doi handle monthly notices of the royal astronomical …

2020Publication Year

2021-08-27T122633ZAcceptance in OAINAF

VVV-WIT-01 highly obscured classical nova or protostellar collisionTitle

Lucas P W Minniti D Kamble A Kaplan D L Cross N et alAuthors

101093mnrasstaa155DOI

httphdlhandlenet205001238630987Handle

MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETYJournal

492Number

MNRAS 492 4847ndash4857 (2020) doi101093mnrasstaa155Advance Access publication 2020 January 20

VVV-WIT-01 highly obscured classical nova or protostellar collision

P W Lucas 1lsaquo D Minniti234 A Kamble5 D L Kaplan6 N Cross7 I Dekany8

V D Ivanov9 R Kurtev210 R K Saito 11 L C Smith12 M Catelan 213

N Masetti314 I Toledo15 M Hempel3 M A Thompson1 C Contreras Pena16

J Forbrich1 M Krause 1 J Dale1 J Borissova 210 and J Emerson17

1Centre for Astrophysics University of Hertfordshire College Lane Hatfield AL10 9AB UK2Millennium Institute of Astrophysics Av Vicuna Mackenna 4860 782-0436 Macul Santiago Chile3Departamento de Ciencias Fısicas Universidad Andres Bello Fernandez Concha 700 Las Condes Santiago Chile4Vatican Observatory V00120 Vatican City State Italy5Harvard-Smithsonian Center for Astrophysics 60 Garden Street Cambridge MA 02138 USA6Center for Gravitation Cosmology and Astrophysics University of Wisconsin-Milwaukee PO Box 413 Milwaukee WI 53201 USA7Wide-Field Astronomy Unit Institute for Astronomy University of Edinburgh Royal Observatory Blackford Hill Edinburgh EH9 3HJ UK8Astronomisches Rechen-Institut Zentrum fur Astronomie der Universitat Heidelberg Monchhofstr 12-14 D-69120 Heidelberg Germany9European Southern Observatory Karl-Schwarszchild-Str 2 D-85748 Garching bei Muenchen Germany10Instituto de Fısica y Astronomıa Universidad de Valparaıso ave Gran Bretana 1111 Casilla 5030 Valparaıso Chile11Departamento de Fısica Universidade Federal de Santa Catarina Trindade 88040-900 Florianopolis SC Brazil12Institute of Astronomy University of Cambridge Madingley Road Cambridge CB3 0HA UK13Instituto de Astrofisica Pontificia Universidad Catolica de Chile Av Vicuna Mackenna 4860 7820436 Macul Santiago Chile14INAFndashOsservatorio di Astrofisica e Scienza dello Spazio via Gobetti 933 I-40129 Bologna Italy15ALMA Observatory Alonso de Cordova 3107 Vitacura Santiago Chile16School of Physics University of Exeter Stocker Road Exeter EX4 4QL UK17Astronomy Unit School of Physics and Astronomy Queen Mary University of London Mile End Road London E1 4NS UK

Accepted 2020 January 14 Received 2020 January 14 in original form 2019 November 5

ABSTRACTA search of the first Data Release of the VISTA Variables in the Via Lactea (VVV) Surveydiscovered the exceptionally red transient VVV-WIT-01 (H minus Ks = 52) It peaked beforeMarch 2010 then faded by sim95 mag over the following 2 yr The 16ndash22 μm spectral energydistribution in March 2010 was well fit by a highly obscured blackbody with T sim 1000 Kand AKs

sim 66 mag The source is projected against the Infrared Dark Cloud (IRDC) SDCG331062minus0294 The chance projection probability is small for any single event (p asymp 001ndash002) which suggests a physical association eg a collision between low mass protostarsHowever blackbody emission at T sim 1000 K is common in classical novae (especially COnovae) at the infrared peak in the light curve due to condensation of dust sim30ndash60 d after theexplosion Radio follow-up with the Australia Telescope Compact Array detected a fadingcontinuum source with properties consistent with a classical nova but probably inconsistentwith colliding protostars Considering all VVV transients that could have been projectedagainst a catalogued IRDC raises the probability of a chance association to p = 013ndash024After weighing several options it appears likely that VVV-WIT-01 was a classical nova eventlocated behind an IRDC

Key words stars formation ndash novae cataclysmic variables ndash ISM clouds ndash infrared stars

1 IN T RO D U C T I O N

Very high-amplitude near-infrared variable sources detected in theMilky Way are typically cataclysmic variables (eg Saito et al 20152016 De et al 2019) or Young Stellar Objects (YSOs) undergoingepisodic accretion variable extinction or both (eg Kospal et al

E-mail pwlucashertsacuk

2013 Contreras Pena et al 2017 Lucas et al 2017) The VVVsurvey (Minniti et al 2010 Saito et al 2012) is the first panoramicnear-infrared time domain survey of the Milky Way offering thechance to improve our knowledge of infrared variability and todiscover new types of variable star Studies of VVV transients mayassist extragalactic transient work where events such as LuminousRed Novae have uncertain physical origin (eg Kasliwal et al 2011Pastorello et al 2019) In the Milky Way the remnants of pastexplosive events can readily be detected providing examples of

