna baseline lengths, atmospheric precipitable water vapour content (PWV), total time on source,and the bandpass, flux, and phase calibrators used for each observation.Date Array Antennas Baselines PWV Time on source Bandpass Flux Phase(d/m/y) configuration G a l a c t i c L a t i t u d e Figure 1.
Three-colour image of G0.253+0.016.
Red : ALMA3 mm dust continuum (Rathborne et al. 2014b),
Green :Spitzer/GLIMPSE 8 µ m emission (Churchwell et al. 2009), Blue :Herschel/HiGAL dust column density (Battersby et al. 2011;Molinari et al. 2016). The white crosses indicate the positionsof known water masers (Lu et al. 2019b). The white circle corre-sponds to the primary beam field of view of the ALMA observa-tion reported in this paper.
Table 2.
Overview of the spectral setup used for our ALMA ob-servation. The specific line(s) targeted per spectral window aregiven, along with the corresponding central frequency ( ν cent ),bandwidth (BW), and spectral resolution in terms of velocity(∆ v ). While these are the lines that were specifically chosen, thereare many more lines observed within these spectral windows.Spectral ν cent BW ∆ v window (GHz) (GHz) (km s − )SiO (5-4) 217.105 0.234 0.78H CO (3 , – 2 , ) 218.222 0.234 0.78H CO (3 , – 2 , ) 218.476 0.234 0.78H CO (3 , – 2 , ) 218.760 0.234 0.78 CO (2-1)/CH CN (12-11) 220.709 0.934 0.77Continuum 232.500 1.875 2.50Continuum 235.000 1.875 2.47 dust continuum and molecular line observations towardsthe ‘maser core’ in G0.253+0.016 with ALMA at ∼
230 GHz (Band 6, 1.3 mm) as part of the Cycle 4 project2016.1.00949.S (PI: D. Walker). The observations were takenas a single pointing centred on the source (G0.261+0.016,see Figure 1), using only the main 12 m array. The correla-tor was configured to target 7 spectral windows, 5 of whichtargeted specific molecular transitions in the lower sidebandwith a spectral resolution of ∼ − . The remaining 2spectral windows were dedicated to broad-band continuumdetection in the upper sideband, with a spectral resolutionof ∼ − . The total aggregate bandwidth is approx-imately 5.6 GHz. The project was observed across 3 indi-vidual execution blocks between April and July 2017. Eachexecution used 40 antennas, with baselines ranging from 15– 3696 m. Full details concerning the observations and spec-tral setup are given in Tables 1 & 2, respectively. The ALMA pipeline calibrated data sets for each executionblock were combined to obtain final data products, whichwere then imaged in CASA (McMullin et al. 2007). Priorto final imaging, dirty cubes were created for each spectralwindow, and ran through the findContinuum routine inCASA in pipeline mode to determine the continuum-onlychannels in each window (Humphreys et al. 2016). The con-tinuum was then imaged in tclean by combining all spectralwindows and specifying the previously identified channelsto be considered when generating the continuum. The finalcontinuum image that is used throughout this paper wasimaged using the Briggs weighting scheme with a robustparameter of 0.5, multi-scale deconvolution, and with the auto-multithresh masking option (Kepley et al. 2020),using the default auto-masking parameters. The resultantimage has a synthesised beam size of 0.17 (cid:48)(cid:48) × (cid:48)(cid:48) ( ∼ × ∼ µ Jy beam − (0.07 K). The largest angular scale is ∼ (cid:48)(cid:48) (0.4 pc).Along with the continuum, we also imaged all spectralwindows to produce full data cubes. The cubes were imagedusing mostly the same parameters as for the continuum,with the exception of a higher cleaning threshold, and anauto-masking negativethreshold parameter of 7.0 (default https://almascience.nrao.edu/documents-and-tools/alma-science-pipeline-users-guide-casa-5-6.1 https://casaguides.nrao.edu/index.php/Automasking_Guide MNRAS000
230 GHz (Band 6, 1.3 mm) as part of the Cycle 4 project2016.1.00949.S (PI: D. Walker). The observations were takenas a single pointing centred on the source (G0.261+0.016,see Figure 1), using only the main 12 m array. The correla-tor was configured to target 7 spectral windows, 5 of whichtargeted specific molecular transitions in the lower sidebandwith a spectral resolution of ∼ − . The remaining 2spectral windows were dedicated to broad-band continuumdetection in the upper sideband, with a spectral resolutionof ∼ − . The total aggregate bandwidth is approx-imately 5.6 GHz. The project was observed across 3 indi-vidual execution blocks between April and July 2017. Eachexecution used 40 antennas, with baselines ranging from 15– 3696 m. Full details concerning the observations and spec-tral setup are given in Tables 1 & 2, respectively. The ALMA pipeline calibrated data sets for each executionblock were combined to obtain final data products, whichwere then imaged in CASA (McMullin et al. 2007). Priorto final imaging, dirty cubes were created for each spectralwindow, and ran through the findContinuum routine inCASA in pipeline mode to determine the continuum-onlychannels in each window (Humphreys et al. 2016). The con-tinuum was then imaged in tclean by combining all spectralwindows and specifying the previously identified channelsto be considered when generating the continuum. The finalcontinuum image that is used throughout this paper wasimaged using the Briggs weighting scheme with a robustparameter of 0.5, multi-scale deconvolution, and with the auto-multithresh masking option (Kepley et al. 2020),using the default auto-masking parameters. The resultantimage has a synthesised beam size of 0.17 (cid:48)(cid:48) × (cid:48)(cid:48) ( ∼ × ∼ µ Jy beam − (0.07 K). The largest angular scale is ∼ (cid:48)(cid:48) (0.4 pc).Along with the continuum, we also imaged all spectralwindows to produce full data cubes. The cubes were imagedusing mostly the same parameters as for the continuum,with the exception of a higher cleaning threshold, and anauto-masking negativethreshold parameter of 7.0 (default https://almascience.nrao.edu/documents-and-tools/alma-science-pipeline-users-guide-casa-5-6.1 https://casaguides.nrao.edu/index.php/Automasking_Guide MNRAS000 , 1–23 (2021)
D. L. Walker et al. is 0.0) to account for any absorption. In contrast to the con-tinuum, we opted to perform the cleaning prior to contin-uum subtraction. This was done as we found that using the uvcontsub task prior to cleaning did not perform a satisfac-tory continuum subtraction for the more line-rich spectralwindows. Instead, we used the statcont Python package,which is specifically designed to determine the continuumlevel in line-rich data and perform continuum subtraction(Sanchez-Monge et al. 2018). The resulting line sensitivityin a 0.78 km s − channel is ∼ µ Jy beam − (1.25 K). Figure 2a displays the 230 GHz continuum image of thefull ALMA field. This observation reveals that, while the tar-get field is still dominated by a bright central source on ∼ astrodendro Python package. In brief,dendrograms are hierarchical clustering algorithms, in whichstructure in a data set is represented as a ‘tree’, where sub-structures are classified as ‘branches’, and local maxima atthe highest level of the branch structures are called ‘leaves’.Using this nomenclature in the context of our continuumdata, each ‘leaf’ represents a continuum source or core.To compute the dendrogram, a threshold of 3 σ , an in-crement between structures of 1 σ , and a minimum numberof pixels in a source of 100 are specified (which is ∼ σ ∼ µ Jy beam − . Thenumber of sources and their properties are not strongly de-pendent on the choice of parameters, with the exception ofthe central source, which is more extended. We discuss thenature of the central source and its embedded structure laterin this section. A total of 17 compact continuum sources aredetected using dendrograms, which are highlighted in thezoom-ins in Figure 2(b - d). The general properties of thesesources are presented in Table 3, including their integratedfluxes, sizes and estimated masses.Assuming that the 1.3 mm continuum flux arises fromoptically thin dust emission (which is likely justified, see Luet al. 2019a), the masses of the detected sources are esti-mated using the following equation: M = d κ ν B ν ( T ) (cid:90) I ν d Ω = d F ν κ ν B ν ( T ) (1)where M is the mass, B ν is the Planck function, T is the dusttemperature, κ ν is the dust opacity, F ν is the integrated fluxand d is the distance. The dust opacity ( κ ν ) is not obser-vationally constrained here, and so we estimate this using κ ν = κ ( ν/ν ) β , where κ is taken to be 0.9 cm g − at ν = 230 GHz (Ossenkopf & Henning 1994), and β is as-sumed to be 1.75 (Battersby et al. 2011). The distance isassumed to be 8.1 kpc (Gravity Collaboration et al. 2019;Reid et al. 2019). We note that Zoccali et al. (2021) recentlyreported a distance of 7.2 kpc based on near-infrared star https://hera.ph1.uni-koeln.de/~sanchez/statcont counts towards the cloud. If true, this would have the ef-fect of decreasing our mass estimates by ∼ ∼ (cid:48)(cid:48) scales fromHerschel (Battersby et al. in prep). The average dust tem-perature towards this source is 22 K, and this is the valueused in the estimation of the dust masses. We acknowledgethat the masses reported here contain these uncertainties,and we explore the possibility of constraining these massesfurther with gas temperature estimates in section 3.5. How-ever, we find no evidence for significant line emission towardsthe majority of sources in our field, which suggests that theyare likely not significantly heated internally. This does notmean that the assumed dust temperature of 22 K is correct,but rather it is the best, and only measurement that we havefor the majority of the sources.Taking the aforementioned assumptions, we find thatthe sources range in mass from ∼ (cid:12) , with a medianof 2 M (cid:12) . In addition to our assumptions, these masses arepotentially lower limits due to the fact that the large scaleemission, some of which may be associated with the cores,is filtered out by the interferometer.We also note that new results from the AzTEC surveyof the CMZ measure higher values of β of ∼ (cid:48)(cid:48) (Tang et al. 2020a,b). Sub-stituting the upper value of this range over our assumedvalue of 1.75 would increase our reported masses by a fac-tor of ∼ (cid:48)(cid:48) ) using the PPMAP pro-cedure. The average dust temperature using this techniquereduces to ∼
17 K. Assuming this value combined with β =2.5 would increase our reported dust continuum masses anddensities by a factor of 1.42.Although dendrograms pick out the central source asa large (R ∼ ∼ − , suggesting that they may be ofsimilar mass, assuming equal temperatures. A 2D Gaussianfit to the central objects yields deconvolved mean radii of ∼ (cid:12) and18.0 M (cid:12) at 22 K). This would suggest that this is a massiveprotostellar binary, but there is a large uncertainty in the MNRAS , 1–23 (2021) tar formation in ‘the Brick’ Figure 2.
