Arxiv:2108.12554V1

Home , Infrared Science Archive

arXiv:2108.12554v1 [astro-ph.SR] 28 Aug 2021 [email protected] orsodn uhr hwiChen Zhiwei author: Corresponding (e.g., a ik thg crto ae ( rates accretion high circumstel- continu- at gas-rich progenitors the disks from lower-mass lar material as accreting form ously accumu- stars high- and that favour accretion massive mass to appear disk The evidence for- of observational basic lated models stars. the formation high-mass on star consensus of no mechanism still feedbacks. mation various is their there through However, environment their on et´ n&d i 2016 Wit de Beltrán & ∗ ihms tr ( stars High-mass yee sn L using 2021 Typeset 31, August version Draft aoy hl,Porm 7.-10AB,ad077.C-0174( and 073.C-0170(A+B), Ob Southern Programs European Chile, the vatory, at collected observations on Based ooaae l 2010 al. et Hosokawa Keywords: earlies outb the in at frequent rate are accretion ph bursts The ourbusrt accretion and minor quiescent q yr. that between yr suggests 26 6 contrast over the a by than 83% lower separated is much mag, ourbusrt 2 accretion about of of magnitudes outburst with yr) siae rmteknmtc fH of kinematics the from estimated fteseta nrydsrbtoso 1 I nte20 (quiesc 2005 the wh MIR, in M17 phase MIR of parameters re-brightening M17 ( stellar of basic and the distributions decreasing constrains energy outburst) the 20 spectral during mid the since MIR of phase M17 re-brightening onto the 1993.03 rate and during 2010, phase mid decreasing H to the 2004 stages, mid three significant during into show split 2019 to be 1993 can during data multi-wavelength The M17. nteaceinotus,alre tla aisi eurdt pro to required is radius stellar larger a outburst, accretion the In h 07eoh h enhanced The epoch. 2017 the 3 M hwiCe ( Chen Zhiwei nttt eAtooá nvria a´lc e Norte Católica Astronom´ıa, del Universidad de Instituto 2 erpr h icvr famsiepootrM7MRembedded MIR M17 protostar massive a of discovery the report We ae msinvrain oehrwt h I aito,indicate variation, MIR the with together variation, emission maser O ∗ ∼ 1 upeMuti bevtr,CieeAaeyo Sciences, of Academy Chinese Observatory, Mountain Purple 2 srnmshsIsiu,Rh–nvri¨tBcu,Univ Ruhr–Universität Bochum, Institut, Astronomisches 5 A T . 4 1. E X M niiulojcs(1 MIR) (M17 objects individual crto,aceindss–Sas asv tr:frain–S – formation Stars: – massive Stars: – disks accretion Accretion, 1 I:AMsiePootrwt utpeAceinOutbur Accretion Multiple with Protostar Massive A MIR: M17 INTRODUCTION M twocolumn ⊙ with ) ∗ ; 陈 & ot ta.2018 al. et Motte 志 8 维 ; M M ) ˙ , acc amelee l 2016 al. Haemmerlé et ⊙ 1 tl nAASTeX631 in style aesrn impacts strong have ) e u ( Sun Wei ∼ 1 & . 1 M × ˙ 10 acc 10 .Observa- ). − 2 孙 aes 1 I hw w crto ubrt (∆ outbursts accretion two shows MIR M17 masers. O 4 − uigotus nue h uioiyotus ∆ outburst luminosity the induces outburst during 玮 upn hn( Chen Xuepeng M 5 M ), ⊙ 1 ⊙ yr ofChini, Rolf yr − ABSTRACT 1 − ) A). ; 1 ser- nte20 pc and epoch 2005 the in vnd nao 60 ail 20Atfgsa Chile Antofagasta, 1280 Casilla 0610, Angamos Avenida , ierneo rtselrms,fo o-as(e.g., low-mass from mass, a protostellar across of extends range accretion episodic wide for evidence tional eetfo h pcrsoial ofimdbloating comfirmed dif- spectroscopically evolution, pre-main-sequence the the from of times ferent end multiple the occur to can MYSOs up of bloating short ( This stars massive the evolved usually of by are occupied regions which colder (HRD), diagram but Hertzsprung-Russel luminous, more toward excur- rapid sions experience MYSOs accretion, during bursting MYSOs; each of tracks evolutionary influence cretions of mass ( main-sequence MYSOs zero-age a of stellar contribute fraction young can substantial massive accretions percent bursting of few (MYSOs), phases a objects accretion only entire are the accretions of bursting multiple of 2017 20 2014 al. et Audard ,3 2, 陈 ri¨ tsrse10 40 ohm Germany Bochum, 44801 ersitätsstrasse 150, 学 M atnHaas, Martin ; 鹏 ⊙ 0YaHaRa,202 ajn,China Nanjing, 210023 Road, YuanHua 10 utre l 2017 al. et Hunter ) (e.g., 1 tgso asv trformation. star massive of stages t ee ta.2019b al. et Meyer s.Teeteeyuho 1 MIR M17 of youth extreme The ase. c sa nemdaems protostar intermediate-mass an is ich duce ai ta.2015 al. et Tapia isetpae h rcintime fraction The phase. uiescent omd20,teqisetphase quiescent the 2004, mid to n)ad21 pc (accretion epoch 2017 and ent) rti aibewt amplitude with variable is urst i-R(I)vrain,which variations, (MIR) mid-IR aitv rnfrmodeling transfer Radiative . M M 2 ˙ ˙ o eiw ohigh-mass to review) a for , acc hb in ( Jiang Zhibo acc 0utl o.Te2 GHz 22 The now. untill 10 nahtmlclrcr in core molecular hot a in nehne ikaccretion disk enhanced an ossetwt h value the with consistent ∼ as rtsas–Stars: – protostars tars: .Atog h oa periods total the Although ). 1 . 7 .Mroe,brtn ac- bursting Moreover, ). × ; 10 sts aat aat tal. et Garatti o Caratti 江治波 − L 3 ∗ ee ta.2019a al. et Meyer M ≈ t ⊙ 7600 ∼ ), yr 1 9 − and − L 1 20 ⊙ in . ∼ 10 − ). 2 Chen, Zhiwei et al.

′ of MYSOs during the contraction phase toward the The wide-field JHKs- and L -band images were ob- main sequence (Chen et al. 2015; Ram´ırez-Tannus et al. tained with ISAAC at the Very Large Telescope (VLT) 2017). The discovery of burst-induced photometric vari- in September 2002 and May to September 2004, respec- ability for MYSOs is impossible at optical and difficult tively. The pixel scale is 0′′. 148 and 0′′. 07, and the Full at IR wavelengths, due to severe extinction and large Width at Half Maximum (FWHM) is 0′′. 6 and 0′′. 5 for ′ distances of MYSOs. the JHKs- and L -band images, respectively (see more At a distance ≈ 2.0kpc (Xuetal. 2011; details of observations in Hoffmeister et al. 2008). These ′ Chibueze et al. 2016; Wu et al. 2019), M17 is among JHKsL -band images and photometric data were pub- the best laboratories in the Galaxy for investigating lished by Hoffmeister et al. (2008). formations of high-mass stars. Several efforts focused The MIR images were carried out toward the M17 on the high-mass stars in M17 attempt to measure SW PDR (e.g., Crétéet al. 1999; Pérez-Beaupuits et al. the physical properties of MYSOs, e.g., Hanson et al. 2012) with the VLT Imager and Spectrometer for (1997); Nielbock et al. (2001); Hoffmeister et al. (2006); mid-InfRared (VISIR; Lagage et al. 2004) at the VLT Ram´ırez-Tannus et al. (2017). Meanwhile detailed stud- in 2006 April through the SiC filter (λ = 11.8 µm, ies for several peculiar MYSOs provide better con- ∆λ = 2.1 µm). The data were taken with the straints on their evolutionary stages and circumstellar VISIR img obs AutoChopNod template. The photomet- environments, e.g., Chini et al. (2000); Kassis et al. ric standard star was HD 178345 during this observation. (2002); Chini et al. (2004, 2006); Nielbock et al. (2007); The data were reduced and calibrated using the EsoRe- Chen et al. (2015); Lim et al. (2020). Despite the clas- flex workflows (Freudling et al. 2013) and the VISIR sified MYSOs (mostly IR bright), it is very likely that pipeline recipes 4.3.71. The image is of good quality unknown high-mass protostars are deeply embedded in with FWHM ≈ 0′′. 4, reaching the diffraction limit. dense cores within M17. There are nine sites of 22GHz 2 H2O masers in M17 (Johnson et al. 1998; Breen et al. 2.2. The Infrared Space Observatory (ISO) data 2010), however, only one is likely associated with the The archival imaging data of ISO (Kessler et al. 1996) known massive protostar IRS5A (Chen et al. 2015); the were obtained for the M17 SW PDR from the ISO driving sources of the remaining H2O masers are still Data Archive3. The ISO observations for the M17 SW unclear. PDR were carried out in 1997 March and published in The expanding shell traced by the H2O maser spots Crétéet al. (1999), which aim at the interstellar medium (Chibueze et al. 2016, hereafter CJO16) motivates the inside the PDR. The calibrated ISO/ISOCAM images search for the driving source. The coordinate of the pu- at the three broad filters (LW1, LW4, LW6) with cen- tative driving source is determined as R.A.(J2000) = tral wavelengths of 4.5, 6.0, and 7.7 µm are used in this h m s ◦ ′ ′′ 18 20 23.017, Decl.(J2000) = −16 11 47. 98 with posi- paper. tional errors of the order of 1mas (CJO16), which co- incides with a faint point source (hereafter M17 MIR) 2.3. The Spitzer data with Spitzer 4.5 and 5.8 µm magnitudes of 9.97 and The Galactic Legacy Infrared Mid-Plane Sur- 7.82 mag, respectively (see more descriptions about vey Extraordinaire (GLIMPSE; Benjamin et al. 2003; GLIMPSE in Sect. 2.3). On the other hand, none of Churchwell et al. 2009) is a legacy science program of the compact radio sources reported by Rodr´ıguez et al. the Spitzer Space Telescope that surveyed the inner (2012) matches M17 MIR, which eliminates the associ- Galactic plane in the 3.6, 4.5, 5.8 and 8.0 µm bands ation with a hyper- or ultra-compact H II region, and with the IR camera IRAC. The angular resolution at thus indicates a very early stage for M17 MIR. the four bands (I1,I2,I3,and I4) are about 2′′. 5. The This paper reports the analyses for the multi- GLIMPSE mosaicked images of M17 were obtained from wavelenth data available for M17 MIR. The multi- the NASA/IPAC Infrared Science Archive4; the median wavelength data are described in Sect. 2. The analyses observation date was September 2005. The GLIMPSE for these data and the results are presented in Sect. 3. The variability of M17 MIR is classified in Sect. 4. The derived properties of M17 MIR are discussed in Sect. 5, 1 http://www.eso.org/sci/software/pipelines/ and the conclusions are summarized in Sect. 6. 2 Based on observations with ISO, an ESA project with instruments funded by ESA Member States (especially the PI countries: France, Germany, the Netherlands and the United Kingdom) and with the participation of ISAS and NASA. 2. OBSERVATIONS 3 https://www.cosmos.esa.int/web/iso/access-the-archive 2.1. IR Imaging Data from the Very Large Telescope 4 http://irsa.ipac.caltech.edu/frontpage/ Bursting massive protostar M17 MIR 3