Ccopy 2020 The Author(s)Published by Oxford University Press on behalf of the Royal Astronomical Society

Dow

nloaded from httpsacadem

icoupcomm

nrasarticle492448475709937 by Biblioteca di Scienze Universitagrave degli studi di Firenze user on 27 August 2021

4848 P W Lucas et al

Figure 1 The VVV DR1 survey area in the southern Galactic plane showing the locations of all the candidate high-amplitude variable stars returned by theSQL query The total area explored covers about 150 deg2 within plusmn2 of the Galactic plane The green circle represents the position of VVV-WIT-01

events that have not been seen in our galaxy in modern times Forexample the BecklinndashNeugebauer explosion in the Orion MolecularCloud 1 (OMC 1) likely the consequence of a protostellar collisionoccurred sim500 yr ago and the optically obscured supernova remnantG19 + 03 appears to be only sim100 yr old (Reynolds et al 2008) Inthis paper we report the discovery and analysis of VVV-WIT-01 anextremely red transient that was first observed by the VVV surveyin March 2010

2 VVV-WIT-01 D ISCOVERY DATA

Since 2010 the VVV survey has mapped and monitored the Galacticbulge and the adjacent southern plane in the near-infrared withVIRCAM (VISTA InfraRed CAMera) mounted on the 41-m wide-field Visible and Infrared Survey Telescope for Astronomy (VISTASutherland et al 2015) at ESO Paranal Observatory There are nowtypically between 75 and 100 epochs of data in the Ks bandpass foreach field across a total area of 560 deg2 From 2016 the surveyarea was extended to cover new fields (the rest of the southernGalactic plane at longitudes l gt 230 and l lt 20 along with theouter bulge (Minniti 2016) leading to less frequent observations ofthe original VVV fields Initial data reduction and archival mergingare carried out using the pipelines at the Cambridge AstronomicalSurvey Unit (CASU Irwin et al 2004) and VISTA Science Archive(VSA) in Edinburgh (Cross et al 2012) The VVV photometrypresented here are profile fitting photometry performed with anupdated version of DOPHOT (Schechter Mateo amp Saha 1993Alonso-Garcıa et al 2012) for all data in the original VVV fields(Smith et al in preparation) The absolute calibration is basedon the VVV photometric catalogue of Alonso-Garcıa et al (2018)which is in turn derived from the CASU v13 VISTA pipelineThe VVV Survey produces near-IR multicolour photometry in fivepassbands Z (088 μm) Y (102 μm) J (125 μm) H (165 μm)and Ks (215 μm) as well as multi-epoch photometry in Ks

The VVV 1st Data Release (VVV DR1) an SQL-based relationaldata base available at the VSA contains the VVV data taken up to2010 September 30 It includes about 58 million sources detectedat four epochs or more in a 150 deg2 area at Galactic latitudes b lt

plusmn2 (see Fig 1) VVV DR1 was created in early 2012 and an SQLquery was very soon run to search for transients and variable sourceswith gt3 magnitudes of variation The query selected candidateshaving at least four epochs of observation images in both theH and Ks bandpasses and a median Ks magnitude at least twomagnitudes brighter than the expected limiting magnitude for eachfield There were a total of sim5000 candidates the great majorityof them false positives with unremarkable colours (H minus Ks lt 2)but one extremely red candidate with (H minus Ks = 52) stood outin the colour versus amplitude diagram (see Fig 2) After passingvisual inspection it was classified as the first lsquoWITrsquo source (an

Figure 2 Ks amplitudes (Ks max minus Ks min) versus H minus Ks colour forcandidate high-amplitude variable stars returned by the SQL query of VVVDR1 The green circle indicates VVV-WIT-01

acronym for lsquoWhat is Thisrsquo) deemed worthy of further study Itwas later renamed as VVV-WIT-01 (False positives were due tomultiple causes such as saturated stars blends of two or more starsor cosmic rays)

VVV-WIT-01 is the only high-amplitude variable with such a redH minus Ks colour among 58 million sources in DR1 with four or moreepochs of photometry (Minniti et al 2012) Moreover additionalsearches of a data base of 2010ndash2018 light curves for all sources inthe original 560 deg2 VVV area have found no additional transientswith H minus Ks gt 3 (see Section 46) It was at Ks asymp 122 when firstdetected in 2010 March but faded to Ks sim 167 by July of that yearand was then undetected in VVV images taken in every subsequentyear up to 2018 to a typical 3σ upper limit of Ks = 1885 (usingannual stacks of all the relevant images taken each year) The sourcewas not detected in the VVV Z Y and J images taken in 2010 andagain in 2015 to 3σ limits of 205 205 and 200 mag in Z Y and Jrespectively Finder images showing the location of the source arepresented in Fig 3 and the time series of profile fitting photometryis given in Table 1