The 1.3 mm dust continuum image towards the ‘maser core’ in G0.253+0.016 as seen with ALMA at an angular resolutionof 0.13 (cid:48)(cid:48) ( ∼ displays the full field. The red contours show the 3mm dust continuum from (Rathborne et al. 2014b), andthe dotted boxes highlight the zoom-in regions shown in sub-figures (b) and (d). and show zoomed-in images of the compactsources detected via dendrograms. shows a zoom-in of the bright central source, denoted as core ‘1’, overlaid with contours of [3, 5,7, 9] mJy beam − or [4.4, 7.4, 10.3, 13.3] K. mass estimates due to the lack of dust temperature measure-ments. In the following section, we demonstrate that thesetwo sources are internally heated, and are therefore likelyless massive than the aforementioned estimates.To more clearly resolve this region, we also imaged thecontinuum using the Briggs weighting scheme with a robustparameter of -2.0 (i.e. uniform weighting), which prioritisesresolution over sensitivity. The resulting image is shown inFigure 3. Note that this image is not used for any analyses– all results reported use the image generated with a robustparameter of 0.5. Using this weighting scheme we see thatthe central sources are more clearly resolved into two dis-tinct components. This also reveals a potential third sourceto the upper left of source n > The full details of the spectral setup are given in Table2. Lines that were specifically targeted are SiO (5-4) and CO (2-1) as these are traditionally good outflow trac- ers (e.g. Bally 2016, and references therein), 3 para-H COtransitions, which can be used to measure gas temperaturesin the range ∼
50 – 150 K, and the J=12-11 k-ladder ofCH CN, which can be used to measure higher gas tempera-tures and is often found in the vicinity of protostars.Manual inspection of all spectral windows towards thecontinuum sources reveals that significant compact lineemission is only detected towards the central sources, 1a and1b. We do not find any single emission line in our spectralsetup that can reliably trace all continuum sources. Sucha lack of correspondence between continuum and molecularline emission has been noted previously in G0.253+0.016,and in the CMZ in general (e.g. Rathborne et al. 2015; Kauff-mann et al. 2017a; Henshaw et al. 2019), though Barneset al. (2019) recently reported a suite of molecular lines at ∼
260 GHz which do reliably trace the continuum structureon 1 (cid:48)(cid:48) scales in the CMZ dust ridge clouds D, E, and F. Wedefer detailed analysis of the molecular line emission to afuture publication, and focus only on the SiO, CO, andCH CN emission in the following sections.
As discussed in section 1, this region in G0.253+0.016 hasbeen noted in the literature due to the presence of a bright,
MNRAS000
MNRAS000 , 1–23 (2021)
D. L. Walker et al.
Galactic Longitude +00.0160°+00.0161° G a l a c t i c L a t i t u d e Figure 3.
Comparison of the central region of our ALMA field showing the 1.3 mm dust continuum generated using the cleaningparameter robust = 0.5 (a) and robust = -2.0 (b) . Contours are the same as those in Figure 2c. compact continuum source that is associated with watermaser emission. While this is potentially indicative of ac-tive star formation, no definitive signatures have previouslybeen found. To directly address the star forming nature ofthe source, we searched for outflows, as they are unambigu-ous signatures of active star formation (e.g. Bally 2016). Weexplicitly targeted the SiO (5-4) 217.105 GHz transition, asthis is a well-established outflow tracer. Previous observa-tions of G0.253+0.016 on larger scales with the SMA andALMA did not detect any signatures of outflow emission inthe cloud in SiO (5-4), / CO (2-1) or any other molecu-lar transitions (e.g. Kauffmann et al. 2013; Johnston et al.2014; Rathborne et al. 2014a). More generally, protostellaroutflows have largely eluded detection in the CMZ. To-date,they have only been detected in the massive star-formingregion Sagittarius B2 (Qin et al. 2008; Higuchi et al. 2015)and a few high-mass CMZ clouds (Lu et al. 2021).Figure 4 shows a two-colour map, where the blue andred correspond to the integrated intensity of the SiO (5-4)emission for the blue- and red-shifted emission across ourALMA field. The blue-shifted emission has been integratedover 29 – 42 km s − , and the red-shifted emission over 43 –56 km s − . There is more compact and diffuse SiO emissiondetected at both lower and higher velocities (see Figures 9 –13). The range displayed here has been chosen to highlightthe outflows while minimising confusion from more diffuseemission.We clearly detect multiple bipolar outflows associatedwith many of the continuum sources, along with larger-scaleemission in the field. Thus, we directly confirm that ac-tive star formation is occurring in G0.253+0.016. Overallwe identify outflow signatures associated with sources ∼ ∼ − (seeFigures 9 – 11). We find that the SiO emission integratedover this velocity range spatially coincides with the edgeof an arc-like structure that has previously been observed MNRAS , 1–23 (2021) tar formation in ‘the Brick’ Table 3.
The below table displays the properties of the identified continuum sources. Shown are the cores identified, along with theenclosed flux, central coordinates (as Galactic coordinates, l and b ), radius (calculated by taking the exact area of each dendrogram leafand determining the radius of a circle of equal area), core mass assuming a dust temperature of 22 K, number density ( n , assumingspherical symmetry), and whether or not corresponding outflow emission was detected. All mass estimates assume a gas-to-dust ratio of100 and a distance of 8.1 kpc (Gravity Collaboration et al. 2019; Reid et al. 2019). † The masses reported for CN emission (lower mass limit, see section 3.6.1). For all other sources, we do not detect any molecular lines that can beused to estimate gas temperatures.Source l b
Radius Mass n Outflow(mJy) ( ◦ ) ( ◦ ) (AU) (M (cid:12) ) (10 cm − ) detected?1 47.87 0.261034 0.0160561 4847 64.2 4.8 Yes1a 12.10 0.261008 0.016051 1300 1.7 - 16.2 † † Yes1b 14.34 0.261042 0.016058 1300 2.7 - 18.0 † † Yes2 2.15 0.260737 0.0153934 1475 2.9 7.7 Yes3 0.69 0.261774 0.0156472 1275 0.9 3.7 No4 0.95 0.262961 0.0158849 1543 1.3 3.0 Yes5 2.63 0.260413 0.0152248 1711 3.5 5.9 Yes6 1.48 0.263037 0.0161058 1682 2.0 3.6 No7 0.62 0.260976 0.0162781 1179 0.8 4.1 Maybe8 0.41 0.262593 0.0163765 945 0.6 6.0 Yes9 3.27 0.260611 0.016462 1687 4.4 7.8 Yes10 1.61 0.262265 0.016906 1236 2.2 9.9 No11 1.63 0.262183 0.0169606 1161 2.2 11.9 Yes12 1.06 0.261871 0.0199851 945 1.4 14.1 No13 1.68 0.261911 0.0200653 1139 2.3 13.2 No14 7.17 0.26156 0.0208225 2613 9.6 4.6 Yes15 2.34 0.262223 0.0201618 2206 3.1 2.4 No16 1.34 0.262202 0.0203759 1568 1.8 4.0 Yes17 0.76 0.257135 0.0163917 1302 1.0 3.8 No
Table 4.
Properties of the outflows detected via SiO (5-4) emission determined for each of the blue- and red-shifted outflow lobes. Shownfor each lobe is the projected size (l proj ), velocity range ( v range ), peak intensity ( I peak ), mean integrated SiO intensity ( (cid:104) (cid:82) T mb dv (cid:105) ),mean SiO column density ( (cid:104) N SiO (cid:105) ), mass (M), momentum (P), kinetic energy (E k ), dynamical time ( τ dyn , calculated by taking the fullextent of the velocity range), and the mass outflow rate ( ˙M out ). † A fractional SiO abundance of 1 × − is assumed. This is subjectto large uncertainties, which are discussed in section 3.4. ∗ The properties for source 11 are given both for the bright, compact outflowemission, as well as for the full flow including the faint extended emission, the latter of which are given in parentheses.