3.8 micron 4.5 micron 5.7 micron 7.7 micron 7.8 micron

M17 MIR M17 MIR M17 MIR M17 MIR M17 MIR

Anon 3 Anon 3 Anon 3 Anon 3 Anon 3

38.0 11.8 micron 19.7 micron 37.1 micron 70 micron 100 micron

42.0 M17 MIR M17 MIR M17 MIR M17 MIR M17 MIR 46.0

-16:11:50.0

54.0 Anon 3 Anon 3 Anon 3 Anon 3 Anon 3 18:20:23.6 18:20:22.8

Figure 1. Multi-wavelength images centered on M17 MIR. Proper motions of H2O maser spots (CJO16) are overlaid on the image at 4.5 µm. point sources were retrieved from the GLIMPSE I Spring 6.4, 6.5, and 12.0 arcsec, respectively. Only the WISE 07 Archive5. W1 and W2 data of M17 are considered in this work, M17 was revisited by Spizter in the warm phase in because the M17 H II region is saturated in W3 and W4 July 2017. During these observations, only the two bands. With the primary aim of studying Near-Earth bluest bands (3.6 and 4.5 µm) were available, and the Objects, NEOWISE observations resumed in December flux calibrated images (program ID: 13117, observer: 2013 and continues to date, after the telescope’s cryogen Robert Benjamin) were obtained from the Spitzer Her- tanks were depleted. Due to the orbit of the telescope, itage Archive6. WISE/NEOWISE observations visit M17 twice a year, in March and September. From the NEOWISE 2020 2.4. The Herschel data data release containing W1 and W2 observations from In order to cover the far-IR regime, we incorporated December 2013 until December 2019, the single expo- the publicly available imaging observations from the sure data of M17 collected in 6 years (12 epochs) are Herschel Space Observatory (Herschel) and its PACS in- analyzed in this work, in addition to the WISE all-sky strument at 70, 100, and 160 µm (Poglitsch et al. 2010). survey image of M17 (2 epochs). The Herschel data used in this work are Level 3.0 prod- The single exposure images of WISE/NEOWISE mis- sion toward M17 were retrieved from the WISE hosted ucts (eight observations with IDs from 1342192767 to 8 1342192774) provided by the Herschel Science Archive7. on IRSA , with the service of WISE/NEOWISE Coad- The pixel scales are 1′′. 6, 1′′. 6, and 3′′. 2 at 70, 100, and der. The single exposures were coadded to obtain 160 µm, and the measured resolutions are 8′′. 0, 8′′. 0, and higher-quality W1/W2 images with enhanced resolution ′′ 12′′. 0, respectively. of 0. 6875/pix. Finally WISE/NEOWISE images from 14 epochs of M17 were used in this work. Aperture ′′ 2.5. The WISE/NEOWISE data photometry with a fixed radius of 8.5 pixel (= 5. 8) is applied to all WISE/NEOWISE images with a FOV of The Wide-field Infrared Survey Explorer (WISE, ′ ′ 3 × 3 . Because WISE/NEOWISE W2 band is very Wright et al. 2010) is a 40 cm telescope in a low-earth similar to the Spitzer I2 band, the photometric re- orbit that surveyed the entire sky in 2010 using four IR sults in multi-epochs are checked against the isolated bands at 3.4, 4.6, 12, and 22 µm. The angular resolu- ′′ star G015.0198-00.6768 located 33 south of the im- tion at the four bands (W1, W2, W3, and W4) are 6.1, age center. G015.0198-00.6768 is not cataloged in the ALLWISE catalog (Cutri & et al. 2014) and in the un- 5 see more descriptions in http://irsa.ipac.caltech.edu/Missions/spitzer.html 6 hosted at https://irsa.ipac.caltech.edu/Missions/spitzer.html 8 https://irsa.ipac.caltech.edu/Missions/wise.html 7 http://archives.esac.esa.int/hsa/whsa/ 4 Chen, Zhiwei et al.

WISE catalog (Schlaﬂy et al. 2019), but has a cataloged Table 1. IR ﬂux densities Spitzer I2 magnitude of 8.55 (Spitzer Science 2009).

Telescope λeff Epoch Flux Densities FWHM 2.6. SOFIA Data ′′ (µm) (mJy) ( ) 9 The SOFIA (Young et al. 2012) observations toward VLT 3.78 2004.09 1.91 ± 0.17 0.5 the M17 SW PDR were carried out with the FOR- Spitzer 4.50 2005.07 18.4 ± 2.2 2.0 CAST instrument (Herter et al. 2013) at 19 µm (λeff = Spitzer 5.71 2005.07 85.0 ± 8.0 2.0 19.7 µm; ∆λ = 5.5 µm) and 37 µm (λeff = 19.7 µm; ISO 7.72 1997.03 2.2 ± 0.3 × 102 3.7 ∆λ = 5.5 µm) in 2017 August 02 (PI: James M. De Spitzer 7.85 2005.07 1.4 ± 0.3 × 102 2.0 Buizer). The calibrated Level 3 images at 19 and VLT 11.80 2006.04 1.4 ± 0.07 × 102 0.4 37 µm were downloaded from the SOFIA Data Cy- 2 SOFIA 19.7 2017.08 5.1 ± 1.2 × 10 3.3 cle System 10. The resolutions at 19 and 37 µm are SOFIA 37.1 2017.08 7.3 ± 1.2 × 103 3.5 3′′. 3 and 3′′. 5, respectively (De Buizer et al. 2017). The Herschel 70 2010.03 < 7.7 × 104 8.0 SOFIA/FORCAST images fill the gap between the VLT/VISIR and Herschel images in wavelength cover- Note—Central wavelengths and effective bandwidths are age. adopted from the SVO Filter Profile Service (Rodrigo et al. 2012). 2.7. JCMT SCUBA2 sub-mm data The JCMT SCUBA2 sub-mm observations were taken source at 3 − 37 µm. At Herschel 70µm and 160µm, the at 450 µm and 850 µm in 2016 April 10 for M17, as a far-IR emission is dominated by the contamination from part of the JCMT large program “SCOPE: SCUBA-2 M17 IRS5 and M17 UC1 in northeast. At the IR loca- Continuum Observations of Pre-protostellar Evolution” tion of M17 MIR, M17 MIR is not resolved, likely due (JCMT program code: M16AL003). The pipeline re- to the strong ambient far-IR emission. For M17 MIR, duced and calibrated map at 450 µm and 850 µm of M17 we estimated an upper limit < 77Jy at 70 µm. were retrieved from Canadian Astronomy Data Centre The faint IR brightness of M17 MIR suggest that 11. M17 MIR is a low-luminosity object or the extinction to M17 MIR is quite high. By a quick examine on the IR 3. RESULTS images of M17 MIR, we think the high extinction is more 3.1. A very early high-mass protostellar object driving likely the reason for the faint IR brightness. M17 MIR’s a protostellar outflow brightness shows steep rise in the 3−8 µm range (see also Figure 8). However, M17 MIR is not visible in the ISO The earliest detection of M17 MIR was at 3.9 µm L broad filter image at 9.6 µm (see Figure 2a–f; Crétéet al. by the CFHT observations in December 1992 and May ′′ 1999). We know that the silicate absorption feature at 1993 with a spatial resolution of 1 (Giard et al. 1994). 9.7 µm can cause high and broad absorption to highly This object, named as source 2 by the authors, along obscured object at around this wavelength. The high with other six extremely red objects, are suggested to be extinction to M17 MIR can explain the non-detection young stars in very early evolutionary stages (see Table at 9.6 µm. 1 in Giard et al. 1994). Figure 1 presents the multi- The IR flux densities of M17 MIR are at least two wavelength images centered on M17 MIR. A summary orders of magnitude lower than those of the IR-bright of the IR flux densities obtained by the various facilities high-mass protostar IRS5A (Chen et al. 2015) within except the earliest CFHT observation is presented in the same M17 SW PDR. On the other hand, the out- Table 1. wards motions of the CJO16 H2O masers very likely The source sizes in the various IR images are in pro- trace the shocked gas accompanying the protostellar portion to FWHMs of the corresponding images. In the ′′ outflow driven by M17 MIR. In the earliest stage of 11.8µm image of 0. 4 resolution (= 800 AU), M17 MIR is high-mass protostars, the mass loss might evolve from a still a single source. M17 MIR seems as a compact point highly uncollimated expanding shell to a collimated outflow (e.g., Carrasco-González et al. 2015). The dynam- 9 SOFIA is jointly operated by the Universities Space Research ical age of the expanding shell driven by M17 MIR is Association, Inc. (USRA), under NASA contract NAS2-97001, 12.5 yr (CJO16), younger than the dynamical age (∼ 25 and the Deutsches SOFIA Institut (DSI) under DLR contract 50 OK 0901 to the University of Stuttgart. yr) of W75N(B)-VLA 2 (Carrasco-González et al. 2015). 10 https://dcs.arc.nasa.gov/ The faint IR brightness of M17 MIR and the young 11 http://www.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/en/jcmt/ dynamical age of the CJO16 H2O masers suggest that Bursting massive protostar M17 MIR 5

SCUBA2 850 micron 10:50.0

-16:11:00.0

10.0 3 1 20.0 3 1 30.0 4 2 40.0 2 Declination (J2000) 4

50.0 Intengrated line intensity [K km/s] [K intensity line Intengrated 12:00.0

25.0 18:20:23.0 21.0

Right Ascension (J2000)