The observed light curve (see Fig 4) shows a rate of decline simi-lar to that of a core collapse supernova (Meikle 2000 Leibundgut ampSuntzeff 2003 Mannucci et al 2003) or perhaps a classical novagiven that the latter has widely varying decline time-scales (see Sec-tion 41) A search of the SIMBAD data base revealed that the objectis projected within SDC G331062minus0294 a small Infrared DarkCloud (IRDC) from the catalogue of Peretto amp Fuller (2009) with anequivalent-area radius of 17 arcsec This IRDC catalogue in fact lists

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4849

Figure 3 Finder images (equatorially oriented) Top 3 times 3prime WISE 3-colour image showing VVV-WIT-01 in 2010 marked with a blue circle(blue 33 μm green 46 μm and red 12 μm) The green colour isdue to high extinction and emission from warm circumstellar matterMiddle 90 times 80primeprime VVV 3-colour image (blue J green H and red Ks)from March 2010 VVV-WIT-01 is the very red star at the centre withadditional red YSO candidates adjacent Bottom Ks stacked image madefrom all VVV images of the field in 2011 The transient was no longerdetected

Table 1 VVV photometry of VVV-WIT-01

MJD-24552500 Filter Magnitude σ

8381478 Ks 1224 0028382969 Ks 1223 00211339142 Ks 1228 00211339855 Ks 1230 00221294046 Ks 1244 00221295505 Ks 1243 002187054863 Ks 1642 006187055789 Ks 1632 005189022168 Ks 1647 005189022994 Ks 1657 006213004979 Ks 1692 010213005872 Ks 1669 007214021768 Ks 1670 009214022686 Ks 1687 0118376614 H 1750 0048378080 H 1741 00421289327 H 1760 00221290754 H 1762 005

Note The quoted errors are those given by DOPHOT with a floor of 002 magimposed as the estimated calibration uncertainty

two IRDCs within 2 arcmin (the other being SDC G331030minus0299)and inspection of the SpitzerGLIMPSE 8-μm image for the regionshows that these two appear to be part of the same dark regionsim3 arcmin in extent (see Fig 5) that is seen in projection againstthe inner part of a large bubble of bright mid-infrared emissiondesignated as MWP1G331057minus002239 (Simpson et al 2012)This bubble is coincident with the H II region IRAS 16067minus5145for which Jones amp Dickey (2012) give a kinematic distance of74 plusmn 11 kpc IRDCs are the densest portions of molecular cloudsand the probability of chance association with such an object is onlysim1 per cent (see Section 4) for any single unrelated VVV transientwithin the southern half of the SpitzerGLIMPSE I survey area[where the Peretto amp Fuller (2009) catalogue is defined] We musttherefore consider the case for a prendashmain sequence origin for theevent We should also keep in mind that many transients have beendetected in VVV so the chance of observing at least one transientin an IRDC over the course of the survey is rather higher (discussedlater in Section 46) While a location within or behind an IRDCmay appear to explain the red colour simply as a consequenceof very high extinction this alone cannot explain the slope ofthe spectral energy distribution (SED) in the mid-infrared (seeSection 31)

IRDCs are initially detected simply as dark regions againstthe bright mid-infrared background of the Galactic plane SmallIRDCs from the Peretto amp Fuller (2009) catalogue can be false-positive detections caused by dips in the background emissionso confirmation is typically sought via detection of far-infraredemission from cold dust in the putative dense cloud Perettoet al (2016) performed automated checks on their 2009 catalogueusing the 250-μm HerschelHiGal images (Molinari et al 2010)to statistically assess the fraction of false positives and falsenegatives SDC G331062minus0294 passed only one of the twoprincipal checks far-infrared emission was detected but belowthe 3σ threshold because the measured noise level was inflatedby real variation on small spatial scales However the adjacentcloud SDC G331030minus0299 passed both checks Visual inspectionof the HiGal 250-μm image clearly shows emission at the twoexpected locations (brightest in between them in fact) and the sur-rounding bright and highly structured emission from the [SPK2012]

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



Figure 4 Observed Ks light curve for VVV-WIT-01 We show the VVV Survey observations from 2010 (green circles) until 2011 (upper limit) and theISAAC observation from 2012 May (green diamond) The inset shows the contemporaneous VVV and WISE observations from early 2010 where we haveaveraged data within each day for clarity The error bars are typically smaller than the point sizes

Figure 5 SpitzerGLIMPSE 8 μm 8 times 6prime

false colour image of the IRDCs in the VVV-WIT-01 field with equatorial orientation This image was taken in2004 before the transient event (marked with a white star) in SDC G331062minus0294 We see that this IRDC is part of a larger dark region that includes SDCG331030minus0299 The bright diffuse 8-μm background emission (green or white areas) is typical of the mid-plane of the inner Milky Way The white area atthe top right is the H II region IRAS 16067-5145 at d asymp 74 kpc see text