Source Lobe l proj v range I peak (cid:104) (cid:82) T mb dv (cid:105) (cid:104) N SiO (cid:105) M † P E k τ dyn ˙M out AU) (km s − ) (K) (K km s − ) (10 cm − ) (M (cid:12) ) (M (cid:12) km s − ) (M (cid:12) km s − ) (10 yr) (10 − M (cid:12) yr − )1 Blue 1.7 [10, 34] 5 38 5.1 0.12 2.2 28 3.4 3.7Red 2.3 [34, 71] 6 35 4.6 0.48 7.7 96 2.9 16.22 Blue 1.0 [36, 45] 6 11 1.5 0.02 0.1 0.3 5.2 0.4Red 0.5 [40, 49] 5 13 1.7 0.02 0.1 0.4 2.6 0.84 Blue 0.2 [8, 29] 4 39 5.2 0.01 0.3 4 0.5 2.6Red 0.2 [36, 48] 4 44 5.9 0.02 0.3 4 0.8 2.25 Blue 0.9 [29, 39] 4 12 1.5 0.04 0.2 0.5 4.3 0.5Red 2.8 [37, 49] 6 13 1.7 0.12 0.5 1.5 11.0 1.18 Blue 0.5 [15, 39] 5 48 6.5 0.05 1.0 14 1.0 4.8Red 0.7 [47, 84] 6 54 7.2 0.12 2.8 41 0.9 13.89 Blue 0.3 [25, 34] 3 26 3.4 0.02 0.4 5 1.6 1.1Red 0.7 [39, 58] 5 35 4.8 0.06 1.1 14 1.8 3.711 ∗ Blue 0.7 (6.8) [9, 36] 10 62 (18) 8.2 (2.3) 0.12 (0.78) 2.2 (8.9) 25 (73) 1.2 (11.9) 10.3 (6.6)Red 0.7 (6.8) [36, 59] 8 54 (22) 7.3 (2.9) 0.08 (1.08) 1.0 (12.3) 11 (100) 1.4 (14.0) 5.5 (7.7) in G0.253+0.016 in a number of high-density/temperatureand shock tracers, as well as CH OH masers (Higuchi et al.2014; Mills et al. 2015; Henshaw et al. 2019). The origin ofthis arcuate structure is unknown, but recent results suggestthat it could be a feedback-driven shell due to embedded starformation (Henshaw et al. in prep.). We also detect significant SiO emission at velocities thatare considerably lower than that of the V lsr of the ‘masercore’ (which is ∼
40 km s − ). In the range (-)18 – (+)10 kms − there are linear features that look outflow-like, as wellas arcuate structures that could be tracing shock-fronts (seeFigure 10). Given the low velocities of these features, it’s MNRAS000
40 km s − ). In the range (-)18 – (+)10 kms − there are linear features that look outflow-like, as wellas arcuate structures that could be tracing shock-fronts (seeFigure 10). Given the low velocities of these features, it’s MNRAS000 , 1–23 (2021)
D. L. Walker et al. h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
29 - 4243 - 56 km/s
Figure 4.
Two-colour image highlighting the outflows as traced by SiO (5-4) emission in our ALMA field. The blue-shifted emission isintegrated between 29 – 42 km s − , and the red-shifted emission 43 – 56 km s − . Continuum sources are highlighted by white ellipses,the extent of which corresponds to the structures determined using dendrograms. Each continuum source is also numbered. The yellowcrosses show the position of water masers from Lu et al. (2019b). likely that they are associated with foreground materialalong the line-of-sight, rather than G0.253+0.016. Having detected a population of molecular outflows that areassociated with the continuum sources in G0.253+0.016, wenow estimate their general properties, assuming local ther-modynamic equilibrium and optically thin SiO emission. Wealso assume that the outflow emission is parallel to the planeof the sky, since we do not have any knowledge of possibleinclinations. Though we clearly detect outflows associatedwith sources proj ) of each outflow lobe is calcu-lated assuming a distance of 8.1 kpc, with typical sizes of10 − AU. Combining these sizes with the full extent of themeasured velocity ranges, we measure dynamical timescales( τ dyn ) of ∼ − years. To estimate the column density ofthe outflow emission we follow the formalism presented insection 3.4 of Li et al. (2019). An excitation temperature of30 K is assumed. The column density is not too stronglydependent on the excitation temperature. An increase from30 K to 200 K, which covers the range of measured tem-peratures for the bulk of the gas in G0.253+0.016 (Gins-burg et al. 2016; Immer et al. 2016; Krieger et al. 2017), MNRAS , 1–23 (2021) tar formation in ‘the Brick’ only increases the estimated column density by a factor of ∼
2. Under these assumptions, typical mean column den-sities of ∼ − cm − are measured for the outflows inG0.253+0.016.To determine outflow masses, we sum the emission overthe spatial extent of each lobe and along the velocity axisover the relevant channels, and estimate the outflow massas: M outflow = (cid:88) A,v X − N SiO A pix µm H (2)where N SiO is the SiO column density, A pix is the pixel area, µ is the mean molecular weight which is assumed to be 2.8(Kauffmann et al. 2008), m H is the mass of hydrogen, and X SiO is the fractional abundance of SiO. The SiO abun-dance is subject to high uncertainty, with several orders ofmagnitude of spread reported in the literature. Estimatestowards IRDCs report an abundance of 5 × − (Sanhuezaet al. 2012), and measurements of some CMZ clouds report6 × − (Tsuboi et al. 2015). Li et al. (2019) report av-erage abundances in Galactic massive star-forming regionsof 4 × − , but with a scatter of ∼ ∼ × − to 3 × − for SiO outflows inmassive star-forming regions. The best constraints towardsG0.253+0.016 report an SiO abundance of 10 − from mea-surements on 26 (cid:48)(cid:48) scales (Mart´ın-Pintado et al. 1997). Asthese scales are significantly larger than those probed in thiswork, it is not clear whether this measurement should holdhere. Most recently, Lu et al. (2021) estimate the SiO abun-dances in a sample of 43 outflows in a few CMZ clouds tobe between 10 − and 10 − , with a mean value of 2 × − on scales of ∼ (cid:48)(cid:48) . However, the uncertainty in these abun-dances is at least one order of magnitude, and G0.253+0.016was not included in their sample.As we do not have any direct constraints on the SiOabundance in G0.253+0.016 on the spatial scales discussedin this paper, we assume an abundance of 10 − as a softupper limit. This is consistent with upper limits measuredin star-forming regions both in the Galactic disc and in theCMZ. This assumption means that any masses and depen-dent properties reported are considered to be likely lowerlimits. However, given the already-high abundance of gas-phase SiO in the CMZ, which could be enhanced even fur-ther in the vicinity of protostellar outflows due to high-velocity shocks (e.g. Schilke et al. 1997), it is plausible thatthe abundance may even be as high as a few × − (Gusdorfet al., private communication). Such high SiO abundanceshave been assumed in the extreme star-forming region W51(Goddi et al. 2020).Under these assumptions, we obtain outflow masses oforder 10 − to 1 M (cid:12) , and mass outflow rates ( ˙M out ) of 10 − to 10 − M (cid:12) yr − . We also estimate the outflow momentumand kinetic energy as: P outflow = (cid:88) A,v M | v − v lsr | (3) E k, outflow = 12 (cid:88) A,v M | v − v lsr | (4)where v is the velocity at a given channel, and v lsr is takento be the velocity of the source driving the outflow. In the majority of cases, the source velocity is not known due to thelack of line emission associated with most of the continuumsources (see section 3.2). Thus, for most of the outflows,we assume the velocity of the central source, which is ∼
40 km s − . The estimated momenta range ∼ (cid:12) kms − . and energies range ∼ (cid:12) km s − (1 × –2 × erg) per lobe. The total estimated energy contained in the detected out-flows is ∼ × erg. To investigate what impact these out-flows may have on the local environment, we first estimatethe gravitational energy of the material in our full ALMAfield. Using the dust column density map from Hi-GAL (30 (cid:48)(cid:48) ,Molinari et al. 2010; Battersby et al. 2011; Mills & Battersby2017), we estimate this mass to be ∼ × M (cid:12) within aradius of 19 (cid:48)(cid:48) . This corresponds to a gravitational energyof 5 × erg via E grav = GM /R. We also estimate theturbulent energy in the region as E turb = M( √ σ los,1D ) /2,where σ los,1D is the one dimensional line-of-sight velocitydispersion. To measure this velocity dispersion, we take anaveraged spectrum across our ALMA field using the HNCOemission from the MALT90 survey (Jackson et al. 2013).The angular resolution of the MALT90 data is 38 (cid:48)(cid:48) , whichis approximately the same size as our ALMA field of view.The averaged HNCO spectrum shows two overlapping ve-locity components. Fitting the brightest component with asingle Gaussian yields σ los,1D ∼ − . If we includeboth components, this increases to ∼
12 km s − . Using thisrange we estimate the turbulent energy to be ∼ × –3 × erg.These results suggest that the detected population ofSiO outflows are not energetic enough to drive the local tur-bulence or to significantly disrupt the local material. Wereiterate that the measured outflow masses and energies arepotential lower limits and so the impact of the outflows couldbe larger. However, the fractional SiO abundance wouldneed to be 3 – 4 orders of magnitude lower in order for theoutflow energy to be similar to the gravitational and turbu-lent energy, which seems unlikely given the high abundanceof SiO in the gas phase the CMZ in general ( ∼ − , e.g.Mart´ın-Pintado et al. 1997; Amo-Baladr´on et al. 2009). CO (2-1) emission
CO is the most commonly used tracer for identifying out-flows, due to the high abundance of the molecule along withthe relatively low energies of the lower rotational states (e.g.Bally 2016, and references therein). We therefore searchedfor outflow emission via the CO (2-1) transition ( COwas not covered in our spectral setup). However, the COemission towards G0.253+0.016 is complex and widespreadon large spatial scales, thus making it difficult to isolate anyemission potentially associated with outflows. Despite this,we do detect a significant amount of interesting structurevia the CO emission, which is discussed below.Unlike with the SiO emission, we do not detect clear-cut outflow emission in CO. Though we do identify a largenumber of linear features, particularly towards the centralregion of the field, that may be associated with outflows
MNRAS , 1–23 (2021) D. L. Walker et al. and/or outflow cavities (see Figure 5d). As shown in Figure5b, we also identify large regions lacking in emission thatare roughly centred on source ∼ (cid:48)(cid:48) in extent.We also see an hourglass-shaped emission structure between ∼
56 – 60 km s − (see Figure 6b). This structure is centredon source ∼
50 km s − tail of emission that is associated with the water maser tothe South-East of the centre of the field, as shown in Fig-ure 5d. This CO tail also coincides with a bright, compactknot of SiO emission that is very close to the second watermaser (see Figure 4). This water maser was reported in Luet al. (2019b), where they note that the maser emission wasnot found to coincide with any continuum source on ∼ (cid:48)(cid:48) scales. Our observations confirm this on ∼ CO at this loca-tion, combined with the presence of a water maser, suggest aprotostellar nature. The fact that we do not detect any con-tinuum emission could imply that there truly is no sourceat this location, or if there is a source, then it is below ourdetection limit.At velocities ∼
52 km s − , there is very linear regionof CO emission that is centred on source > ∼
15 degrees (see Figure6a). This emission is constrained to just a few km s − , andshows a ‘braided’ structure, which appears to rotate whenstepping through in velocity. Such a structure may plau-sibly be caused by precession of the central binary sourceshaping the outflow emission (e.g. Fendt & Zinnecker 1998),though it is not clear that the origin of this emission is dueto outflowing material. 2 As shown in Figure 5c, at the ap-proximate V lsr of the central cluster of continuum sources( ∼
41 km s − ), the CO emission is complex. Most notablythere are a series of arc-like filamentary structures with aroughly South-East – North-West orientation. This ‘bear-claw’ like structure is reminiscent of the HCO + absorptionfilaments observed in G0.253+0.016 on larger scales (Ballyet al. 2014). Inspection of the HCO + cube reveals that thetwo prominent broad-line absorption filaments shown in Fig-ure 1 of Bally et al. (2014) directly cross the centre of ourALMA field and coincide with the arcuate CO structures,both in position and velocity. In Figure 5c, we also see someinteresting structure at the location of the sources ∼ -35 km s − ) we also see some emission.In particular we identify two bright, compact knots of COemission just North of the central continuum sources (seeFigure 5a). These knots are confined to only a few km s − ,but they display velocity gradients across their small extent.Given their location in the central cluster, these could bebullets or post-shock clumps due to protostellar outflows inthe region. Though their highly blue-shifted velocities maysuggest that they could be due to some unrelated emissionalong the line-of-sight. Though there is a diverse amount of structure observedvia the CO emission, we stress that caution must be takenwhen interpreting the data. The emission is complex andmuch of it is diffuse. As our observations do not utilise the7 m or Total Power arrays of ALMA, we are missing shortspacing data and so much of the larger-scale diffuse emis-sion will be filtered out by the long baselines of the inter-ferometer. Thus, while we can speculate on the origin of theemission, we do not present any strong conclusions based onthe CO data. CN (12-11) emission CH CN (methyl cyanide) is commonly used to trace small-scale gas kinematics towards hot protostellar cores, and therelative ratios of the k-components can be used to estimategas temperatures and column densities (e.g. some recent re-sults include: Beuther et al. 2017; Ilee et al. 2018; Maud et al.2018). With this in mind, we targeted the J=12-11 k-ladderof CH CN at ∼
221 GHz as another means of identifyingsigns of star formation via locally-heated gas.We detect CH CN only towards the central cores(sources CN emission is dominated by source CN emission towards CN and k=0-6 components of the isotopologue CH
CNare clearly detected towards the core. The upper energy lev-els of the CH CN k=0-8 components are 69, 76, 97, 133,183, 247, 326, 419, and 525 K. Their detection therefore in-dicates that the gas is hot, and is likely internally heated byan embedded protostar(s).