Figure 2. JCMT-SCUBA2 850 µm map of M17 SW in left, and the HCO+ J = 9 → 8 integrated intensity map of M17 SW in right (Pérez-Beaupuits et al. 2015) . The 22 GHz water maser sources detected by Johnson et al. (1998) are denoted by crosses with numbers as same as Johnson et al. (1998). The white circle in both maps denote the position of M17 UC1. 7 −3 M17 MIR is a promising candidate for a high-mass pro- sity ncrit =9 × 10 cm and an excitation temperature tostar in the earliest stage. The analysis of the spectral Eup = 192.58K of the 802.458GHz HCO+ J = 9 → 8 energy distribution (SED) constructed from Table 1 will line (Pérez-Beaupuits et al. 2015), the dense core asso- help to better constrain the fundamental properties of ciated with M17 MIR is likely a candidate for hot molec- M17 MIR (Sect. 5.3). ular core (HMC; Cesaroni 2005). HMCs are commonly 23 The warm area close the central object emits radia- corresponding to a hydrogen column density NH & 10 tion mostly at MIR wavelengths, while the surround- or even higher (e.g. Bonfand et al. 2017). Although ing envelope of much larger size is expected to radiate a high-mass protostar embedded in an HMC is unde- at sub-mm to mm wavelengths due to a much lower tectable at short wavelengths due to high extinction, effective temperature. Figure 2 presents the 850µm the detection of the central protostar becomes possible dust continuum emission map and the HCO+ J = at MIR wavelengths. For instance, the IR counterpart 9 → 8 line integrated intensity map of M17 SW PDR. F4 of the HMC G9.62+0.19-Fappears at 3.8 µm L′-band M17 MIR, represented by the H2O maser source 2 de- and longer wavelengths (Linz et al. 2005). tected by VLA observations (Johnson et al. 1998, here- The multi-wavelength data of M17 MIR and the sub- after JDG98), is exactly located in the secondary peak mm to mm data of its surrounding medium indicates of the 850 µm dust continuum emission. This dust con- that M17 MIR is a very early high-mass protostar em- tinuum emission peak, along with the other two peaks, bedded in an HMC and is also driving a protostellar make up the northern condensation in M17 SW PDR outflow. that was firstly noticed by Stutzki & Guesten (1990). 4. VARIABILITY The three main components in the northern condensation in M17 SW PDR were studied in detail through 4.1. Infrared variability sub-mm and mm observations. Their physical proper- For M17 MIR, the photometric magnitude in the 1993 ties were estimated by Hobson et al. (1993) (their Ta- CFHT observation was 10.2 ± 0.2 (source 2 in Table 2; ble 4). M17 MIR is located within the FIR3 source Giard et al. 1994), while the magnitude in the 2004 VLT in Hobson et al. (1993). On the other hand, the inte- observation was 12.78 ± 0.14 (Hoffmeister et al. 2008). 1 grated line intensity map (range: 0 − 40kms− , peak The filter profiles in the two observations are almost 1 velocity ∼ +20kms− ) of the HCO+ J = 9 → 8 line identical. Table 2 lists the L-band magnitudes of the shows clumpy structures similar to the dust continuum embedded YSOs in the two observations. Interestingly, emission. A dense core whose size is comparable to M17 MIR is about 10 times fainter in the 2004.09 epoch the beam size of the observations (∼ 8′′, or 0.08 pc) than in the 1993.03 epoch, whereas the other sources are is coinciding with M17 MIR. Given the critical den- fairly stable. 6 Chen, Zhiwei et al. Spitzer/I1 2005-09 Spitzer/I1 2017-07 Table 2. L-band magnitudes of embedded YSOs in M17 7 7 SW during the 1993.03 and 2004.09 epochs. 5 5 3 M17 MIR 3 M17 MIR Name R.A. Decl. 1993.03 2004.09 (J2000) (J2000) (mag) (mag) 2=M17 MIR 18:20:23.02 -16:11:47.9 10.2 12.78 3 18:20:23.91 -16:11:43.4 10.8 10.38 5 18:20:24.10 -16:11:40.1 11.2 10.99 Spitzer/I2 2005-09 7 Spitzer/I2 2017-07 7 7 18:20:22.34 -16:11:28.7 8.8 8.76 11:30.0 35.0 Note—See Figure 3 for the positions of these sources 5 5 40.0 3 M17 MIR 3 M17 MIR on the Spitzer I1 and I2 images. 45.0 50.0 The VLT L′-band observations in 2004.09 were one 55.0 year earlier than the Spitzer observations in 2005.09. -16:12:00.0 By interpolating the Spitzer flux densities at I2 and I3 05.0 24.0 18:20:23.0 22.0 ′ to that at L under the assumption of a blackbody radi- Figure 3. Top panel: Spitzer/I1 images in 2005.09 (left) ′ ation, the flux density of M17 MIR at L band is about and 2017.07 (right). Bottom panel: Spitzer/I2 images in 2.2 mJy in the 2005.09 epoch. This predicted flux den- 2005.09 (left) and 2017.07 (right). The embedded YSOs 3, sity in 2005.09 is close to the 1.91 ± 0.17 mJy measured 5, and 7 are the same as in Table 2. in 2004.09. We therefore speculate that M17 MIR’s lu- ISO/LW1 1997-03 7 Spitzer/I2 2005-09 7 minosity is stable during 2004.09–2005.09. 11:30.0 M17 MIR shows variabilities in the Spitzer I1 and I2 35.0 images from 2005.09 and 2017.07 (see Figure 3). Com- 5 5 40.0 3 M17 MIR 3 M17 MIR paring the flux densities at I1 in the two epochs for an 45.0 ′′ aperture radius of 1. 6 we find an increase of a factor of 50.0 2.9. Similarly, the flux densities at I2 show an increase 55.0 of a factor of 3.3 in an aperture radius of 3′′. 2. In con- -16:12:00.0 trast, the other sources are stable at the two bands (see 05.0 24.0 18:20:23.0 22.0 Table 3). Figure 4. ISO/LW1 image in 1997.03 (left) and Spitzer/I2 The MIR variabilities of M17 MIR found for differ- image in 2005.09 (right). The embedded YSOs 3, 5, and 7 ent epochs motivate to compare its flux densities from are identical to those in Figure 3. the ISO observations in 1997.03 and the Spitzer observations in 2005.09 and 2017.07. Table 4 shows the flux I1 and I2 bands, are about 0.6 and 1.8mJy during the densities of M17 MIR and other non-variable sources 2005.09 and 2017.07 epochs, respectively. The extreme in the ISO/LW1 and Spitzer/I2 bands, which have al- faintness at W1 is consistent with the non-detection of most identical filter profiles centered at 4.5 µm. Figure 4 M17 MIR in the corresponding W1 images. compares the ISO LW1 and Spitzer I2 images obtained Figure 5 presents the WISE/NEOWISE W2 images in 1997.03 and 2005.09, respectively. Direct compar- centered on M17 MIR in 14 epoches. The pixel reso- ison from Figure 4 clearly indicate that M17 MIR is lution of 0′′. 6875 clearly separates M17 MIR from the significantly brighter than source 3 and 5 in the ISO nearby IR source Anon 3 in the southwest. The pho- LW1 image in 1997.03 epoch, while it becomes fainter tometric results of M17 MIR, source 7, and G015.0198- than source 3 and 5 in the Spitzer I2 image in 2005.09 00.6768 are tabulated in Table 5 for 14 epoches. In epoch. Given that source 3 and 5 are unchanged (see 2010.03, the W2 flux density of M17 MIR was 17.7 ± 4.4 context above), M17 MIR is varying between 1997.03 mJy, which is very close to the value 18.4 ± 2.0 at the and 2005.09. The aperture photometry of M17 MIR at Spitzer I2 band in 2005.09. Given the fact that the 4.5 µm band shows variability from 1997 to 2017. WISE/NEOWISE W2 band is very similar to Spitzer For M17, WISE/NEOWISE images in 14 epochs, in- I2 band, M17 MIR was roughly stable in brightness cluding 2010.03 and 2010.09, March and September in from 2004 to 2010. However, 6 months later M17 MIR 2014 to 2019, were analyzed in this work. M17 MIR was 2 times brighter in 2010.09, with a flux density of is only detected in the W2 band. The WISE W1 flux 33.7±4.4 mJy. It is very likely, that the re-brightening of densities, extrapolated from its flux densities at Spitzer M17 MIR started between 2010.03 and 2010.09. In the Bursting massive protostar M17 MIR 7

2010.03 2010.09 2014.03 2014.09 2015.03

7 7 7 7 7

M17 MIR M17 MIR M17 MIR M17 MIR M17 MIR

Anon 3 Anon 3 Anon 3 Anon 3 Anon 3

G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768

2015.09 2016.03 2016.09 2017.03 2017.09

7 7 7 7 7

M17 MIR M17 MIR M17 MIR M17 MIR M17 MIR

Anon 3 Anon 3 Anon 3 Anon 3 Anon 3

G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768

2018.03 2018.09 2019.03 2019.09

11:20.0 7 7 7 7

40.0 M17 MIR M17 MIR M17 MIR M17 MIR

Anon 3 Anon 3 Anon 3 Anon 3 -16:12:00.0

G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768 G015.0198-00.6768

24.5 18:20:23.0 21.5

Figure 5. WISE/NEWOWISE W2 images centered on M17 MIR in 14 epochs. observations of NEOWISE, M17 MIR generally shows 1997.03, 2005.09, and 2017.07. The multi-wavelength a steady rise in brightness, but with a jump and fall flux densities of M17 MIR in 1997.03, 2005.09, 2017.07 motion epoch by epoch in its light curve (see Figure 6). epoch yield a reasonable estimate for the continuum in Between 2016 and 2017, M17 MIR underwent a steep the range 3 − 8 µm, assuming blackbody emission. The rise in brightness with an amplitude of 1.7. In 2018 flux density of M17 MIR in the 3.8 µm L′ filter is the M17 MIR’s brightness dropped back to the level of 2016. continuum emission inside the transmission curve of the and was followed by a rise with a mean amplitude of 1.5 L′ filter. Note that the blackbody temperatures for the between 2018. and 2019. three different epochs are different. For the L obser- ′ vations in 1993, the three blackbody temperatures are 4.2. Combined light curve at L and W2 used for estimating the L′ flux density of M17 MIR. We The MIR flux measurements of M17 MIR span sev- find that the largest difference among the three values eral wavelengths and observation epochs. The 3.8 µm is smaller than 15%, thus the mean value is used as the ′ L filter is closest to the L filter of the 1993.03 epoch L′ filter flux density. For 1993.03, 1997.03, 2005.09, and observation (Giard et al. 1994), while the observations 2017.07 we obtain L′ flux densities; for 2004.09 there of other epochs provided flux densities in at least two is a direct L′-band photometry for M17 MIR. Figure 6 filters. For a direct comparison from 1993 to 2017, the presents the combined light curve of M17 MIR at L′ and flux density at 3.8 µm L′ filter is estimated for 1993.03, 8 Chen, Zhiwei et al.

Table 3. Flux densities of M17 MIR measured from the Spitzer observations in 2005 and 2017.

Name Spitzer 2005 2017 Aper. band (mJy) (mJy) (′′) I1 1.0 ± 0.4 2.9 ± 0.5 1.6 M 17 MIR I2 19.0 ± 2.5 52.9 ± 4.7 3.6 I1 9.7 ± 0.9 9.6 ± 0.8 2.0 3 I2 33.8 ± 3.7 33.7 ± 3.9 2.5 I1 10.7 ± 0.7 8.3 ± 0.6 2.0 5 I2 23.8 ± 1.5 20.9 ± 1.3 2.0 I1 37.8 ± 5.0 31.3 ± 3.7 2.7 7 I2 201.2 ± 17.2 193.0 ± 6.9 4.0 Note—Aperture photometry on the sources in this table are ﬁxed in positions, aperture size, and circular annulus size in the 2005 and 2017 epochs, for direct Figure 6. Top: Light curve of M17 MIR at 3.8 µm L′ band; comparison in the two epochs. Bottom: Light curve at 4.6 µm W2 band. The time span is over 26 yr from 1993 to 2019.