MWP1G331057minus002239 bubble dominates the image In view ofthe caveats described by Peretto et al (2016) for their automatedanalysis and after visual comparison with other small IRDCs inthe same HiGal image that pass or fail the automated tests wedo not regard this non-confirmation of SDC G331062minus0294 as

significant Given that high infrared extinction is required to fit theSED of VVV-WIT-01 (see Section 31) and YSO candidates arepresent (see Section 34 and Appendix A) it seems very probablethat SDC G331062minus0294 is a bona fide IRDC located in front ofthe H II region IRAS 16067-5145

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4851

Table 2 AllWISE multi-epoch photometry of VVV-WIT-01

MJD-55250 W1 σW1 W2 σW2 W3 σW3 W4 σW4

5564351 7918 0018 5592 0018 4764 0028 4295 01495696656 7908 0015 5669 0022 4753 0027 4117 01115828960 7929 0016 5685 0022 4746 0024 4066 01165961264 7922 0015 5775 0019 4747 0031 4115 01646027479 7915 0017 563 0016 4795 003 4211 01936093695 7934 0017 5603 0024 4743 0024 4025 01376159783 7922 0013 5667 0021 4761 0023 4157 01616225999 7975 0016 5672 0029 4696 0023 4137 01676292087 7945 0014 5621 0017 4745 0027 4153 0166424391 7974 0017 5546 0019 4722 0029 4113 0166424519 7939 0016 5575 0015 475 0033 4253 01656556823 7945 0015 5742 0028 4718 0023 414 00856689127 7942 0015 5627 0017 4698 0027 4196 01951257761 8078 0018 5805 0016 4806 0026 42 01661257773 8057 0021 5754 0021 4819 0026 4243 0171

In summary VVV-WIT-01 is a very red object (H minus Ks = 522plusmn 005 mag) with mean Ks = 123 mag in 2010 March (Fig 4)We have measured an accurate absolute position (uncertaintysim6 milliarcsec) based on the VVV profile fitting photometrycatalogues astrometrically transformed to the Gaia DR2 referenceframe (Lindegren et al 2018) using numerous stars in the vicinityof the transient (Smith et al in preparation) The J20000 equatorialcoordinates are RA = 16h10m534293s DEC = minus51d55m32439s

(J2000) and the Galactic coordinates are l = 331061347 b =0295905

3 MU LT I WAV E BA N D F O L L OW-U P A N DA R C H I VA L O B S E RVAT I O N S

31 Infrared and optical data

We searched various archival optical-infrared data bases for thissource The source is absent in the optical plates taken with the UKSchmidt Telescope available at the SuperCosmos Sky Survey (www-wfauroeacuksss) There were no detections in the followingadditional image data sets JHKs data taken on 1999 July 15 by theTwo Micron All Sky Survey (Skrutskie et al 2006) IJK data takenon 1996 April 16 by the Deep Near Infrared Survey of the SouthernSky (Epchtein et al 1999) riHα data taken on 2015 June 8 byVPHAS+ (Drew et al 2014) 36ndash80 μm mid-infrared data takenon 2004 April 2 by SpitzerGLIMPSE (Benjamin et al 2003) a24-μm image taken on 2006 April 13 by SpitzerMIPSGAL (Riekeet al 2004 Carey et al 2009) and Midcourse Space Experiment8-μm and 21-μm images taken in 1996ndash1997 (Egan et al 1998)

However the object is present and very bright in the Wide-field Infrared Survey Explorer (WISE) mid-infrared data set (withseveral images between 2010 February 28 and March 7 fortuitouslyoverlapping with the dates of the first VVV images) The WISEphotometry from the AllWISE Multi-Epoch Photometry Data baseis given in Table 2 The tabulated W2 data in fact require a positivecorrection for saturation of 007 mag (Cutri et al 2012) this appliesin a similar fashion for WISE Allsky and AllWISE data takenduring the cryogenic portion of the mission After making thiscorrection the mean magnitudes are W1(33 μm) = 795 plusmn 001W2(46 μm) = 573 plusmn 005 W3(12 μm) = 475 plusmn 001 andW4(22 μm) = 416 plusmn 002 mag The brightness of VVV-WIT-01steadily declined in March 2010 (see Fig 4) matching the slope ofthe near-simultaneous VVV observations in Ks