Measuring gas temperatures to be used as a proxy for thedust temperature is a fairly common practice in star for-mation studies. However, measurements in the CMZ haveshown that the gas and dust are thermally decoupled onlarge scales ( ∼ (cid:48)(cid:48) ), with gas temperatures that are of-ten many factors greater than measured dust temperatures(Ginsburg et al. 2016; Immer et al. 2016). As we begin toprobe sub-parsec scales in the CMZ, we no longer have dusttemperature measurements, as these scales are not accessi-ble with facilities such as Herschel or SOFIA at the distanceof the CMZ. This leaves a large uncertainty in the dust tem-peratures, and hence mass estimates. However, SPH mod-els of G0.253+0.016 suggest that the gas and dust shouldbe close to thermalised in the density regime of the sourcesthat we detect with our ALMA observations ( ∼ − cm − ,Clark et al. 2013). Thus, in the following we estimate gastemperatures (where possible) and use these as a proxy forthe dust temperatures, while acknowledging the caveat thatthey may still be weakly decoupled. Any uncertainties inreported mass estimates are likely dominated by this.We use the eXtended CASA Line Analysis Software MNRAS , 1–23 (2021) tar formation in ‘the Brick’ (a) (b) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J ) (c) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J ) (d) Figure 5.
Three-colour figures showing the CO emission at different velocities. (a) : -34.5 (blue), -33.8 (green), and -33.1 (red) kms − , (b) : 31.9 (blue), 33.3 (green), and 34.6 (red) km s − , (c) : 42.6 (blue), 43.9 (green), and 45.2 (red) km s − , (d) : 47.9 (blue), 49.2(green), and 50.5 (red) km s − . Continuum sources are highlighted by white ellipses, the extent of which corresponds to the structuresdetermined using dendrograms. Each continuum source is also numbered. The yellow crosses show the position of water masers from Luet al. (2019b). The red arrows in (a) highlight two bright, compact knots of CO emission. The grey dashed line in (b) shows the axisof symmetry of a cavity-like region.
Suite (XCLASS , M¨oller et al. 2017) software package withinCASA to simultaneously fit the CH CN and CH
CN emis-sion to obtain an estimate of the average gas temperaturetowards sources https://xclass.astro.uni-koeln.de/ is not well sampled by the synthesised beam. We assume afilling factor of unity.The resulting single-component fit for × cm − , and 3.6 km s − , respectively. This tem-perature is likely greater on smaller scales, close to theprotostar(s), where the relative intensities of the higher k-components are likely to be greater. Note that there are sev- MNRAS000
CN emis-sion to obtain an estimate of the average gas temperaturetowards sources https://xclass.astro.uni-koeln.de/ is not well sampled by the synthesised beam. We assume afilling factor of unity.The resulting single-component fit for × cm − , and 3.6 km s − , respectively. This tem-perature is likely greater on smaller scales, close to theprotostar(s), where the relative intensities of the higher k-components are likely to be greater. Note that there are sev- MNRAS000 , 1–23 (2021) D. L. Walker et al. h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J ) (a) (b) Figure 6.
As in Figure 5. Three-colour figures showing the CO emission at different velocities. (a) : 50.5 (blue), 51.9 (green), and 53.2(red), (b) : 56.5 (blue), 57.8 (green), and 59.2 (red) km s − . The grey dashed line in (b) shows the axis of symmetry of an hourglass-shapedregion that may be tracing an outflow cavity from source 8. Figure 7.
Moment maps of the CH CN J=12-11 k=3 emission towards the central sources (cores CN emission isstrongly dominated by core 1a. Contours are at the same levels as those described for Figure 1c. eral peaks that are not fit by the model. This is because thoseemission lines are not associated with CH CN/CH
CN .If we take this estimated gas temperature of 167 K andassume that the gas and dust are thermalised at these densi-ties, we can use this to better constrain the lower limit of thedust mass. In section 3.1 we estimated an upper mass limitfor source (cid:12) at 22 K. Assuming a dust tem-perature of 167 K, this mass estimate decreases to ∼ (cid:12) .We also estimate the average gas temperature of (cid:12) to3 M (cid:12) . If this system is a protostellar binary, then another methodof constraining their masses is through a dynamical argu-ment. Looking at the two right-most panels in Figure 7 – the1 st and 2 nd moments – there is a small difference betweenthe velocities and velocity dispersions of the two sources. Ifwe assume that the sources are of equal mass, and that theirmeasured velocities are the maximum line-of-sight velocities(i.e. that we are viewing the system edge-on), then we canuse a simple gravitational-kinetic energy balance argumentto estimate their masses. For a binary separation of 1150 AU,and a velocity difference between the sources of 1.5 km s − ,we estimate that the dynamical mass of each source is ∼ (cid:12) . If we now consider the possible inclination, whichis observationally unconstrained, this calculation is modu- MNRAS , 1–23 (2021) tar formation in ‘the Brick’ .
45 0 .
50 0 .
55 0 .
60 0 .
65 0 . Sky Frequency (GHz) +2 . × T ν ( K ) ∆ V = 3.6 km s − Figure 8.