Table 4. Flux densities of M17 MIR at 4.5 µm measured with ISO/ISOCAM in 1997 and Spitzer in 2005 and 2017. Table 5. WISE/NEOWISE flux densities of M17 MIR and other IR sources Name 1997 2005 2017 Aper. (mJy) (mJy) (mJy) (′′) Epoch MJD M17 MIR source 7 G015.0198 a M17 MIR 32.1 ± 11.1 19.0 ± 2.5 52.9 ± 4.7 3.6 W2 [mJy] W2 [mJy] W2 [mJy] 7 202.7 ± 31.9 201.2 ± 17.2 193.0 ± 6.9 4.0 2010.03 55280.5 17.6 (4.4) 250.0 (5.1) 70.7 (2.4) Note—The aperture photometry of M17 MIR in the 2010.09 55463.5 33.7 (4.4) 257.8 (5.7) 72.4 (2.3) ISO/LW1 image might be affected by the row of bad 2014.03 56745.5 50.2 (4.1) 210.3 (6.4) 71.9 (2.3) pixels lying below M 17 MIR, thus its flux density 2014.09 56924.6 82.0 (4.1) 180.5 (6.4) 72.0 (2.4) might be underestimated. 2015.03 57107.5 62.0 (4.4) 177.3 (6.4) 69.9 (2.4) 2015.09 57284.5 81.7 (4.6) 203.3 (7.7) 69.0 (3.2) W2 band. The time span of the multi-epoch data is 26.5 2016.03 57471.5 71.5 (4.7) 180.2 (7.5) 68.9 (2.8) years. 2016.09 57645.5 82.5 (4.5) 212.8 (7.7) 66.3 (3.3) M17 MIR shows decreasing, quiescent and re- 2017.03 57838.5 95.6 (5.3) 199.4 (7.7) 69.2 (2.9) brightening phases in its MIR light curve, as shown in 2017.09 58006.5 126.0 (4.2) 193.4 (6.3) 72.2 (2.4) Figure 6. Due to the lack of 3 − 4 µm observations be- 2018.03 58202.6 72.7 (4.3) 195.0 (5.8) 67.3 (2.4) fore 1993 and during 1997 − 2004, the duration of the 2018.09 58366.5 110.5 (4.4) 189.2 (6.3) 70.0 (3.5) decreasing phase is poorly constrained. From the avail- 2019.03 58566.5 118.0 (3.9) 185.7 (5.1) 74.9 (1.6) able data, the quiescent phase between the decreasing 2019.09 58730.5 95.7 (4.2) 228.5 (7.8) 68.3 (2.4) and re-brightening phase might last about 6 years, from a Spitzer mid 2004 to mid 2010. Because M17 MIR seems to The I2 magnitude of this source is 8.55, keep the trend of re-brightening, the timescale of the corresponding to 68.05 mJy, while its re-brightening phase is & 9 yr. The continuing NEO- WISE/NEOWISE W2 mean flux density is 70.49 mJy. WISE observations will provide vital information on the re-brightening timescale. ration in June 1984 (Forster & Caswell 1989, hereafter FC89). The complete catalog of H2O masers found by Forster & Caswell (1989) was compiled by Forster & Caswell (1999). Three FC89 H2O masers are 4.3. H2O Maser Variability associated with M17 MIR, and show two velocity compo- −1 −1 The earliest H2O maser detection associated with nents of 16−18kms and 23.3kms . One H2O maser M17 MIR was achieved at the VLA in C configu- coinciding with M17 MIR was detected with the VLA Bursting massive protostar M17 MIR 9

Table 6. Water masers observations

Epoch Obs. R.A. Decl. Sep. V peak S peak Beam Ref.

(J2000) (J2000) (arcsec) (km s−1) (Jy)

(1) (2) (3) (4) (5) (6) (7) (8) (9)

1984.06 VLA-C 18:20:22.91 -16:11:48.29 1.7 23.32 1.5 4′′. 1 × 2′′. 2 FC89 1984.06 VLA-C 18:20:23.12 -16:11:48.29 1.5 18.05 6.7 4′′. 1 × 2′′. 2 FC89 1984.06 VLA-C 18:20:23.05 -16:11:48.20 0.5 16.73 6.9 4′′. 1 × 2′′. 2 FC89 1984.08 VLA-D 18:20:23.07 -16:11:49.2 17.4 18.57 5′′. 6 × 7′′. 1 MCF88 1996.09 VLA-D 18:20:22.98 -16:11:47.97 0.6 15.8 121 5′′. 0 × 3′′. 4 JDG98 1996.09 VLA-D 18:20:22.98 -16:11:47.97 0.6 27.7 17 5′′. 0 × 3′′. 4 JDG98 2003.10 ATCA-EW352 18:20:23.04 -16:11:48.4 0.6 19 197 8′′ × 29′′ B10 2003.10 ATCA-EW352 18:20:23.04 -16:11:48.4 0.6 24 60 8′′ × 29′′ B10 2004.07 ATCA-H168 18:20:23.04 -16:11:48.4 0.6 20 67 13′′ × 9′′ B10 2004.07 ATCA-H168 18:20:23.04 -16:11:48.4 0.6 24 17 13′′ × 9′′ B10 2008.08 ATCA-6B 18:20:23.02 -16:11:47.8 0.2 21 4.2 1.7 × 0.5 BE11

in D configuration and shows a flux density of 18.57Jy north to M17 MIR (see Figure 7). This feature is consis- (Massi et al. 1988, hereafter MCF88). One JDG98 H2O tent with the stronger blue-shifted component and the maser (source 2 in JDG98) with two velocity compo- weaker red-shifted component observed by the VLA and nents is detected around the same source. The veloci- ATCA. ties of the FC89 H2O masers are consistent with those of the two components of the JDG98 H2O maser source 5. DISCUSSION 2. Both the FC89 and JDG98 H2O masers are likely 5.1. The connection between the IR variability and pumped by the interaction between the outflow from H2O maser emission variability M17 MIR and the surrounding gas. Interestingly, the flux density of the JDG98 H2O maser source 2 in 1996.9 The IR variability of M17 MIR overlaps with the sur- epoch is ten times higher than the one of the FC89 H2O rounding H2O maser variability in time space. The con- maser in 1984.6. nection between them helps to clarify the physical ori- Several years after the last VLA observations, three gin of these variabilities. The flux density of the JDG98 water maser observations were taken by ATCA dur- H2O maser source 2 in 1996.09 is ten times higher than ing three epochs. A fast fading of the water maser in- the FC89 H2O maser source in 1984.6. This H2O maser tensity from 197 Jy in 2003.10 (Breen et al. 2010, here- variation from 1984 to 1996 and the decreasing phase in after B10), through 67Jy in 2004.07 (B10), to 4.2Jy in IR light curve during 1993 and 2004 suggests that the 2008.08 (Breen & Ellingsen 2011, hereafter BE11) was brightness peak likely occurred between 1984 and 1993. observed. In the earlier two epochs, the spectra of wa- Moreover, the fast fading of the H2O maser emission ter maser emission show two components with a stronger from 2003 to 2008 and the quiescent phase in IR light blue-shifted one at ∼ 19 − 20kms−1 and a weaker red- curve during 2004 and 2010 together constrain the oc- shifted one at 24kms−1, which are consistent with the currence of the quiescent phase between mid 2004 and VLA observations in the earlier epochs. In 2008.08 mid 2010, characterized by both low IR brightness and epoch, two components were observed in the spectrum – H2O maser emission. a weaker blue-shifted one at ∼ 18kms−1 and a stronger VERA observations in the earliest two epochs red-shifted one at 21kms−1. (2009.10 and 2010.01) were made in the quiescent phase, The latest water maser observations were conducted while the later observations in the epochs 2010.04, by VLBI with VERA in 2009.10 − 2010.12 consisting of 2010.09, 2010.11, and 2010.12 took place in the re- 6 epochs (CJO16). The VERA observations detected brightening phase. The proper motions of the H2O 203 water maser spots in the 6 epoch observations. The masers spots are traced in more than three epochs (priv. −1 dominant blue-shifted maser spots (Vlsr . 22kms ) communication with J. O. Chibueze). The proper mo- are mostly distributed south to M17 MIR, while few red- tions of the H2O maser spots are measured in the tran- −1 shifted maser spots (Vlsr & 22kms ) are distributed sient phase between quiescent and re-brightening phase from 2009.10 to 2010.12. The 3D motions of the H2O 10 Chen, Zhiwei et al. maser spots detected by the VERA observations show an expanding bubble originating from M17 MIR. Together with the re-brightening we suggest that both phenomena are due to the accretion/outflow activity of M17 MIR. Similar flaring of the H2O masers associated with the outflow from the high-mass protostar NGC 6334I- MM1B shows a mean factor of 6.5 before and after the burst of MM1B in the submillimeter range (Brogan et al. 2018). They attribute the enhanced accretion rate to the coincident flaring in the submillimeter continuum and the H2O maser emission of NGC6334I- MM1B. For M17 MIR, we suggest an accretion outburst to explain both the MIR variations and the correlated H2O maser variations. Accompanied with the expanding motions of H2O maser spots measured during the transient phase from quiescent to re-brightening and the close positive relation between mass loss and accretion Figure 7. The proper motions and VLSR of the CJO16 H2O maser. The background gray scale image is the VLT L′ rate, the most promising explanation for the outburst in image. MIR brightness since 2010.03 is an enhanced accretion rate. −3 −1 −1 outflow velocity. Given P˙ ∼ 4 × 10 M⊙ yr kms −1 5.2. Outflow rate and disk accretion rate of M17 MIR estimated in this work and V10 ≈ 1.9kms determined by CJO16, the outflow mass rate M˙ out is ∼ The proper motions and radial velocities of the CJO16 −4 −1 ′ 2 × 10 M⊙ yr . Assuming a ratio of outflow mass H2O masers are overlaid on the VLT L image, as shown rate to disk accretion rate of 30% (Beltrán & de Wit in Figure 7. The 3D motions of the H2O maser spots 2016), the disk accretion rate onto M17 MIR is M˙ acc ∼ can be used to estimate the momentum rate of the gas −4 −1 7 × 10 M⊙ yr . flow, given the assumption that all the momentum of the This disk accretion rate is estimated from the H2O ejected gas is transferred to the surrounding molecular maser observations during 2009 to 2010, when M17 MIR environment. The momentum rate P˙ is then estimated just started to re-brighten. The MIR emission between from the relation (Goddi et al. 2011): 2010.09 and 2010.03 varies by a factor of about 2 in −3 2 2 −1 −1 P˙ =1.5 × 10 V10 R100 (Ω/4π) n8 M⊙ yr kms , the WISE W2 band. Johnstone et al. (2013) simulate the MIR flux variation induced by an enhanced disk where V10 is the mean maser velocity in units of accretion rate and find that the MIR flux variation is −1 10kms , R100 is the average distance of H2O masers to proportional to the disk accretion rate. The W2 flux the driving source in units of 100 AU, Ω is the solid angle between 2010 and 2019 varies by an average factor of of the gas flow, and n8 is the gas volume density in units 3. We speculate that the disk accretion rate in 2019 8 −3 of 10 cm . R100 is calculated as 0.85 by averaging the is likely 3 times the rate estimated above, reaching to −3 −1 distances of the H2O masers to the dynamical center, ∼ 2 × 10 M⊙ yr . which is determined at the position of R.A.(J2000) = 18h20m23s.017, Decl.(J2000) = −16◦11′47′′. 98 (CJO16). 5.3. Modelling of the SED in quiescent and The gas density is about 108 cm−3 according to the dis- re-brightening phase cussion in Sect. 3.1. Ω is about 4π for an expanding bub- During the quiescent phase from mid 2004 to mid ble. The above physical properties from observations 2010 (hereafter 2005 epoch), we collect IR flux densities −3 −1 −1 yield a momentum rate P˙ ∼ 4 × 10 M⊙ yr kms at six wavelengths to construct the SED of M17 MIR. for the protostellar outflow of M17 MIR. This is compa- During the re-brightening phase, the flux densities ob- −3 0 −1 −1 rable to outflows of 10 − 10 M⊙ yr kms driven tained in 2017.07 (Spitzer I1 and I2 band) and 2017.08 by MYSOs (Moscadelli et al. 2016), and one magni- (SOFIA/FORCAST 19 µm and 37 µm band) are sep- tude higher than the outflow from a low-mass YSO in arated by 14 days. The NEOWISE W2 flux density IRAS20231+3440 (Ogbodo et al. 2017). was obtained at the end of September 2017, nearly two The protostellar outflow provides an indirect measure months later than the Spitzer and SOFIA observations. of the disk accretion rate. The outflow mass rate is de- We collect only the Spitzer and SOFIA observations in rived by dividing the outflow momentum rate P˙ with the 2017.07/08 to construct the SED of M17 MIR during Bursting massive protostar M17 MIR 11