After March 2010 the source was no longer detected in thesensitive WISE W1 or W2 passbands (we inspected the deeperunWISE W2 stacked image constructed from data taken on 2010August 28 to August 31 [Meisner Lang amp Schlegel 2018] butthe location is severely blended with two adjacent stars in thesim6 arcsec WISE beam preventing detection of stars with W2 ge11 mag) However the stacked W3 image taken at that time shows aclear but uncatalogued detection of a point source at the position ofthe transient There are no blending issues in this waveband becausefewer stars are detected Moreover this source is not present in theGLIMPSE 8-μm image taken in 2004 (see Fig 5) so we can be con-fident that it is VVV-WIT-01 We suspect that the AllWISE pipelinedid not detect it because the sky background level is measured witha coarse spatial grid so the flux peak within the small IRDC did notstand out sufficiently above the (overestimated) background levelWe performed aperture photometry on the W3 FITS image obtainedfrom the NASA Infrared Science Archive with IRAFDAOPHOTbootstrapping the zero-point using several nearby sources in theAllWISE source catalogue We found W3 = 678 plusmn 014 magon 2010 August 29 (the mid-point date of the image stack) themeasurement error being dominated by the uncertain backgroundlevel within the IRDC This compares with Ks = 1645 on thesame date (averaging VVV data from August 28 and August 30)indicating that the transient had faded by 42 mag at 215μm but only20 mag at 12μm NB the 2010 August data were from the lsquo3-BandCryorsquo phase of the WISE mission so no data were taken in W4

The near simultaneity of the VVV and WISE measurements inMarch 2010 allows us to define an SED see Fig 6 (upper panel)We attempted to fit the SED using a reddened uniform temperatureblackbody ie the model is log10(λBλ(T)10minus04A(λ)) We constructeda library of blackbody SEDs with a range of temperatures andextinctions (60lt T lt 6000 K and 0 lt AKs

lt 20) and located thearea of parameter space with the minimum value of the reduced χ2

parameter χ2ν computed with 5 degrees of freedom Each model

was scaled to the same average flux as the data measured as theaverage of log(λFλ) at the six wavelengths To relate extinctionin Ks and the WISE passbands we adopted the mid-infraredextinction law defined by Koenig amp Leisawitz (2014) for fields withhigh extinction separately the Cardelli Clayton amp Mathis (1989)extinction law was used to relate the extinction in H and Ks1 This

1We also considered the modified near-infrared extinction law of Alonso-Garcıa et al (2017) derived for VVV data in the inner bulge Steeper

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



Figure 6 Fit to the SED of VVV-WIT-01 (upper panel) The SED around2010 March 1 is well fit by a blackbody model with high extinction (T =1020 K AKs = 66) Blue points show the SED on 2010 August 29including the upper limit at 46 μm Lower panel A plot of χ2

ν as a functionof extinction and temperature

simple model produced a very good fit for T = 1020 K AKs= 66

(see upper panel of Fig 6) In the lower panel of Fig 6 we illustratehow χ2

ν varies in the parameter space combinations of temperaturesand extinction in the ranges 720 lt T lt 2100 K 36 lt AKs

lt 101can produce values of χ2

ν less than four times the minimum value(As expected χ2

ν gt 1 in all models due to systematic uncertaintiesin the mid-infrared extinction law that become important whenextinction is high) The best-fitting extinction value AKs

= 66 isunusually high for an IRDC in the Peretto amp Fuller (2009) catalogueso within the likely range of 36 lt AKs

lt 101 a lower extinctionslightly lower temperature model is perhaps more likely unlessthere is some extinction associated with the transient itself

If we assume that the same extinction was present in early Marchand late August 2010 then the ratio of Ks and W3 fluxes from29 August can be used to calculate a blackbody temperature Thedata indicate significant cooling eg from 1020 K to 713 K (ifAKs

= 66) or from 720 K to 602 K (if AKs= 36) However the

upper limit of W2 gt 11 at this epoch corresponds to a fade of over5 mag at 46 μm larger than the 42-mag fade at 215 μm andthe 20 mag at 12 μm After including this upper limit in W2 the

near-infrared extinction laws such as this lead to best-fitting solutions withslightly lower values of both extinction and temperature not significantlychanging our conclusions

SED is not well described by a single temperature blackbody in lateAugust (see the blue points in Fig 6) for any value of extinctionNonetheless the relative brightness at 12 μm suggests a shift tolower temperatures

Soon after the discovery of the transient in VVV DR1 inearly 2012 we obtained deeper and higher resolution follow-upobservations on 2012 May 10 using the Infrared Spectrometer AndArray Camera (ISAAC Moorwood et al 1998) at the ESO VeryLarge Telescope (ESO programme 289D-5009) We also obtainedSpitzerIRAC images on 2012 May 29 (MJD 56076787) underSpitzer programme 80259 The ISAAC data showed that VVV-WIT-01 had faded to Ks = 2034 plusmn 009 mag The IRAC data alsoshowed a point source with I2(45 μm) = 1522 plusmn 030 mag (Thesource is clearly detected but there is uncertainty in the backgroundlevel because two much brighter stars are present on either sideof VVV-WIT-01 and there are also gradients in the surroundingnebulosity) In the I1 passband we measure a 3σ upper limitI1(36 μm) gt165 mag The I2 magnitude can be compared withthe earlier WISE W2 magnitude as the passbands are similar Thecomparison shows that the variable source has faded by sim95 magat 45 μm over the 2-yr interval