Beam-averaged spectrum of the CH CN J=12-11 emis-sion towards the central source (core CN, and the underlined numbers correspondto the k=0-6 components of isotopologue CH
CN. The best-fitted temperature and line-width are shown at the bottom right.There are a few lines that are not fitted here, as they are not fromCH CN or its isotopologues. lated by a sin(i) term. For 15 ◦ (cid:54) i (cid:54) ◦ , the estimateddynamical mass range is 5.6 M (cid:12) (cid:62) M dyn (cid:62) (cid:12) .While this is a simple argument with several caveats,and one that requires that the sources are actually in a bi-nary system, it demonstrates that the dynamical masses insuch a scenario are consistent with the dust masses that havebeen estimated assuming that T dust = T gas . This suggeststhat this assumption may be valid, and that the dust andgas may be thermalised at these densities ( ∼ − cm − )in the CMZ. If instead we assume the upper mass limit forthese sources of ∼
18 M (cid:12) (for T dust = 22 K), this wouldrequire a velocity difference between them of 3 – 5.5 km s − for 15 ◦ (cid:54) i (cid:54) ◦ , which is several factors greater than whatis observed. Given the extreme conditions in the CMZ, particularly theelevated gas temperatures and high gas densities (e.g Gins-burg et al. 2016; Immer et al. 2016; Krieger et al. 2017;Mills et al. 2018), it is pertinent to investigate the contin-uum structure and the nature of the fragmentation that oc-curs in the molecular clouds here. The thermal Jeans massis given by: M J = π / c s (cid:112) G ρ (5)where c s = ( k b T /µm H ) / is the sound speed, G is the gravi-tational constant, and ρ is the volume density. For the ‘masercore’ on 1 (cid:48)(cid:48) scales, Rathborne et al. (2014b) estimate a rangeof densities from 1 – 3 × cm − , for a dust temperaturerange of 50 – 20 K, respectively. Taking this range of param-eters, the resulting thermal Jeans mass ranges from 0.35 –2.15 M (cid:12) . This range is broadly consistent with the massesthat we estimate for the 1.3 mm continuum sources in our data, for which the median mass is 2 M (cid:12) . We reiterate thatour mass estimates may be affected by significant uncer-tainties, primarily due to spatial filtering and unconstraineddust temperatures on these scales. Nonetheless, we find thatthe observed structure on small spatial scales is generallyconsistent with thermal Jeans fragmentation within the un-certainties.While all of the individual core masses and the medianmass are consistent with thermal Jeans fragmentation, thefull extent of the central source ∼
35 M (cid:12) (assuming T dust = 22 K), which is 16 – 100 times greaterthan the thermal Jeans mass. This suggests that thermalfragmentation alone is not sufficient to explain the propertiesof core c s ) in Equation 5 with the velocity dispersion of the gas( σ ), under the assumption that the total velocity dispersionis a suitable proxy for the turbulent linewidth (e.g. Wanget al. 2014; Li et al. 2019). As explained in Section 3.4.1,the velocity dispersion measured with single-dish data forthis region of the cloud is in the range of 4 – 12 km s − .Substituting this range in place of the sound speed yieldsturbulent fragmentation masses of thousands of M (cid:12) for therange of densities given. This is significantly larger than themasses of the fragments that we observe. Even if we assumethe small-scale velocity dispersion as measured towards thecentral sources via the CH CN emission (∆V = 3.6 km s − , σ = 1.5 km s − , see Section 3.6.1), the turbulent fragmen-tation mass is ∼
50 – 100 M (cid:12) . This further supports theconclusion that the thermal pressure is likely dominatingthe fragmentation process in this region of G0.253+0.016,not the turbulent pressure.We also consider the thermal Jeans length, given by: λ J = c s (cid:114) πGρ (6)Taking the same range of densities and temperaturesas for the Jeans mass estimation, we obtain a range of 0.8– 1.4 × AU for the thermal Jeans length. Using a near-est neighbour algorithm, we find that the mean projectedseparation of the continuum sources in our ALMA field is4.5 × AU. If we restrict this to only consider the sourcesin the central region of the field (sources 1 – 11), this re-duces slightly to 3.4 × AU. This is generally consistentwith the expected separations from thermal Jeans fragmen-tation, though the observed separations are larger by a fac-tor of a few.These results suggest that even in a cloud that is so tur-bulent (Henshaw et al. 2019, 2020) and in such an extremeenvironment as the CMZ, thermal Jeans fragmentation maystill dominate the fragmentation properties on protostellarscales. Results outside of the CMZ conclude that thermalJeans fragmentation is sufficient to explain the observed coreproperties in a variety of Galactic disc environments on sim-ilar scales (e.g. Alves et al. 2007; Lada et al. 2008; Beutheret al. 2018).While we cannot draw definitive conclusions from asmall number of cores in a region of a single cloud, this result
MNRAS000
MNRAS000 , 1–23 (2021) D. L. Walker et al. is important. It suggests that while large-scale gas propertiesin the CMZ are strongly shaped by turbulence (e.g. Gins-burg et al. 2016; Henshaw et al. 2016), the small-scale prop-erties may be less sensitive to this, in which case star forma-tion may proceed ‘normally’ once it is underway. Lu et al.(2019b) recently came to a similar conclusion when com-paring star formation efficiencies on large and small scalesin CMZ clouds. Using ALMA data similar in resolution andsensitivity to those presented here, Lu et al. (2020) also con-cluded that the structure in four other massive CMZ molec-ular clouds is consistent with thermal Jeans fragmentation.Ultimately, these results suggest that the process of star for-mation on protostellar scales in this Galactic environmentis not fundamentally different, other than the fact that theinitial fragmentation towards protostellar cores is inhibitedbelow a higher critical density threshold when compared tothe Galactic disc (e.g. Ginsburg et al. 2018; Walker et al.2018; Barnes et al. 2019). + yet G0.253+0.016 is one of the most massive ( > M (cid:12) ) anddense ( > cm − ) molecular clouds known to exist in theGalaxy that appears to be largely quiescent (e.g. Immeret al. 2012; Kauffmann et al. 2013; Johnston et al. 2014).Given these properties, it has been proposed to represent anideal candidate precursor to a massive stellar cluster (Long-more et al. 2012; Rathborne et al. 2014a; Walker et al. 2015,2016). Indeed, some of the most extreme star clusters existin this region of the Galaxy, such as the Arches and Quin-tuplet, with the former being the most dense young clusterin the Galaxy (Espinoza et al. 2009). As these clusters arerelatively young ( ∼ (cid:48)(cid:48) , Rathborne et al. (2015) estimated thatthis region contained 72 M (cid:12) within a radius of 0.04 pc, cor-responding to a volume density of 3 × cm − . If high-massstars are presently forming anywhere in G0.253+0.016, thenthis is the most likely location.Our ALMA observations show that, while star forma-tion is unambiguously underway in this region of the cloud,there are no obvious high-mass protostars residing within orin the vicinity of the ‘maser core’. The mass range of the de-tected cores is 0.6 – 9.6 M (cid:12) with a median of 2 M (cid:12) , thoughthis is subject to several caveats, particularly due to uncer-tain dust temperatures and missing flux (see section 3.1).We acknowledge that the upper limit for the masses of thecentral sources ( (cid:12) ) due to the uncertainty in the dust temperature, which would putthem in high-mass protostar territory. However, as discussedin section 3.6.1, their bright CH CN emission combined withtheir high densities means that it is likely that the dust andgas should be close to thermalised (Clark et al. 2013), andtheir masses would therefore be on the lower end of the es-timated range (2 – 3 M (cid:12) ). This is also consistent with therange of dynamical masses estimated in section 3.6.2 (1.5 –5.6 M (cid:12) ).In addition to the lack of high-mass cores, there are alsono signatures of ongoing high-mass star formation. Masersurveys have only found water maser emission towards thisregion (e.g. Lu et al. 2019b), but no class II methanol masers,which are typically indicative of high-mass star formation,have been detected. No HII regions have been detected to-wards the ‘maser core’ in the radio continuum either (Immeret al. 2012; Rodr´ıguez & Zapata 2013; Mills et al. 2015; Luet al. 2019a).Despite the low-intermediate masses of the continuumsources in this region, the properties of the detected SiOoutflows, namely their masses, energies, momenta, and massoutflow rates, are similar to those observed in intermediateand high-mass star-forming regions (e.g. Beuther et al. 2002;Arce et al. 2007; Zhang et al. 2005; Bally 2016; Beltr´an & deWit 2016, and references therein). These properties are alsoconsidered to be soft lower limits (see section 3.4), and maybe larger if the true SiO abundance is lower and if we aremissing flux due to spatial filtering. The typical dynamicalage is ∼ years, suggesting that the embedded protostarsare young. This, coupled with the high outflow rates of ∼ − – 10 − M (cid:12) yr − , means that these sources may havethe potential to become intermediate or high-mass stars inthe future.Of particular interest in this context are the two centralsources in this field. The cores are situated in a larger-scaleclump of dense material (see Figure 2). Assuming a dusttemperature of 22 K and subtracting the flux contributionfrom the embedded sources, we estimate that this envelopehas a mass of ∼
35 M (cid:12) within 5000 AU. If the embeddedprotostars were able to efficiently accrete from this materialand the larger scale reservoir, they could potentially grow tobecome high-mass stars. The lower limit to the total massoutflow rate of these sources is ∼ × − M (cid:12) yr − , whichis consistent with outflow rates observed in high-mass star-forming regions. Assuming that this is a lower limit to themass infall rate, and that the accretion rate onto the proto-star(s) is some fraction of the infall rate, then it would take10 − yrs to form a ∼
10 M (cid:12) star, and potentially less ifthe accretion rate is variable and increases as the protostarsgrow in mass (e.g Zhang et al. 2005, 2015; Li et al. 2020).Given the densities presented in Table 3, the typical free-fall time of the detected sources is a few thousand years.This means that in the aforementioned scenario, the centralsources would have to accrete over many free-fall times.With the data presented in this paper, we are not able toresolve the central sources well, and it is not possible to de-termine the contribution that either of the sources is makingto the outflowing material. Follow-up molecular line obser-vations at higher angular resolution are required to betterconstrain the infall and accretion rates onto the central pro-tostars, and ultimately determine whether this could repre-sent the early stages of a high-mass binary in G0.253+0.016.
MNRAS , 1–23 (2021) tar formation in ‘the Brick’ We have presented high-resolution (0.13 (cid:48)(cid:48) , 1000 AU) ALMABand 6 (1.3 mm, 230 GHz) observations towards the ‘masercore’ in the extreme Galactic centre cloud G0.253+0.016.The main results are summarised as follows: • The ‘maser core’ region fragments significantly on1000 AU scales, revealing a small cluster of at least 18compact sources that are detected in the dust contin-uum. The median mass of the cores is 2 M (cid:12) , with typicalradii of ∼ − cm − . • We detect at least 9 bi-polar outflows via SiO (5-4)emission that are associated with the observed dust con-tinuum sources. We also find potential evidence for out-flows and outflow cavities traced by CO (2-1) emis-sion. This constitutes unambiguous evidence of activestar formation in G0.253+0.016. • The central source of the ‘maser core’ dominates thecontinuum flux on small-scales, and is revealed to bea protostellar binary system (projected separation ∼ • Despite the high densities towards this region ( > cm − ) and the large reservoir of dense gas inG0.253+0.016 as a whole, we find no evidence of high-mass protostars. However, the observed SiO outflowproperties are consistent with those observed towardsintermediate and high-mass protostars, and so some ofthe detected cores may potentially grow to become high-mass stars in the future. The central protostellar binaryis a promising candidate for a future high-mass stellarbinary, as it is embedded in a dense reservoir of mate-rial. Direct measurements of infall/accretion rates arenecessary to determine whether they could potentiallybecome high-mass stars in the future. • The masses and distribution of the detected contin-uum sources are found to be generally consistent withthermal Jeans fragmentation. This suggests that thelarge-scale turbulence may not play a significant rolein shaping the cloud structure on protostellar scales,and that the mechanisms governing the fragmentationof protostellar-scale structure in this extreme Galacticenvironment are similar to those in the nearby star-forming regions at the individual core scale.