0 0 )−--s-i )−--s-i 0 2010.03 0 2010.03 Herschel Herschel

0 0 2017.08 2017.08 SOFIA 2 2 SOFIA

0 0 2017.08

2005.07 2017.08

SOFIA Spitzer 2006.04 SOFIA

2017.07 2006.04 0 VLT 0 Spitzer VLT

2004.09 10 VLT 0

1 10 100 1000 1 10 100 1000 Wavelength [μm] Wavelengthμ

2 2 Figure 8. SEDs of the R17 sp–s-i models that satisfy χ − χbest < 3ndata for M 17 MIR. the outburst phase (hereafter 2017 epoch). The flux To sustain such high accretion rate, we require this cri- densities of M17 MIR in the 2005 and 2017 epochs are terion. tabulated in Table 7. The 3σ upper limit at Herschel Disk inclination angle >= 30◦. Because the blue- and 70µm image is 77 Jy for M17 MIR, which is consid- red-shifted H2O maser spots are distinctly separated ered as the upper limit in both 2005 and 2017 epochs. in the vicinity of M17 MIR, the disk inclination angle The additional input parameters besides the flux den- should not be small. sities are distance and foreground extinction AV . The Exclude unphysical models. Some models are unphysi- distance to M17 MIR is fixed to 2.0kpc. We assume a cal with too small radii but too high temperatures. We 2 4 wide range of 10–200mag for the foreground extinction compute the Stefan–Boltzmann luminosity (4πR∗σT∗ ) AV . from radius and temperature of each model. These un- We employed the SED model grid for YSOs that spans physical models are below the ZAMS track in the HRD. a wide range of evolutionary stages, from the youngest We keep only the models that are lying on/above the deeply embedded protostars to pre-main-sequence stars ZAMS track in the HRD. with debris or no disk, and which covers a wide and uni- form region of parameter space (Robitaille 2017a, here- The above criteria to the four model sets return rea- after R17). We fit the SED of M17 MIR using several sonable good fits for the SEDs of M17 MIR in the R17 model sets of different model components, using the 2005 and 2017 epoch. Because of the degeneracy of python-based fitting tool (1.3 version; Robitaille 2017b). the four model sets, one has to score the four model Based on the prior knowledge of M17 MIR, M17 MIR sets not only based on χ2 values. In order to assess is a highly embedded massive protostar with an accre- the model sets in comparison to each other in a sta- tion disk. The model sets we used are sp-s-i,sp-smi,spu- tistically robust way, Robitaille (2017a) suggested cal- smi,spubsmi. All the four model sets have a passive disk culating P (D|M) ∝ Ngood/N, where Ngood is the total with an inner radius equal to the dust sublimation ra- number of models from a given model set, and N is the dius. sp-s-i is the simplest model set including only the total number of models in that set. Towner et al. (2019) central star and passive disk. The other three model sets used the R17 model sets to find the best YSO models include ambient gas (sp-smi), envelope+ambient gas for 12 MYSOs based on the overall assessment of χ2 (spu-smi), and envelope+ambient+cavity (spubsmi). and P (D|M) scores. For the good fits of M17 MIR, we 2 Fitting the SED of M17 MIR to the individual mod- showed the χbest and P (D|M) scores of the four model els of the four model sets returns a χ2 value for every sets in Table 8. In the 2005 and 2017 epoch, sp-s-i is 2 2 individual model. The criterion χ − χbest < 3ndata the ”best representation” model set according to both 2 can generally split good and bad fits. We used this the χbest and P (D|M) scores. criterion to select good fits from the four model sets. Among the sp-s-i model set, 44 and 13 models remain Among these good fits of the four model sets, we use for the SEDs of M17 MIR in the 2005 and 2017 epoch, the following criteria to pick reasonable models. respectively. In the 2017 epoch, the SOFIA observa- Disk mass >=0.01 M⊙. The disk accretion of M17 MIR tions at 19.7µm and 37.1µm are very useful in distin- −4 −1 estimated from H2O maser motions is 7×10 M⊙ yr . guishing the various models of a given model set (Fig- 12 Chen, Zhiwei et al.

Table 7. Flux densities of M17 MIR in the 2005 and 2017 epoch

Epoch Spitzer VLT Spitzer Spitzer Spitzer VLTa SOFIAb SOFIAb Herschelc State 3.6µm 3.78µm 4.5µm 5.71µm 7.8µm 11.8µm 19.7µm 37.1µm 70µm [mJy] [mJy] [mJy] [Jy] [Jy] [Jy] [Jy] [Jy] 2005 - 2.29 ± 0.21 21.3 ± 2.9 85.0 ± 8.0 0.14 ± 0.03 0.139 ± 0.007 < 0.51 ± 0.12 < 7.3 ± 1.2 < 77 2017 3.8 ± 0.7 - 59.7 ± 5.9- - > 0.139 ± 0.007 0.51 ± 0.12 7.3 ± 1.2 < 77 aThe ﬂux density at 11.8µm in the 2005 epoch is considered as the lower limit in the 2017 epoch. bThe ﬂux densities at 19.7µm, 37.1µm in the 2017 epoch are considered as the upper limits in the 2005 epoch. cThe measurement at 70µm in the 2010.03 epoch is considered as the upper limit.

Table 8. χ2 and P (D|M) scores of R17 model sets

2005 epoch 2017 epoch Model χ2 P (D|M) Model χ2 P (D|M) sp-s-i 9.034 0.0044 sp-s-i 2.41 0.0013 sp-smi 8.241 0.0022 sp-smi 15.99 0.0017 spu-smi 14.54 0.0012 spu-smi 14.46 0.0009 spubsmi 9.42 0.00065 spubsmi 11.14 0.0005

2 ure 8). Among the 13 models in the 2017 epoch (Ta- have large circumstellar disks. Moreover, the χbest sp- ble 9), two models correctly reproduce the SED slope -s-i model in the 2017 epoch also suggests a large cir- 2 disk between 19.7µm and 37.1µm. Indeed, the χbest model is cumstellar disk for M17 MIR. The Rmax > 1000 AU different from the second fit only in the disk inclination family of the sp-s-i models is preferred for the SED angle. Aided by the observations at 19.7µm and 37.1µm of M17 MIR in the 2005 epoch. Among the 9 sp-s-i 2 disk in the 2017 epoch, the χbest model is distinctly better models with Rmax > 1000 AU in Table 9, two mod- than the other sp-s-i models. els with Teff = 24540 K likely produce significant 22 In the 2005 epoch, the 44 sp-s-i models show a de- GHz continuum emission that could be detected by generacy in the long-wavelength SED beyond 12µm, the ATCA observations during 2003–2008 (B10,BE11). due to the lack of observations at long wavelengths. However, no 22 GHz continuum source was reported Povich et al. (2011) found similar SED degeneracy at around M17 MIR according to the ATCA observations. long wavelengths for a deeply embedded YSO with hard We excluded the two very hot sp-s-i models in the fol- X-ray emission (PCYC source 699 from Povich et al. lowing analysis. On the other hand, stellar luminosity (2011)) in the Carina nebula region, when they fit- in the 2005 epoch should be much lower than that in ted the SED of this object to YSO model SEDs from the 2017 epoch. This constraint leads to a final collec- Robitaille et al. (2006). In Figure 9 of Povich et al. tion of only 3 reasonable sp-s-i models. For the 3 sp-s-i (2011), the distributions of the well-fit YSO models show models, we calculated the mean hTeff i, hR∗i,hL∗i, and 1 two distinct groups: intermediate-mass YSOs with rela- foreground ground hAV i, weighted by χ2 . These values tively high circumstellar extinction and low-mass YSOs are listed in Table 10 for the 2005 epoch. with lower circumstellar extinction. The high circum- In the above analysis for the R17 models satisfying 2 2 stellar extinction group is consistent with the high ab- χ − χbest < 3ndata, no constraint on the foreground ex- sorption of the X-ray source (Povich et al. 2011). For tinction AV was applied. The good SED models in the the 44 sp-s-i models of M17 MIR in the 2005 epoch, 2005 and 2017 epoch converge on a narrow foreground two distinct families are found in the plot of infrared extinction AV range 120 − 150 mag, despite a broad AV luminosity integrated between 12 − 500µm versus disk range 10 − 200 mag used in SED fitting procedure. This disk maximum radius Rmax: high infrared luminosity with narrow AV range still holds for the other ’bad’ SED disk Rmax > 1000 AU and lower infrared luminosity with models. The observations at 7.8µm and 11.8µm in the disk disk Rmax < 1000AU. The Rmax > 1000 AU family is more 2005 epoch that cover the broad absorption feature at consistent with the general understanding of MYSOs 9.7µm yield rather strong constrain on the foreground which are luminous at long infrared wavelengths and extinction. This is consistent with the deep absorption Bursting massive protostar M17 MIR 13 feature at 9.7 µm caused by the large absorption of sili- value estimated from observations. This fact is consis- cate grains, which is commonly observed for deeply em- tent with the simulation prediction that large accretion bedded high-mass protostars (e.g., Morales et al. 2009; rate may cause the protostellar photosphere to bloat, Pitts et al. 2018). resulting in larger radius (Yorke & Bodenheimer 2008; Hosokawa & Omukai 2009; Meyer et al. 2019a).