32 High energy and neutrino data

We note that the ANTARES neutrino project (Ageron et al 2011)did not detect any non-solar astronomical neutrino sources inthe 2007ndash2012 period (see httpantaresin2p3frpublicdata2012html) The Super-Kamiokande and SNEWS neutrino projects(Fukuda et al 2003 Antonioli et al 2004) also did not detect anyevents in a 9-month interval prior to 2012 March (M Vagins and KScholberg private communication) Similarly there are no reporteddetections of γ -ray or hard X-ray sources that can be connected withthis transient in alert data and catalogue data from the Fermi Swiftand INTEGRAL satellites (Krivonos et al 2012 Ackermann et al2013 Oh et al 2018)

We moreover obtained targeted X-ray follow-up observationsof the VVV-WIT-01 field with SwiftXRT from 2012 April 21 toMay 13 totalling 8123 s on source in the 03ndash10 keV passband NoX-ray source was detected with a 3σ upper limit of 3 times 10minus14 ergcmminus2 sminus1

33 Radio cm continuum data

Following the initial announcement of this transient event (Minnitiet al 2012) the field was observed with the Australia TelescopeCompact Array (ATCA) at 21 GHz 55 GHz and 9 GHz on 2012May 21 some 2 yr after the first detections (ATCA programmeCX240) The observations used the longest baseline array configu-ration (6 km) and lasted for a full track providing good uv coverageand a spatial resolution θ asymp 1 arcsec at the two higher frequencieswith a symmetric beam The phase and gain calibrator was 1613-586 The MIRIAD software package was used to reduce the datausing standard procedures to flag and remove bad data VVV-WIT-01 was clearly detected at 55 GHz and 9 GHz with fluxes of276 plusmn 10 μJy and 236 plusmn 37 μJy respectively It was not detected at21 GHz with a 2σ upper limit of approximately 400 μJy A powerlaw fit to the detections and upper limit yields a spectral index α =minus015 plusmn 022 for Fν prop να essentially a flat radio spectrum

A second set of ATCA observations at 55 GHz and 9 GHzwas obtained 7 yr later on 2019 June 2425 with a duration ofalmost a full track (ATCA programme C3309) Again the 6-kmconfiguration was used The phase and gain calibrators observed

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853

on this occasion were 1646-50 and 1600-48 the latter being usedonly at the start of the run before 1646-50 rose high enough Anadjacent uncatalogued continuum source at 1610594 -5149240was used for self-calibration at 55 GHz but this was not possible at9 GHz VVV-WIT-01 was no longer detected with 3σ upper limitsof approximately 40 μJy at both frequencies

34 Radio mmsubmm continuum data

The transient was also followed up with the Atacama LargeMillimetersubmillimeter Array (ALMA) using the 12-m antennasin the main array Observations were made in the continuum inALMA band 3 (26ndash36 mm) band 6 (11ndash14 mm) and band 7 (08ndash11 mm) in 2013 and 2015 (ALMA project codes 2012100463Sand 2013A00023S) These were reduced with the standard datareduction pipeline by observatory staff No source was detected atthe coordinates of VVV-WIT-01 to 3σ upper limits of approximately01 mJy in all three bands In bands 6 and 7 two sources weredetected within the area of SDC G331062minus0294 both of themoffset from VVV-WIT-01 by a few arcseconds One of these sourcesis coincident with a likely YSO VVV J16105393-5155328 thathappens to show an sim1 mag infrared outburst in 2015 This is oneof the two red stars visible to either side of VVV-WIT-01 in themiddle panel of Fig 3 both of which are YSO candidates We givedetails of VVV J16105393-5155328 in Appendix A The otherALMA detection has no infrared counterpart and may well be anextragalactic object

4 D ISCUSSION AND SEARCHES FORA D D I T I O NA L EV E N T S

41 Protostellar collision interpretation

Owing to the location projected against an IRDC we must considerthe case for a genuine physical association and a prendashmain sequenceorigin IRDCs are numerous in the inner Galactic plane but relativelysmall Using the equivalent area radii from Peretto et al (2016) forIRDCs that passed their far-infrared confirmation test we derive acovering fraction f = 12 per cent within the southern portion ofthe GLIMPSE-I survey area at 295 lt l lt 350 |b| lt 1 where theIRDC catalogue of Peretto et al (2016) was defined This risesto f = 23 per cent if we include IRDCs that did not pass theautomated confirmation test These small covering fractions led usto explore the possibility that the transient was a collision betweenYSOs within the IRDC SDC G331062minus0294 Collisions whichwill often lead to mergers can cause high-amplitude transient eventssuch as the lsquomergeburstrsquo observed in V1309 Sco (arising in acontact eclipsing binary system Tylenda et al 2011) and other likelyexamples such as V838 Mon and V4332 Sgr (Martini et al 1999Tylenda amp Soker 2006 Kochanek Adams amp Belczynski 2014) Asnoted earlier such events are a possible explanation for lsquoluminousred novaersquo seen in external galaxies