ACKNOWLEDGEMENTS
This paper makes use of the following ALMA data:ADS/JAO.ALMA
Software : This research primarily made use of thefollowing software packages:
CASA (McMullin et al.2007), astropy , a community-developed core Pythonpackage for Astronomy (Astropy Collaboration et al.2013; Price-Whelan et al. 2018), astrodendro , aPython package to compute dendrograms of astro-nomical data ( ), APLpy ,an open-source plotting package for Python (Robitaille &Bressert 2012), spectral-cube ( https://spectral-cube.readthedocs.io/en/latest/ ), radio-beam ( https://radio-beam.readthedocs.io/en/latest/ ), statcont ( https://hera.ph1.uni-koeln.de/~sanchez/statcont ),and XCLASS ( https://xclass.astro.uni-koeln.de/ ). Data availability : The data products used to conduct theresearch presented in this paper are made publicly availableat the following Harvard Dataverse repository: https://doi.org/10.7910/DVN/FXORIW .Uncompressed versions of the figures included in the paperare also made available in the same Dataverse at: https://doi.org/10.7910/DVN/O6GN1T . REFERENCES
Alves J., Lombardi M., Lada C. J., 2007, A&A, 462, L17Amo-Baladr´on M. A., Mart´ın-Pintado J., Morris M. R., MunoM. P., Rodr´ıguez-Fern´andez N. J., 2009, ApJ, 694, 943Arce H. G., Shepherd D., Gueth F., Lee C. F., Bachiller R., RosenA., Beuther H., 2007, in Reipurth B., Jewitt D., Keil K., eds,Protostars and Planets V. p. 245 ( arXiv:astro-ph/0603071 )Armillotta L., Krumholz M. R., Di Teodoro E. M., McClure-Griffiths N. M., 2019, MNRAS, 490, 4401Astropy Collaboration et al., 2013, A&A, 558, A33Bally J., 2016, ARA&A, 54, 491Bally J., et al., 2014, ApJ, 795, 28MNRAS000
Alves J., Lombardi M., Lada C. J., 2007, A&A, 462, L17Amo-Baladr´on M. A., Mart´ın-Pintado J., Morris M. R., MunoM. P., Rodr´ıguez-Fern´andez N. J., 2009, ApJ, 694, 943Arce H. G., Shepherd D., Gueth F., Lee C. F., Bachiller R., RosenA., Beuther H., 2007, in Reipurth B., Jewitt D., Keil K., eds,Protostars and Planets V. p. 245 ( arXiv:astro-ph/0603071 )Armillotta L., Krumholz M. R., Di Teodoro E. M., McClure-Griffiths N. M., 2019, MNRAS, 490, 4401Astropy Collaboration et al., 2013, A&A, 558, A33Bally J., 2016, ARA&A, 54, 491Bally J., et al., 2014, ApJ, 795, 28MNRAS000 , 1–23 (2021) D. L. Walker et al.
Barnes A. T., Longmore S. N., Battersby C., el al 2017, MNRAS,469, 2263Barnes A. T., Longmore S. N., Avison A., et al 2019, MNRAS,486, 283Battersby C., Bally J., Ginsburg A., et al 2011, A&A, 535, A128Battersby C., et al., 2020, ApJS, 249, 35Beltr´an M. T., de Wit W. J., 2016, A&ARv, 24, 6Beuther H., Schilke P., Sridharan T. K., Menten K. M., WalmsleyC. M., Wyrowski F., 2002, A&A, 383, 892Beuther H., Walsh A. J., Johnston K. G., el al 2017, A&A, 603,A10Beuther H., Mottram J. C., Ahmadi A., et al 2018, A&A, 617,A100Churchwell E., et al., 2009, PASP, 121, 213Clark P. C., Glover S. C. O., Ragan S. E., et al 2013, ApJ, 768,L34Dale J. E., Kruijssen J. M. D., Longmore S. N., 2019, MNRAS,486, 3307Espinoza P., Selman F. J., Melnick J., 2009, A&A, 501, 563Federrath C., Klessen R. S., 2012, ApJ, 761, 156Federrath C., et al., 2016, ApJ, 832, 143Fendt C., Zinnecker H., 1998, A&A, 334, 750Figer D. F., McLean I. S., Morris M., 1999, ApJ, 514, 202Figer D. F., Najarro F., Gilmore D., et al 2002, ApJ, 581, 258Giannetti A., et al., 2017, A&A, 606, L12Ginsburg A., Henkel C., Ao Y., et al 2016, A&A, 586, A50Ginsburg A., Bally J., Barnes A., et al 2018, ApJ, 853, 171Goddi C., Ginsburg A., Maud L., Zhang Q., Zapata L., 2020,arXiv e-prints, p. arXiv:1805.05364Gravity Collaboration et al., 2019, A&A, 625, L10Hatchfield H. P., et al., 2020, ApJS, 251, 14Hennebelle P., Chabrier G., 2013, ApJ, 770, 150Henshaw J. D., Longmore S. N., Kruijssen J. M. D., et al 2016,MNRAS, 457, 2675Henshaw J. D., et al., 2019, MNRAS, 485, 2457Henshaw J. D., et al., 2020, Nature Astronomy, 4, 1064Higuchi A. E., Chibueze J. O., Habe A., Takahira K., Takano S.,2014, AJ, 147, 141Higuchi A. E., Hasegawa T., Saigo K., Sanhueza P., ChibuezeJ. O., 2015, ApJ, 815, 106Humphreys E., Miura R., Brogan C. L., Hibbard J., Hunter T. R.,Indebetouw R., 2016, in Proceedings of the 2016 ALMA Con-ference. p. 1Ilee J. D., Cyganowski C. J., Brogan C. L., Hunter T. R., ForganD. H., Haworth T. J., Clarke C. J., Harries T. J., 2018, ApJ,869, L24Immer K., Menten K. M., Schuller F., Lis D. C., 2012, A&A, 548,A120Immer K., Kauffmann J., Pillai T., Ginsburg A., Menten K. M.,2016, A&A, 595, A94Jackson J. M., et al., 2013, Publ. Astron. Soc. Australia, 30, e057Johnston K. G., Beuther H., Linz H., et al 2014, A&A, 568, A56Kauffmann J., Bertoldi F., Bourke T. L., et al 2008, A&A, 487,993Kauffmann J., Pillai T., Zhang Q., 2013, ApJ, 765, L35Kauffmann J., Pillai T., Zhang Q., Menten K. M., GoldsmithP. F., Lu X., Guzm´an A. E., 2017a, A&A, 603, A89Kauffmann J., Pillai T., Zhang Q., Menten K. M., GoldsmithP. F., Lu X., Guzm´an A. E., Schmiedeke A., 2017b, A&A,603, A90Kepley A. A., Tsutsumi T., Brogan C. L., Indebetouw R., YoonI., Mason B., Donovan Meyer J., 2020, PASP, 132, 024505Krieger N., et al., 2017, ApJ, 850, 77Kruijssen J. M. D., Longmore S. N., Elmegreen B. G., et al 2014,MNRAS, 440, 3370Kruijssen J. M. D., Dale J. E., Longmore S. N., 2015, MNRAS,447, 1059Kruijssen J. M. D., et al., 2019, MNRAS, 484, 5734 Krumholz M. R., Kruijssen J. M. D., 2015, MNRAS, 453, 739Krumholz M. R., McKee C. F., 2005, ApJ, 630, 250Krumholz M. R., Kruijssen J. M. D., Crocker R. M., 2017, MN-RAS, 466, 1213Lada C. J., Muench A. A., Rathborne J., Alves J. F., LombardiM., 2008, ApJ, 672, 410Lada C. J., Lombardi M., Alves J. F., 2010, ApJ, 724, 687Lada C. J., Forbrich J., Lombardi M., Alves J. F., 2012, ApJ,745, 190Leurini S., Codella C., L´opez-Sepulcre A., Gusdorf A., CsengeriT., Anderl S., 2014, A&A, 570, A49Li S., Wang J., Fang M., et al. 2019, ApJ, 878, 29Li S., et al., 2020, ApJ, 903, 119Lis D. C., Menten K. M., Serabyn E., Zylka R., 1994, ApJ, 423,L39Lis D. C., Serabyn E., Zylka R., Li Y., 2001, ApJ, 550, 761Longmore S. N., Rathborne J., Bastian N., et al 2012, ApJ, 746,117Longmore S. N., Bally J., Testi L., et al 2013, MNRAS, 429, 987Lu X., et al., 2019a, ApJS, 244, 35Lu X., et al., 2019b, ApJ, 872, 171Lu X., Cheng Y., Ginsburg A., et al. 2020, ApJ, 894, L14Lu X., et al., 2021, arXiv e-prints, p. arXiv:2101.07925Marsh K. A., Ragan S. E., Whitworth A. P., Clark P. C., 2016,MNRAS, 461, L16Marsh K. A., et al., 2017, MNRAS, 471, 2730Mart´ın-Pintado J., de Vicente P., Fuente A., Planesas P., 1997,ApJ, 482, L45Maud L. T., Cesaroni R., Kumar M. S. N., et al 2018, A&A, 620,A31McMullin J. P., Waters B., Schiebel D., et al 2007, in Shaw R. A.,Hill F., Bell D. J., eds, Astronomical Society of the PacificConference Series Vol. 376, Astronomical Data Analysis Soft-ware and Systems XVI. p. 127Mills E. A. C., Battersby C., 2017, ApJ, 835, 76Mills E. A. C., Butterfield N., Ludovici D. A., Lang C. C., OttJ., Morris M. R., Schmitz S., 2015, ApJ, 805, 72Mills E. A. C., Ginsburg A., Immer K., Barnes J. M., WiesenfeldL., Faure A., Morris M. R., Requena-Torres M. A., 2018, ApJ,868, 7Molinari S., et al., 2010, PASP, 122, 314Molinari S., et al., 2016, A&A, 591, A149M¨oller T., Endres C., Schilke P., 2017, A&A, 598, A7Morris M., Serabyn E., 1996, ARA&A, 34, 645Ossenkopf V., Henning T., 1994, A&A, 291, 943Padoan P., Nordlund ˚A., 2011, ApJ, 730, 40Padoan P., Federrath C., Chabrier G., Evans N. J. I., JohnstoneD., Jørgensen J. K., McKee C. F., Nordlund ˚A., 2014, inBeuther H., Klessen R. S., Dullemond C. P., Henning T.,eds, Protostars and Planets VI. p. 77 ( arXiv:1312.5365 ),doi:10.2458/azu˙uapress˙9780816531240-ch004Pillai T., Kauffmann J., Tan J. C., et al 2015, ApJ, 799, 74Price-Whelan A. M., et al., 2018, AJ, 156, 123Qin S.