The difference of luminosity between the 2005 and 5.4. Accretion history of M17 MIR 2017 epoch indicates the luminosity burst of ∆L = The amplitude in MIR flux variation is directly pro- 7600±900L⊙ during an accretion outburst of M17 MIR. portional to the enhanced L . Nevertheless, the en- If ∆L is entirely attributed to increased accretion, we acc hanced L is a reduced version of enhanced M˙ acc, due can predict the flux density of M17 MIR at Spitzer I1 acc to the significant inflation of the central object during and I2 bands during the outburst. We assume that burst accretion phase. In the burst accretion phase, the blackbody temperature of ∆L during outburst is the M˙ acc is amplified by a factor of tens compared 1000 K. Reddenning the blackbody radiation from ∆L with that in quiescent phase. M17 MIR even shows a with a certain amount of extinction essentially matches WISE/NEOWISE W2 flux variation in a cadence of six the flux density at Spitzer I1 and I2 bands. We assume months, despite the overall trend of becoming continu- a total extinction to M17 MIR as 150 mag, which is the ously brighter. The overall trend is suggested to origi- sum of a foreground extinction about 130mag accord- nate from the increasing M˙ acc. On the other hand, the ing to the good SED models for M17 MIR and an addi- smaller variation along the cadence of six months might tional extinction of 20 mag from the circumstellar disk. imply unsteady M˙ acc with a smaller amplitude. For in- The blackbody radiation from ∆L = 7600 ± 900 L⊙ suf- stance, the variation in the W2 flux density is −42% be- fers from a total extinction of 150mag, equivalent to tween 2017.09 and 2018.03. This variation corresponds 0.066 × 150mag and 0.052 × 150mag at Spitzer I1 and to ∆M˙ acc = −42%, if we assume the stellar radius to I2 bands. This toy model yields flux density of 5.8 mJy keep unchanged. This is the largest variation in M˙ acc and 52.5 mJy at Spitzer I1 and I2 bands, respectively, during the burst accretion phase. which are comparable to the observed values 3.8 ± 0.7 Considering the MIR variation during the decreas- mJy and 59.7 ± 5.9 mJy in the 2017 epoch with a factor ing phase 1993.03–2004.09, M17 MIR is about 3 times of less than 2. brighter in 1993.03 than in 2017.08. We speculate that The luminosities of M17 MIR in Table 10 are indeed −3 −1 M˙ acc in 1993.03 is ∼ 5 × 10 M⊙ yr , 3 times that in the total luminosites as the combination of photospheric the re-brightening phase. However, M˙ acc in 1993.03 is luminosity (L∗) and accretion luminosity (Lacc), probably even higher, since M17 MIR may expand to 2 4 L =4πR∗σTeff + GM∗M˙ acc/R∗. even larger size than that in the re-brightening phase. The MIR light curve of M17 MIR exhibits three phases In the 2005 epoch, we assume that half of L2005 = characterized by M17 MIR’s MIR flux density variation 1400 L⊙ arises from a ZAMS photosphere with solar over 26yr. From the above results and analyses, we con- metalicity, the progenitor’s properties are: mass M∗ = clude that the decreasing and re-brightening phase of 5.4 M⊙, radius R∗ = 2.7 R⊙, and effective temperature M17 MIR indeed reflect the two accretion bursts, which Teff = 18000K (Tout et al. 1996), which lies between are separated by the quiescent phase lasting about 6 spectral type B2.5V and B3V (Pecaut & Mamajek yr. The first accretion burst likely launched in a time 2013). The accretion rate M˙ acc in the 2005 epoch needed between 1984 and 1993, and lasted longer than 11 yr −5 −1 to produce another half of L2005 is ∼ 1.1×10 M⊙ yr , but less than 20 yr. Thus we use a middle value of comparable to the steady accretion rates for the burst- 15 yr. We fortunately catched the initial stage of the sec- ing MYSO of Hunter et al. (2021) and for the 6 M⊙ ond accretion burst, aided by the multi-epoch observa- formation model of Haemmerléet al. (2013). In the tions of Spitzer and WISE/NEOWISE space telescopes. 2017 epoch, we assume Lacc ≈ 8300 L⊙, which yields The duration of the second accretion burst is definitely −3 −1 M˙ acc ≈ 1.7 × 10 M⊙ yr in the 2017 epoch, where longer than 9 yr. It is very likely that M17 MIR has yet M∗ =5.4 M⊙ and R∗ = 34.1 R⊙ are assumed. The value not reached to its peak brightness, and the duration of of M˙ acc in the 2017 epoch estimated from the stellar pa- the second accretion burst is similar to that of the first −3 −1 rameters agrees with the value of ∼ 2 × 10 M⊙ yr epoch. M17 MIR is the first MYSO showing multiple obtained from the proper motions of H2O maser spots accretion bursts with durations of tens of yr and burst (Sect. 5.2). In the 2017 epoch, a larger stellar radius is magnitude of about 2mag. The observed properties of required to produce an accretion rate comparable to the the accretion bursts of M17 MIR agree with the theoret- 14 Chen, Zhiwei et al.

2 2 Table 9. The parameters of R17 sp-s-i models that satisfy χ − χbest < 3ndata for M 17 MIR

2 disk Epoch model name χ AV R∗ Teﬀ Mdisk Rmax inclination Comments ◦ (mag) (R⊙) (K) (M⊙) (AU) ( )

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

7Px1sJu9 07 11.387 137.259 8.678 11330.0 0.03119 2707.0 60.37 good fit ORya6h0F 07 13.715 138.992 18.77 12710.0 0.01624 1362.0 68.83 L2005 ≈ L2017 e6RRFGdF 04 15.246 143.522 8.979 17560.0 0.03541 1359.0 38.46 L2005 ≈ L2017 1EHMCdxi 07 15.652 135.411 17.09 13690.0 0.07064 4221.0 63.84 L2005 ≈ L2017 2005 e6RRFGdF 05 16.029 143.569 8.979 17560.0 0.03541 1359.0 41.11 L2005 ≈ L2017 wRz4n9OE 04 16.72 120.122 7.19 10480.0 0.02166 3444.0 37.77 good fit LwIekVA6 04 19.687 147.335 16.18 10510.0 0.03299 2554.0 31.59 good fit kskLMN4Y 05 20.179 149.157 5.494 24540.0 0.0622 3640.0 47.86 too hot kskLMN4Y 06 20.859 148.937 5.494 24540.0 0.0622 3640.0 51.65 too hot 8UXTjAhl 08 2.412 134.181 34.14 9631.0 0.07883 4067.0 72.06 good fit 8UXTjAhl 07 7.531 134.866 34.14 9631.0 0.07883 4067.0 63.34 good fit yTvNrh2o 04 10.444 117.76 4.789 22810.0 0.07566 316.5 32.15 bad fit at 19.7, 37.1µm REwCcjZc 05 10.863 132.046 39.91 8132.0 0.017 4539.0 48.53 bad fit at 19.7, 37.1µm REwCcjZc 04 11.187 134.554 39.91 8132.0 0.017 4539.0 32.8 bad fit at 19.7, 37.1µm yTvNrh2o 05 11.204 119.925 4.789 22810.0 0.07566 316.5 42.77 bad fit at 19.7, 37.1µm 2017 REwCcjZc 06 11.619 130.341 39.91 8132.0 0.017 4539.0 56.39 bad fit at 19.7, 37.1µm REwCcjZc 07 12.488 129.402 39.91 8132.0 0.017 4539.0 60.18 bad fit at 19.7, 37.1µm V5ttV0Uz 07 12.517 145.476 59.43 7264.0 0.0982 4549.0 67.08 bad fit at 19.7, 37.1µm h2UvYxug 09 13.188 139.651 76.3 9036.0 0.01695 1770.0 82.7 bad fit at 19.7, 37.1µm LZEBw0DZ 08 13.556 138.967 38.38 11690.0 0.01377 941.7 71.79 bad fit at 19.7, 37.1µm LZEBw0DZ 07 14.313 139.677 38.38 11690.0 0.01377 941.7 60.78 bad fit at 19.7, 37.1µm jzCvPxfT 06 14.323 126.138 47.92 8185.0 0.0576 4076.0 52.93 bad fit at 19.7, 37.1µm

Table 10. Basic stellar parameters of M17 MIR in the 2005 and 2017 epochs.

Epoch R∗ Teﬀ L∗ AV M∗ M˙ acc Lacc −1 (R⊙) (K) (L⊙) (mag) (M⊙) (M⊙ yr ) (L⊙) 2005 10.1(3.6) 10863(415) 1400(897) 134.7(10.4) 5.4 1.1 × 10−5 7.0 × 102 2017 34.1 9631 9034 134 5.4 1.7 × 10−3 8.3 × 103 ical simulation prediction that 2 mag bursts have mean (Caratti o Garatti et al. 2017) and two orders higher duration of about 20yr (Meyer et al. 2019b). than that onto G358.93-MM1 (Stecklum et al. 2021). We assume a bursting accretion timescale of 15yr The quiescent phase of M17 MIR lasts about 6yr. Dur- which is comparable to the duration of the decreasing ing the time span of 26 yr, the fraction time in accre- phase in the MIR light curve (somewhere between 1984 tion burst is about 83%. The main accretion phase of and 1993 to mid 2004). If M17 MIR can sustain this a high-mass protostellar object in a HMC is a few 104 −3 −1 enhanced disk accretion rate of ∼ 2 × 10 M⊙ yr , it yr (Furuya et al. 2005; Bonfand et al. 2017). A crude 4 could gain about 0.03 M⊙ during one accretion burst. estimate, assuming tacc = 1 × 10 yr, this fraction time This amount of accreted mass is an order of magni- of accretion burst yields ∼ 14 M⊙ mass that M17 MIR tude higher than the mass accretion onto S255IR-NIRS3 could accrete from its circumstellar disk. Bursting massive protostar M17 MIR 15

Table 11. Key parameters of six MYSOs with accretion burst

pre burst post Name d L L L ∆FIR ∆Fmm ∆t Maser Flaring Note 3 3 3 (kpc) (10 L⊙) (10 L⊙) (10 L⊙) (yr) V723 Car 2.5 - 4 ∼ 10 – 5 – K & 12.9