We note that the high amplitude of the event (ge95 mag at45 μm) tends to argue against some form of episodic accretionevent in a protostar because infrared variations in the most extremeexamples of such events have not exceeded sim6 to 7 mag hithertoeg Audard et al (2014) An apparent protostellar eruption withan unprecedented mid-infrared amplitude of 85 mag has recentlybeen discovered in the public WISENEOWISE GLIMPSE andVVV data (Lucas et al in preparation) but the event has a longplateau and a slow decline over several years not resembling therapid decline seen in VVV-WIT-01

Bally et al (2017) provide three examples of explosion remnantswithin Galactic star formation regions that are likely due to historicalcollisions between YSOs The best-studied example is the BecklinndashNeugebauer event in OMC-1 where recent data indicate that threeYSOs are moving away from a common origin some 500 yr agoat the centre of the explosion remnant (Kuiper et al 2010b) Thephysics of collisions between protostars and mergers is discussedin detail by Bally amp Zinnecker (2005) The topic is interestingin part because mergers are a possible mode of formation formassive stars (Bonnell Bate amp Zinnecker 1998) though it isnow thought likely that massive stars can form without recourseto this mechanism (Kuiper et al 2010a Baumgardt amp Klessen2011 Kuiper amp Hosokawa 2018) If collisions in star-formingregions do occur the relatively low space density in such regionswould require some explanation other than random motions withinthe cluster potential even when gravitational focusing is allowedfor (Shara 2002) Possible explanations might include focusing ofYSOs within relatively small volumes due to formation within hub-filament systems of dense gas within molecular clouds (Myers 2009Peretto et al 2013 Williams et al 2018) ongoing accretion oflow angular momentum gas in a star-forming cluster (Gieles et al2018) or perhaps the unstable dynamical decay of newborn multiplesystems (Rawiraswattana Lomax amp Goodwin 2012 Moeckel ampBonnell 2013 Railton Tout amp Aarseth 2014)

No massive protostar remained visible in the vicinity of VVV-WIT-01 after the transient event so a collision would have to involvelow mass YSOs This is not an obstacle because the great majority ofYSOs have low masses and they can appear faint when embedded inan IRDC at a distance of several kpc IRDCs have typical distancesof 1ndash8 kpc (Simon et al 2006) and we inferred in Section 2 thatSDC G331062minus0294 is probably located in front of the bright H II

region IRAS 16067-5145 at d asymp 74 kpc The recent 12CO-basedmolecular cloud catalogue of Miville-Deschenes Murray amp Lee(2017) lists at least four massive molecular clouds that appear tooverlap the location of VVV-WIT-01 with likely distances rangingfrom 15 to 85 kpc The distance to SDC G331062minus0294 can beconstrained using blue foreground stars with parallaxes from GaiaGaia source 5933412592646831872 is projected in this IRDC andhas a likely minimum parallax-based distance of 24 kpc in Bailer-Jones et al (2018) Source 5933458016223623552 projected in theadjacent apparently connected IRDC SDC G331030minus0299 has alikely minimum parallax-based distance of 31 kpc (based on a 7σ

parallax and neglecting the small systematic parallax error in GaiaDR2) Combining these lower limits with the upper limit from thebackground H II region we can assume 31 lt d lt 74 kpc for theseIRDCs Unfortunately it was not possible to calculate a red clumpgiant distance (Lopez-Corredoira et al 2002) because the IRDCsare too small to have a projected population of such stars The giantbranch in the 2MASS and VVV Ks versus J minus Ks colour magnitudediagrams for the larger region within 3 arcmin of VVV-WIT-01 (notshown) shows a fairly smooth increase in extinction with distance atd gt 3 kpc consistent with the existence of several large molecularclouds along this line of sight

42 Supernova hypotheses

The non-detections of neutrinos and γ -rays (see Section 32) ruleout a highly obscured core collapse supernova on the far side of theGalactic disc Such events are very luminous γ -ray and hard X-raysources we would also expect to detect the neutrino signal becauseit is very rare for all three of the neutrino observatories listed inSection 32 to be offline simultaneously

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



These neutrino and high-energy radiation checks were neces-sary because a relatively low-luminosity core collapse supernovawith peak absolute magnitude MKs

= minus16 subject to extinctionAKs

= 10 mag (the maximum plausible value see above) mighthave apparent magnitude mKs

= 105 if located at d = 20 kpc fromthe Sun the infrared data alone would not entirely rule out thispossibility because VVV-WIT-01 was discovered after the peakSimilarly while a core collapse supernova would typically be afar brighter cm continuum radio source than was observed (seeSection 33) rare examples with blue supergiant progenitors can beradio-quiet (Weiler amp Sramek 1988)