-L., Zhao J.-H., Moran J. M., Marrone D. P., Patel N. A.,Wang J.-J., Liu S.-Y., Kuan Y.-J., 2008, ApJ, 677, 353Rathborne J. M., Longmore S. N., Jackson J. M., et al 2014a,ApJ, 786, 140Rathborne J. M., Longmore S. N., Jackson J. M., et al 2014b,ApJ, 795, L25Rathborne J. M., et al., 2015, ApJ, 802, 125Reid M. J., Menten K. M., Brunthaler A., et al. 2019, ApJ, 885,131Robitaille T., Bressert E., 2012, APLpy: Astronomical Plot-ting Library in Python, Astrophysics Source Code Library(ascl:1208.017)Rodr´ıguez L. F., Zapata L. A., 2013, ApJ, 767, L13S´anchez-Monge ´A., L´opez-Sepulcre A., Cesaroni R., WalmsleyC. M., Codella C., Beltr´an M. T., Pestalozzi M., MolinariMNRAS , 1–23 (2021) tar formation in ‘the Brick’ S., 2013, A&A, 557, A94Sanchez-Monge A., Schilke P., et al 2018, A&A, 609, A101Sanhueza P., Jackson J. M., Foster J. B., et al. 2012, ApJ, 756,60Schilke P., Walmsley C. M., Pineau des Forets G., Flower D. R.,1997, A&A, 321, 293Schneider F. R. N., et al., 2014, ApJ, 780, 117Tang Y., et al., 2020a, arXiv e-prints, p. arXiv:2008.12351Tang Y., Wang Q. D., Wilson G. W., 2020b, arXiv e-prints, p.arXiv:2008.12361Tsuboi M., Miyazaki A., Uehara K., 2015, PASJ, 67, 90Walker D. L., Longmore S. N., Bastian N., et al 2015, MNRAS,449, 715Walker D. L., Longmore S. N., Bastian N., et al 2016, MNRAS,457, 4536Walker D. L., Longmore S. N., Zhang Q., et al 2018, MNRAS,474, 2373Wang K., et al., 2014, MNRAS, 439, 3275Zhang Q., Hunter T. R., Brand J., Sridharan T. K., Cesaroni R.,Molinari S., Wang J., Kramer M., 2005, ApJ, 625, 864Zhang Q., Wang Y., Pillai T., Rathborne J., 2009, ApJ, 696, 268Zhang Q., Wang K., Lu X., Jim´enez-Serra I., 2015, ApJ, 804, 141Zoccali M., Valenti E., Surot F., Gonzalez O. A., Ren-zini A., Valenzuela Navarro A., 2021, arXiv e-prints, p.arXiv:2101.04022MNRAS000
Barnes A. T., Longmore S. N., Battersby C., el al 2017, MNRAS,469, 2263Barnes A. T., Longmore S. N., Avison A., et al 2019, MNRAS,486, 283Battersby C., Bally J., Ginsburg A., et al 2011, A&A, 535, A128Battersby C., et al., 2020, ApJS, 249, 35Beltr´an M. T., de Wit W. J., 2016, A&ARv, 24, 6Beuther H., Schilke P., Sridharan T. K., Menten K. M., WalmsleyC. M., Wyrowski F., 2002, A&A, 383, 892Beuther H., Walsh A. J., Johnston K. G., el al 2017, A&A, 603,A10Beuther H., Mottram J. C., Ahmadi A., et al 2018, A&A, 617,A100Churchwell E., et al., 2009, PASP, 121, 213Clark P. C., Glover S. C. O., Ragan S. E., et al 2013, ApJ, 768,L34Dale J. E., Kruijssen J. M. D., Longmore S. N., 2019, MNRAS,486, 3307Espinoza P., Selman F. J., Melnick J., 2009, A&A, 501, 563Federrath C., Klessen R. S., 2012, ApJ, 761, 156Federrath C., et al., 2016, ApJ, 832, 143Fendt C., Zinnecker H., 1998, A&A, 334, 750Figer D. F., McLean I. S., Morris M., 1999, ApJ, 514, 202Figer D. F., Najarro F., Gilmore D., et al 2002, ApJ, 581, 258Giannetti A., et al., 2017, A&A, 606, L12Ginsburg A., Henkel C., Ao Y., et al 2016, A&A, 586, A50Ginsburg A., Bally J., Barnes A., et al 2018, ApJ, 853, 171Goddi C., Ginsburg A., Maud L., Zhang Q., Zapata L., 2020,arXiv e-prints, p. arXiv:1805.05364Gravity Collaboration et al., 2019, A&A, 625, L10Hatchfield H. P., et al., 2020, ApJS, 251, 14Hennebelle P., Chabrier G., 2013, ApJ, 770, 150Henshaw J. D., Longmore S. N., Kruijssen J. M. D., et al 2016,MNRAS, 457, 2675Henshaw J. D., et al., 2019, MNRAS, 485, 2457Henshaw J. D., et al., 2020, Nature Astronomy, 4, 1064Higuchi A. E., Chibueze J. O., Habe A., Takahira K., Takano S.,2014, AJ, 147, 141Higuchi A. E., Hasegawa T., Saigo K., Sanhueza P., ChibuezeJ. O., 2015, ApJ, 815, 106Humphreys E., Miura R., Brogan C. L., Hibbard J., Hunter T. R.,Indebetouw R., 2016, in Proceedings of the 2016 ALMA Con-ference. p. 1Ilee J. D., Cyganowski C. J., Brogan C. L., Hunter T. R., ForganD. H., Haworth T. J., Clarke C. J., Harries T. J., 2018, ApJ,869, L24Immer K., Menten K. M., Schuller F., Lis D. C., 2012, A&A, 548,A120Immer K., Kauffmann J., Pillai T., Ginsburg A., Menten K. M.,2016, A&A, 595, A94Jackson J. M., et al., 2013, Publ. Astron. Soc. Australia, 30, e057Johnston K. G., Beuther H., Linz H., et al 2014, A&A, 568, A56Kauffmann J., Bertoldi F., Bourke T. L., et al 2008, A&A, 487,993Kauffmann J., Pillai T., Zhang Q., 2013, ApJ, 765, L35Kauffmann J., Pillai T., Zhang Q., Menten K. M., GoldsmithP. F., Lu X., Guzm´an A. E., 2017a, A&A, 603, A89Kauffmann J., Pillai T., Zhang Q., Menten K. M., GoldsmithP. F., Lu X., Guzm´an A. E., Schmiedeke A., 2017b, A&A,603, A90Kepley A. A., Tsutsumi T., Brogan C. L., Indebetouw R., YoonI., Mason B., Donovan Meyer J., 2020, PASP, 132, 024505Krieger N., et al., 2017, ApJ, 850, 77Kruijssen J. M. D., Longmore S. N., Elmegreen B. G., et al 2014,MNRAS, 440, 3370Kruijssen J. M. D., Dale J. E., Longmore S. N., 2015, MNRAS,447, 1059Kruijssen J. M. D., et al., 2019, MNRAS, 484, 5734 Krumholz M. R., Kruijssen J. M. D., 2015, MNRAS, 453, 739Krumholz M. R., McKee C. F., 2005, ApJ, 630, 250Krumholz M. R., Kruijssen J. M. D., Crocker R. M., 2017, MN-RAS, 466, 1213Lada C. J., Muench A. A., Rathborne J., Alves J. F., LombardiM., 2008, ApJ, 672, 410Lada C. J., Lombardi M., Alves J. F., 2010, ApJ, 724, 687Lada C. J., Forbrich J., Lombardi M., Alves J. F., 2012, ApJ,745, 190Leurini S., Codella C., L´opez-Sepulcre A., Gusdorf A., CsengeriT., Anderl S., 2014, A&A, 570, A49Li S., Wang J., Fang M., et al. 2019, ApJ, 878, 29Li S., et al., 2020, ApJ, 903, 119Lis D. C., Menten K. M., Serabyn E., Zylka R., 1994, ApJ, 423,L39Lis D. C., Serabyn E., Zylka R., Li Y., 2001, ApJ, 550, 761Longmore S. N., Rathborne J., Bastian N., et al 2012, ApJ, 746,117Longmore S. N., Bally J., Testi L., et al 2013, MNRAS, 429, 987Lu X., et al., 2019a, ApJS, 244, 35Lu X., et al., 2019b, ApJ, 872, 171Lu X., Cheng Y., Ginsburg A., et al. 2020, ApJ, 894, L14Lu X., et al., 2021, arXiv e-prints, p. arXiv:2101.07925Marsh K. A., Ragan S. E., Whitworth A. P., Clark P. C., 2016,MNRAS, 461, L16Marsh K. A., et al., 2017, MNRAS, 471, 2730Mart´ın-Pintado J., de Vicente P., Fuente A., Planesas P., 1997,ApJ, 482, L45Maud L. T., Cesaroni R., Kumar M. S. N., et al 2018, A&A, 620,A31McMullin J. P., Waters B., Schiebel D., et al 2007, in Shaw R. A.,Hill F., Bell D. J., eds, Astronomical Society of the PacificConference Series Vol. 376, Astronomical Data Analysis Soft-ware and Systems XVI. p. 127Mills E. A. C., Battersby C., 2017, ApJ, 835, 76Mills E. A. C., Butterfield N., Ludovici D. A., Lang C. C., OttJ., Morris M. R., Schmitz S., 2015, ApJ, 805, 72Mills E. A. C., Ginsburg A., Immer K., Barnes J. M., WiesenfeldL., Faure A., Morris M. R., Requena-Torres M. A., 2018, ApJ,868, 7Molinari S., et al., 2010, PASP, 122, 314Molinari S., et al., 2016, A&A, 591, A149M¨oller T., Endres C., Schilke P., 2017, A&A, 598, A7Morris M., Serabyn E., 1996, ARA&A, 34, 645Ossenkopf V., Henning T., 1994, A&A, 291, 943Padoan P., Nordlund ˚A., 2011, ApJ, 730, 40Padoan P., Federrath C., Chabrier G., Evans N. J. I., JohnstoneD., Jørgensen J. K., McKee C. F., Nordlund ˚A., 2014, inBeuther H., Klessen R. S., Dullemond C. P., Henning T.,eds, Protostars and Planets VI. p. 77 ( arXiv:1312.5365 ),doi:10.2458/azu˙uapress˙9780816531240-ch004Pillai T., Kauffmann J., Tan J. C., et al 2015, ApJ, 799, 74Price-Whelan A. M., et al., 2018, AJ, 156, 123Qin S.