S255IR-NIRS3 1.8 29 160 – & 10 2 2 CH3OH, H2O K & 9

NGC6334I-MM1 1.3 2.9 47.6 – 16.3 3.9 ± 0.6 > 6 CH3OH, H2O K-band invisible

G323.46–0.08 4.8 – – – – – – CH3OH K-band visible

G358.93-MM1 6.75 7.6 19.3 12.7 & 2 – . 0.5 CH3OH K & 15

M17 MIR 2.0 1.4 9.0 1.4 & 10 – ∼ 2 × 15 H2O K & 22

5.5. Comparisons to other variable MYSOs its multiple accretion bursts suggest that minor accre- The MYSO accretion burst sample is still small. Six tion bursts are frequent at very early stages of massive MYSOs show a luminosity increase due to an accretion star formation. burst, including V723 Car (Tapia et al. 2015), S255IR- NIRS3 (Caratti o Garatti et al. 2017; Liu et al. 2018), 6. CONCLUSIONS NGC6334I-MM1 (Hunter et al. 2017, 2021), G323.46– The multi-wavelength and multi-epoch data of the 0.08 (Proven-Adzri et al. 2019), G358.93–0.03 MM1 massive protostar M17 MIR and the subsequent anal- (Stecklum et al. 2021), and M17 MIR in the present yses on these data yield the following major results: work. However, these MYSOs are in different PMS 1. By analyzing the IR data of M17 SW in the evolutionary stages. The accretion burst durations are 3 − 70 µm range, M17 MIR is identified as the driv- poorly constrained except for M17 MIR. To qualita- ing source of the expanding structure traced by H2O tively describe the relation between the PMS evolution- maser motions. The sub-mm dust continuum emission ary stage and an accretion burst, their key parameters and molecular line emission indicates that M17 MIR is are summarized in Table 11. Relatively evolved MYSOs embedded within a HMC. are commonly brighter in the near-IR K-band, while 2. The unified flux density of M17 MIR at the 3.8 µm the earliest MYSOs are very faint or even invisible in L′ band and the 4.6 µm WISE W2 band clearly shows the same band. Generally the K-band brightness of three stages in the light curve: i) the decreasing phase these MYSOs is representative for the PMS evolution- during 1993.03 to mid 2004 with an amplitude of 10 ary stage. Among the six MYSOs tabulated in Table 11, at 3.8 µm, ii) the quiescent phase during mid 2004 to four sources show variabilities in the K-band, indicat- mid 2010, and iii) the re-brightening phase since mid ing relative later PMS stages. Particularly, G323.46– 2010 until now with an amplitude of 5 at 4.6 µm. The 0.08 was the first one which showed periodic 6.7 GHz 22 GHz H2O maser emission associated with M17 MIR methanol maser emission with a period of about 90 also shows flux variations. They are correlated with the days (Proven-Adzri et al. 2019). Later – on the basis of MIR flux variation in time space, which together indi- multi-epoch IR data – it was confirmed as an accretion cate enhanced disk accretion rate onto M17 MIR during burst MYSO (see Stecklum et al. 2021, and references the decreasing and re-brightening phase. therein). 3. The kinematics of H2O maser spots, measured in For V723 Car, S255IR-NIRS3, and G358.93-MM1 2009.10–2010.12, in the close vicinity of M17 MIR yield −4 −1 the accretion burst duration ∆t was constrained from a disk accretion rate M˙ acc ∼ 7 × 10 M⊙ yr . In the their IR luminosity changes. The durations of their lu- later stage of the re-brightening phase after 2014, M˙ acc is −3 −1 minosity increase are on the order of years and thus ∼ 2 × 10 M⊙ yr , given by the 4.6 µm flux increased comparable to each other. The two earliest MYSOs, by a factor of ∼ 3 compared to the initial stage in 2010. NGC6334I-MM1 and M17 MIR, are both embedded in 4. Radiative transfer modeling for the SEDs in the HMC. The progenitor of NGC6334I-MM1 in its pre- 2005 and 2017 epoch constrains the basic stellar param- burst phase might be able to produce an HCHII region eters of M17 MIR. In the quiescent phase in the 2005 MM1B (Brogan et al. 2016), implying that NGC6334I- epoch, M17 MIR likely has total luminosity L2005 = MM1 is more massive and at later evolutionary stage 1400 ± 897 L⊙, stellar mass M∗ ∼ 5.4 M⊙, stellar ra- than M17 MIR. Acoording to the skematic evolution- dius R∗ = 10.1 ± 3.6 R⊙, and accretion rate M˙ acc ∼ −5 −1 ary sequence of massive protostar (Motte et al. 2018), 1.1 × 10 M⊙ yr . In the accretion burst phase (e.g. M17 MIR might be the earliest MYSO with accretion in the 2017 epoch), L2017 = 9034 L⊙( ∆L ≈ 7600 L⊙), −3 −1 outbursts. Considering the extreme youth of M17 MIR, R∗ = 34.1 R⊙, and M˙ acc ∼ 1.7 × 10 M⊙ yr . In the 16 Chen, Zhiwei et al. accretion outburst, a larger stellar radius is required to produce M˙ acc consistent with the value estimated from The authors like to thank the anonymous referee for a the kinematics of H2O maser spots. detailed and thoughtful review that has improved the 5. The decreasing and re-brightening phase of scientific contents of the paper. This work is supported M17 MIR reflect two accretion bursts (∆t ∼ 9 − 20 yr), by National Key Research & Development Program of which are separated by the quiescent phase lasting about China (2017YFA0402702). We acknowledge support 6 yr. During the time span of 26 yr, the fraction time from the general grants 11903083, U2031202, 11873093, of accretion burst is about 83%. M17 MIR is the first 11873094, and 12041305 of National Natural Science discovered massive protostar showing multiple accretion Foundation of China, and support from the DFG grant bursts with durations of tens of years and a burst mag- CH 71/33-1. We thank Tie Liu for informing us the nitude of about 2 mag. Among the limited sample of JCMT SCUBA2 observations in 2016 April that used in accretion burst MYSOs, the extreme youth of M17 MIR this research. We thank J. O. Chibueze for the discus- suggest that minor accretion bursts are frequent at very sion of VERA observations. This research has made us early stages of massive star formation. of the observations performed with the Herschel Space 6. Even during the current accretion burst (since mid Observatory, which is an ESA space observatory with 2010 to until now), the W2 flux density of M17 MIR science instruments provided by European-led Principal shows variations of small amplitude up to a factor of 3, Investigator consortia and with important participation indicating variable accretion rate with amplitutde much from NASA. This research has made use of the NASA/ lower than the overall contrast between queiscent and IPAC Infrared Science Archive, which is operated by the outburst state. Long time monitoring at MIR bands will Jet Propulsion Laboratory, California Institute of Tech- record the bursting accretion activity of M17 MIR in the nology, under contract with the National Aeronautics near future. Because M17 MIR is still in its earliest stage and Space Administration. This work is based in part of evolution, it will be a unique testbed for probing the on observations made with the Spitzer Space Telescope, earliest phases of massive star formation. which is operated by the Jet Propulsion Laboratory, Cal- ifornia Institute of Technology under a contract with NASA. This publication makes use of data products from the Near-Earth Object Wide-field Infrared Survey Explorer (NEOWISE), which is a joint project of the Jet Propulsion Laboratory/California Institute of Technol- ogy and the University of Arizona. NEOWISE is funded by the National Aeronautics and Space Administration. This research is based in part on observations made with the NASA/DLR Stratospheric Observatory for In- frared Astronomy (SOFIA). SOFIA is jointly operated by the Universities Space Research Association, Inc. (USRA), under NASA contract NNA17BF53C, and the Deutsches SOFIA Institut (DSI) under DLR contract 50 OK 0901 to the University of Stuttgart. This research used the facilities of the Canadian Astronomy Data Centre operated by the National Research Council of Canada with the support of the Canadian Space Agency. This research has made use of the SVO Filter Profile Service (http://svo2.cab.inta-csic.es/theory/fps/) supported from the Spanish MINECO through grant AyA2014-55216. This research has made use of NASA’s Astrophysics Data System.

REFERENCES

Audard, M., Abrah´am,´ P., Dunham, M. M., et al. 2014, in Beltr´an, M. T., & de Wit, W. J. 2016, A&A Rv, 24, 6, Protostars and Planets VI, ed. H. Beuther, R. Klessen, doi: 10.1007/s00159-015-0089-z C. Dullemond, & T. Henning (University of Arizona Press), 387–410 Bursting massive protostar M17 MIR 17

Benjamin, R. A., Churchwell, E., Babler, B. L., et al. 2003, Fruscione, A., McDowell, J. C., Allen, G. E., et al. 2006, in PASP, 115, 953, doi: 10.1086/376696 Proc. SPIE, Vol. 6270, Society of Photo-Optical Bonfand, M., Belloche, A., Menten, K. M., Garrod, R. T., Instrumentation Engineers (SPIE) Conference Series, & Müller, H. S. P. 2017, A&A, 604, A60, 62701V doi: 10.1051/0004-6361/201730648 Furuya, R. S., Cesaroni, R., Takahashi, S., et al. 2005, ApJ, Breen, S. L., Caswell, J. L., Ellingsen, S. P., & Phillips, 624, 827, doi: 10.1086/429286 C. J. 2010, MNRAS, 406, 1487, Getman, K. V., Feigelson, E. D., Broos, P. S., Townsley, doi: 10.1111/j.1365-2966.2010.16791.x L. K., & Garmire, G. P. 2010, ApJ, 708, 1760, Breen, S. L., & Ellingsen, S. P. 2011, MNRAS, 416, 178, doi: 10.1088/0004-637X/708/2/1760 doi: 10.1111/j.1365-2966.2011.19020.x Getman, K. V., Feigelson, E. D., Garmire, G., Broos, P., & Wang, J. 2007, ApJ, 654, 316, doi: 10.1086/509112 Brogan, C. L., Hunter, T. R., Cyganowski, C. J., et al. Giard, M., Bernard, J. P., Lacombe, F., Normand, P., & 2016, ApJ, 832, 187, doi: 10.3847/0004-637X/832/2/187 Rouan, D. 1994, A&A, 291, 239 —. 2018, ApJ, 866, 87, doi: 10.3847/1538-4357/aae151 Goddi, C., Moscadelli, L., & Sanna, A. 2011, A&A, 535, Broos, P. S., Getman, K. V., Povich, M. S., et al. 2013, L8, doi: 10.1051/0004-6361/201117854 ApJS, 209, 32, doi: 10.1088/0067-0049/209/2/32 Guarcello, M. G., Caramazza, M., Micela, G., et al. 2012, Caratti o Garatti, A., Stecklum, B., Garcia Lopez, R., et al. ApJ, 753, 117, doi: 10.1088/0004-637X/753/2/117 2017, Nature Physics, 13, 276, doi: 10.1038/nphys3942 Güver, T., & Ozel,¨ F. 2009, MNRAS, 400, 2050, Carrasco-González, C., Torrelles, J. M., Cantó, J., et al. doi: 10.1111/j.1365-2966.2009.15598.x 2015, Science, 348, 114, doi: 10.1126/science.aaa7216 Haemmerlé, L., Eggenberger, P., Meynet, G., Maeder, A., Cesaroni, R. 2005, in IAU Symposium, Vol. 227, Massive & Charbonnel, C. 2013, A&A, 557, A112, Star Birth: A Crossroads of Astrophysics, ed. doi: 10.1051/0004-6361/201321359 R. Cesaroni, M. Felli, E. Churchwell, & M. Walmsley —. 2016, A&A, 585, A65, (Cambridge: Cambridge University Press), 59–69 doi: 10.1051/0004-6361/201527202 Chen, Z., Nürnberger, D. E. A., Chini, R., Jiang, Z., & Hanson, M. M., Howarth, I. D., & Conti, P. S. 1997, ApJ, Fang, M. 2015, A&A, 578, A82, 489, 698, doi: 10.1086/304808 doi: 10.1051/0004-6361/201424895 Herter, T. L., Vacca, W. D., Adams, J. D., et al. 2013, Chibueze, J. O., Kamezaki, T., Omodaka, T., et al. 2016, PASP, 125, 1393, doi: 10.1086/674144 MNRAS, 460, 1839, doi: 10.1093/mnras/stw1019 Hobson, M. P., Padman, R., Scott, P. F., Prestage, R. M., Chini, R., Hoffmeister, V., Kimeswenger, S., et al. 2004, & Ward-Thompson, D. 1993, MNRAS, 264, 1025, Nature, 429, 155, doi: 10.1038/nature02507 doi: 10.1093/mnras/264.4.1025 Chini, R., Hoffmeister, V. H., Nielbock, M., et al. 2006, Hoffmeister, V. H., Chini, R., Scheyda, C. M., et al. 2006, ApJL, 645, L61, doi: 10.1086/505862 A&A, 457, L29, doi: 10.1051/0004-6361:20065946 Chini, R., Nielbock, M., & Beck, R. 2000, A&A, 357, L33 —. 2008, ApJ, 686, 310, doi: 10.1086/591070 Hosokawa, T., & Omukai, K. 2009, ApJ, 691, 823, Churchwell, E., Babler, B. L., Meade, M. R., et al. 2009, doi: 10.1088/0004-637X/691/1/823 PASP, 121, 213, doi: 10.1086/597811 Hosokawa, T., Yorke, H. W., & Omukai, K. 2010, ApJ, 721, Crété, E., Giard, M., Joblin, C., et al. 1999, A&A, 352, 277 478, doi: 10.1088/0004-637X/721/1/478 Cutri, R. M., & et al. 2014, VizieR Online Data Catalog, Hunter, T. R., Brogan, C. L., MacLeod, G., et al. 2017, II/328 ApJL, 837, L29, doi: 10.3847/2041-8213/aa5d0e De Buizer, J. M., Liu, M., Tan, J. C., et al. 2017, ApJ, 843, Hunter, T. R., Brogan, C. L., De Buizer, J. M., et al. 2021, 33, doi: 10.3847/1538-4357/aa74c8 ApJL, 912, L17, doi: 10.3847/2041-8213/abf6d9 Feigelson, E. D., Broos, P., Gaffney, III, J. A., et al. 2002, Johnson, C. O., Depree, C. G., & Goss, W. M. 1998, ApJ, ApJ, 574, 258, doi: 10.1086/340936 500, 302 (JDG98), doi: 10.1086/305717 Feigelson, E. D., Townsley, L. K., Broos, P. S., et al. 2013, Johnstone, D., Hendricks, B., Herczeg, G. J., & Bruderer, S. ApJS, 209, 26, doi: 10.1088/0067-0049/209/2/26 2013, ApJ, 765, 133, doi: 10.1088/0004-637X/765/2/133 Forster, J. R., & Caswell, J. L. 1989, A&A, 213, 339 Kamezaki, T., Imura, K., Omodaka, T., et al. 2014, ApJS, —. 1999, A&AS, 137, 43, doi: 10.1051/aas:1999479 211, 18, doi: 10.1088/0067-0049/211/2/18 Freudling, W., Romaniello, M., Bramich, D. M., et al. 2013, Kassis, M., Deutsch, L. K., Campbell, M. F., et al. 2002, A&A, 559, A96, doi: 10.1051/0004-6361/201322494 AJ, 124, 1636, doi: 10.1086/341819 18 Chen, Zhiwei et al.

Kessler, M. F., Steinz, J. A., Anderegg, M. E., et al. 1996, Ram´ırez-Tannus, M. C., Kaper, L., de Koter, A., et al. A&A, 500, 493 2017, A&A, 604, A78, doi: 10.1051/0004-6361/201629503 Lagage, P. O., Pel, J. W., Authier, M., et al. 2004, The Robitaille, T. P. 2017a, A&A, 600, A11 (R17), Messenger, 117, 12 doi: 10.1051/0004-6361/201425486 Lim, W., De Buizer, J. M., & Radomski, J. T. 2020, ApJ, —. 2017b, sedfitter: a Python SED fitting tool - v1.0 888, 98, doi: 10.3847/1538-4357/ab5fd0 release, doi: 10.5281/zenodo.235786. Linz, H., Stecklum, B., Henning, T., Hofner, P., & Brandl, https://doi.org/10.5281/zenodo.235786 B. 2005, A&A, 429, 903, Robitaille, T. P., Whitney, B. A., Indebetouw, R., Wood, doi: 10.1051/0004-6361:200400035 K., & Denzmore, P. 2006, ApJS, 167, 256, doi: 10.1086/508424 Liu, S.-Y., Su, Y.-N., Zinchenko, I., Wang, K.-S., & Wang, Rodrigo, C., Solano, E., & Bayo, A. 2012, SVO Filter Y. 2018, ApJL, 863, L12, doi: 10.3847/2041-8213/aad63a Profile Service Version 1.0, IVOA Working Draft 15 Massi, M., Felli, M., & Churchwell, E. 1988, A&A, 194, 116 October 2012 Meyer, D. M.-A., Haemmerlé, L., & Vorobyov, E. I. 2019a, Rodr´ıguez, L. F., González, R. F., Montes, G., et al. 2012, MNRAS, 484, 2482, doi: 10.1093/mnras/sty3527 ApJ, 755, 152, doi: 10.1088/0004-637X/755/2/152 Meyer, D. M.-A., Vorobyov, E. I., Elbakyan, V. G., et al. Romine, G., Feigelson, E. D., Getman, K. V., Kuhn, M. A., 2019b, MNRAS, 482, 5459, doi: 10.1093/mnras/sty2980 & Povich, M. S. 2016, ApJ, 833, 193, Morales, E. F. E., Mardones, D., Garay, G., Brooks, K. J., doi: 10.3847/1538-4357/833/2/193 & Pineda, J. E. 2009, ApJ, 698, 488, Schlafly, E. F., Meisner, A. M., & Green, G. M. 2019, doi: 10.1088/0004-637X/698/1/488 ApJS, 240, 30, doi: 10.3847/1538-4365/aafbea Moscadelli, L., Sánchez-Monge, A.,´ Goddi, C., et al. 2016, Spitzer Science, C. 2009, VizieR Online Data Catalog, A&A, 585, A71, doi: 10.1051/0004-6361/201526238 II/293 Motte, F., Bontemps, S., & Louvet, F. 2018, ARA&A, 56, Stecklum, B., Wolf, V., Linz, H., et al. 2021, arXiv e-prints, 41, doi: 10.1146/annurev-astro-091916-055235 arXiv:2101.01812. https://arxiv.org/abs/2101.01812 Nielbock, M., Chini, R., Hoffmeister, V. H., et al. 2007, Stutzki, J., & Guesten, R. 1990, ApJ, 356, 513, ApJL, 656, L81, doi: 10.1086/512972 doi: 10.1086/168859 Nielbock, M., Chini, R., Jütte, M., & Manthey, E. 2001, Tapia, M., Roth, M., & Persi, P. 2015, MNRAS, 446, 4088, A&A, 377, 273, doi: 10.1051/0004-6361:20011053 doi: 10.1093/mnras/stu2362 Tout, C. A., Pols, O. R., Eggleton, P. P., & Han, Z. 1996, Ogbodo, C. S., Burns, R. A., Handa, T., et al. 2017, MNRAS, 281, 257, doi: 10.1093/mnras/281.1.257 MNRAS, 469, 4788, doi: 10.1093/mnras/stx1154 Towner, A. P. M., Brogan, C. L., Hunter, T. R., Pecaut, M. J., & Mamajek, E. E. 2013, ApJS, 208, 9, Cyganowski, C. J., & Friesen, R. K. 2019, ApJ, 875, 135, doi: 10.1088/0067-0049/208/1/9 doi: 10.3847/1538-4357/ab1140 Pérez-Beaupuits, J. P., Wiesemeyer, H., Ossenkopf, V., Winston, E., Megeath, S. T., Wolk, S. J., et al. 2010, AJ, et al. 2012, A&A, 542, L13, 140, 266, doi: 10.1088/0004-6256/140/1/266 doi: 10.1051/0004-6361/201218929 Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. Pérez-Beaupuits, J. P., Güsten, R., Spaans, M., et al. 2015, 2010, AJ, 140, 1868, doi: 10.1088/0004-6256/140/6/1868 A&A, 583, A107, doi: 10.1051/0004-6361/201526441 Wu, Y. W., Reid, M. J., Sakai, N., et al. 2019, ApJ, 874, Pitts, R. L., Barnes, P. J., Ryder, S. D., & Li, D. 2018, 94, doi: 10.3847/1538-4357/ab001a ApJL, 867, L7, doi: 10.3847/2041-8213/aae6ce Xu, Y., Moscadelli, L., Reid, M. J., et al. 2011, ApJ, 733, Poglitsch, A., Waelkens, C., Geis, N., et al. 2010, A&A, 25, doi: 10.1088/0004-637X 518, L2, doi: 10.1051/0004-6361/201014535 Yorke, H. W., & Bodenheimer, P. 2008, in Astronomical Povich, M. S., Smith, N., Majewski, S. R., et al. 2011, Society of the Pacific Conference Series, Vol. 387, ApJS, 194, 14, doi: 10.1088/0067-0049/194/1/14 Massive Star Formation: Observations Confront Theory, Preibisch, T., Kim, Y.-C., Favata, F., et al. 2005, ApJS, ed. H. Beuther, H. Linz, & T. Henning, 189 160, 401, doi: 10.1086/432891 Young, E. T., Herter, T. L., Güsten, R., et al. 2012, in Proven-Adzri, E., MacLeod, G. C., Heever, S. P. v. d., et al. Proc. SPIE, Vol. 8444, Ground-based and Airborne Telescopes IV, 844410 2019, MNRAS, 487, 2407, doi: 10.1093/mnras/stz1458