-L., Zhao J.-H., Moran J. M., Marrone D. P., Patel N. A.,Wang J.-J., Liu S.-Y., Kuan Y.-J., 2008, ApJ, 677, 353Rathborne J. M., Longmore S. N., Jackson J. M., et al 2014a,ApJ, 786, 140Rathborne J. M., Longmore S. N., Jackson J. M., et al 2014b,ApJ, 795, L25Rathborne J. M., et al., 2015, ApJ, 802, 125Reid M. J., Menten K. M., Brunthaler A., et al. 2019, ApJ, 885,131Robitaille T., Bressert E., 2012, APLpy: Astronomical Plot-ting Library in Python, Astrophysics Source Code Library(ascl:1208.017)Rodr´ıguez L. F., Zapata L. A., 2013, ApJ, 767, L13S´anchez-Monge ´A., L´opez-Sepulcre A., Cesaroni R., WalmsleyC. M., Codella C., Beltr´an M. T., Pestalozzi M., MolinariMNRAS , 1–23 (2021) tar formation in ‘the Brick’ S., 2013, A&A, 557, A94Sanchez-Monge A., Schilke P., et al 2018, A&A, 609, A101Sanhueza P., Jackson J. M., Foster J. B., et al. 2012, ApJ, 756,60Schilke P., Walmsley C. M., Pineau des Forets G., Flower D. R.,1997, A&A, 321, 293Schneider F. R. N., et al., 2014, ApJ, 780, 117Tang Y., et al., 2020a, arXiv e-prints, p. arXiv:2008.12351Tang Y., Wang Q. D., Wilson G. W., 2020b, arXiv e-prints, p.arXiv:2008.12361Tsuboi M., Miyazaki A., Uehara K., 2015, PASJ, 67, 90Walker D. L., Longmore S. N., Bastian N., et al 2015, MNRAS,449, 715Walker D. L., Longmore S. N., Bastian N., et al 2016, MNRAS,457, 4536Walker D. L., Longmore S. N., Zhang Q., et al 2018, MNRAS,474, 2373Wang K., et al., 2014, MNRAS, 439, 3275Zhang Q., Hunter T. R., Brand J., Sridharan T. K., Cesaroni R.,Molinari S., Wang J., Kramer M., 2005, ApJ, 625, 864Zhang Q., Wang Y., Pillai T., Rathborne J., 2009, ApJ, 696, 268Zhang Q., Wang K., Lu X., Jim´enez-Serra I., 2015, ApJ, 804, 141Zoccali M., Valenti E., Surot F., Gonzalez O. A., Ren-zini A., Valenzuela Navarro A., 2021, arXiv e-prints, p.arXiv:2101.04022MNRAS000 , 1–23 (2021) D. L. Walker et al. h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
16 - 2223 - 2929 - 36 km/s (a) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
23 - 2929 - 3637 - 42 km/s (b) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
43 - 4950 - 5656 - 62 km/s (c) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
29 - 4243 - 5656 - 69 km/s (d)
Figure 9.
Three-colour figures showing the SiO (5-4) emission at different velocities. (a) : 16 – 22 (blue), 23 – 29 (green), and 29 – 36(red) km s − , (b) : 23 – 39 (blue), 29 – 36 (green), and 37 – 42 (red) km s − , (c) : 43 – 49 (blue), 50 – 56 (green), and 56 – 62 (red) kms − , (d) : 29 – 42 (blue), 43 – 56 (green), and 56 – 69 (red) km s − . Continuum sources are highlighted by white ellipses, the extent ofwhich corresponds to the structures determined using dendrograms. Each continuum source is also numbered. The yellow crosses showthe position of water masers from Lu et al. (2019b). MNRAS , 1–23 (2021) tar formation in ‘the Brick’ h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J ) -18 - -12 km/s (a) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J ) -11 - -5 km/s (b) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J ) -4 - 2 km/s (c) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J ) (d) Figure 10.
Figures showing the integrated SiO (5-4) emission in different velocity ranges. (a) : -18 – -12 km s − , (b) : -11 – -5 km s − , (c) : -4 – 2 km s − , (d) : 2 – 9 km s − . Continuum sources are highlighted by cyan ellipses, the extent of which corresponds to thestructures determined using dendrograms. Each continuum source is also numbered. The grey crosses show the position of water masersfrom Lu et al. (2019b).MNRAS000
Figures showing the integrated SiO (5-4) emission in different velocity ranges. (a) : -18 – -12 km s − , (b) : -11 – -5 km s − , (c) : -4 – 2 km s − , (d) : 2 – 9 km s − . Continuum sources are highlighted by cyan ellipses, the extent of which corresponds to thestructures determined using dendrograms. Each continuum source is also numbered. The grey crosses show the position of water masersfrom Lu et al. (2019b).MNRAS000 , 1–23 (2021) D. L. Walker et al. h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
10 - 15 km/s (a) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
16 - 22 km/s (b) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
23 - 29 km/s (c) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
29 - 36 km/s (d)
Figure 11.
As in Figure 10. (a) : 10 – 15 km s − , (b) : 16 – 22 km s − , (c) : 23 – 29 km s − , (d) : 29 – 36 km s − .MNRAS , 1–23 (2021) tar formation in ‘the Brick’ h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
37 - 42 km/s (a) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
43 - 49 km/s (b) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
50 - 56 km/s (c) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
56 - 63 km/s (d)
Figure 12.
As in Figure 10. (a) : 37 – 42 km s − , (b) : 43 – 49 km s − , (c) : 50 – 56 km s − , (d) : 56 – 62 km s − .MNRAS000
As in Figure 10. (a) : 37 – 42 km s − , (b) : 43 – 49 km s − , (c) : 50 – 56 km s − , (d) : 56 – 62 km s − .MNRAS000 , 1–23 (2021) D. L. Walker et al. h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
63 - 69 km/s (a) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
70 - 76 km/s (b) h m s s s -28°42'00"10"20"30" Right Ascension (J2000) D e c li n a t i o n ( J )
77 - 83 km/s (c)
Figure 13.
As in Figure 10. (a) : 63 – 69 km s − , (b) : 70 – 76 km s − , (c) : 77 – 83 km s − .MNRAS , 1–23 (2021) tar formation in ‘the Brick’ AUTHOR AFFILIATIONS Department of Physics, University of Connecticut, 196A Auditorium Road, Storrs, CT 06269 USA National Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo, 181-8588, Japan Joint ALMA Observatory, Alonso de C´ordova 3107, Vitacura, Santiago, Chile Astrophysics Research Institute, Liverpool John Moores University, IC2, 146 Brownlow Hill, Liverpool, L3 5RF, UnitedKingdom CASA, University of Colorado, 389-UCB, Boulder, CO 80309 Department of Astronomy, University of Florida, PO Box 112055, Gainesville, FL 32611, USA Astronomisches Rechen-Institut, Zentrum f¨ur Astronomie der Universit¨at Heidelberg, M¨onchhofstraße 12-14, 69120 Heidel-berg, Germany Center for Astrophysics | Harvard & Smithsonian, 60 Garden Street, Cambridge, MA 02138, USA Max-Planck Institute for Astronomy, K¨onigstuhl 17, 69117 Heidelberg, Germany University of Vienna, Department of Astrophysics, T¨urkenschanzstrasse 17, 1180 Wien, Austria Radcliffe Institute for Advanced Study, Harvard University, 10 Garden Street, Cambridge, MA 02138, USA Argelander-Institut f¨ur Astronomie, Universit¨at Bonn, Auf dem H¨ugel 71, 53121, Bonn, Germany Leiden Observatory, Leiden University, PO Box 9513, NL 2300 RA Leiden, the Netherlands Kavli Institute for Astronomy and Astrophysics, Peking University, Beijing 100871, China Department of Astronomy, School of Physics, Peking University, Beijing 100871, China SOFIA Science Center, USRA, NASA Ames Research Center, Moffett Field CA 94045, USA Haystack Observatory, Massachusetts Institute of Technology, 99 Millstone Road, Westford, MA 01886, USA Department of Physics and Astronomy, University of Kansas, 1251 Wescoe Hall Drive, Lawrence, KS 66045, USA Boston University Astronomy Department, 725 Commonwealth Avenue, Boston, MA 02215, USA
This paper has been typeset from a TEX/L A TEX file prepared by the author.MNRAS000