A Type Ia supernova in the inner Milky Way would be radio-quietand lack detectable neutrino emission (Odrzywolek amp Plewa 2011)However such an event would have been detected in γ -rays by theFermi Swift or INTEGRAL telescopes due to strong emission atspecific energies associated with the decay of 56Ni and 56Co in theejecta see Wang Fields amp Lien (2019) for a full discussion Forexample in the case of SN2014j at 33 Mpc distance the 847 keVline of 56Co was detected by INTEGRAL for more than 4 months(Diehl et al 2015)

43 Helium flash hypothesis

A few examples of a very late thermal pulse in a post-AsymptoticGiant Branch star have been observed over the past century egV4334 Sgr (Sakurairsquos object) and V605 Aql (see eg Hajduk et al2005) In these events the star returns from the white dwarf coolingtrack to giant star status but then remains luminous (L sim 104L)for hundreds of years Such stars can appear almost nova-like atvisible wavelengths because the rise in flux takes place over a timeof order 1 yr and subsequent formation of dust in the ejecta cancause a rapid optical fading as in V605 Aql The near-infrared fluxcan also fade and re-brighten as the ejecta cool and the observedstructure changes (Kimeswenger Koller amp Schmeja 2000 Hinkle ampJoyce 2002 Evans et al 2006 Clayton et al 2013 Hinkle amp Joyce2014) but the mid-infrared flux remains high for several years withmeasured temperatures exceeding 500 K The data for VVV-WIT-01 do not seem consistent with a helium flash event because whilesome cooling apparently occurred over the 6-month interval afterfirst detection the source faded by 2 mag at 12 μm and by muchlarger amounts at 46 μm and 215 μm (see Section 31) indicatingthat the total luminosity probably declined by a large factor duringthis time This rapid decline and the continued decline from 2010 to2012 at 46 μm contrast with the approximately constant luminosityof V4334 Sgr over several years (Hinkle amp Joyce 2014) The relativerarity of these events also means that a chance association with anIRDC would be unlikely

44 Classical nova interpretation

A clear case may be made for a classical nova located behind theIRDC Classical novae are the commonest type of transient seen inVVV at least 20 likely examples have been discovered within theVVV data set many of them in the lsquoVVV discrsquo area at Galactic lon-gitudes 295 lt l lt 350 (eg Saito et al 2015 2016) and many othershave been independently discovered within the Galactic bulge since2010 by VVV WISE or optical surveys such as OGLE and ASASClassical novae (and recurrent novae which are novae that havebeen observed to erupt more than once) have outburst amplitudesfrom 7 to 20 mag in the optical (eg Strope Schaefer amp Henden2010 Osborne 2015) and a very wide range of decay time-scales inthe sample of 95 optical nova light curves presented by Strope et al

(2010) the t3 parameter (the time for the outburst to decline by 3 magfrom the peak at optical wavelengths) ranges from 3 d to 900 dStrope et al (2010) illustrated considerable variety in the morpho-logical features of nova light curves that remains poorly understood

In the infrared only a few well sampled novae have been studiedin detail (see eg Gehrz 1988 1999 Hachisu amp Kato 2006 Gehrzet al 2015) though fairly well sampled K bandpass light curvesare now available for many novae in the SMARTS data base(wwwastrosunysbedufwalterSMARTSNovaAtlas see Walteret al 2012) Many novae have a second peak in the infrared lightcurve some 30ndash60 d after the initial opticalinfrared maximumdue to the condensation of a dust shell within the ejecta that issometimes optically thick and sometimes optically thin perhapsdue to clumpiness (see eg Gehrz 1988 Hachisu amp Kato 2006 Daset al 2013 Burlak et al 2015) The SED of the dust shell is wellfit by emission from a blackbody at temperatures from sim700 to1100 K Examples include Nova Cyg 1978 (Gehrz et al 1980b) andNova Ser 1978 (Gehrz et al 1980a) while in the SMARTS data basethe light curves of eg V906 Car and V5668 Sgr appear to show aprominent K bandpass dust peak A large amount of dust productionis more common in CO novae originating in low mass white dwarfsthan in ONeMg novae wherein the white dwarf has a typical mass Mgt 12 M (Gehrz 1999) Our derived temperature near 1000 K forthe Planckian emission from VVV-WIT-01 is clearly a close matchto a classical nova soon after the epoch of dust condensation

The absolute visual magnitudes of classical novae at peakbrightness typically lie in the range minus10 lt MV lt minus5 (Warner 1987)and Ks magnitudes are similar to V at this time in the absence ofreddening The absolute Ks magnitude of the second peak after theepoch of dust formation can vary widely depending on the opticaldepth and temperature of the dust (eg Hachisu amp Kato 2019) but insome of the above examples it is similar to or only 1 or 2 mag fainterthan the first peak Adopting our best-fitting values of extinctionand temperature AKs

= 66 T asymp 1000 K for VVV-WIT-01 thebrightest Ks magnitudes in 2010 March would imply a distancemodulus m minus M = 106 to 156 or d = 13ndash13 kpc if we wereobserving the initial outburst If we adopt a lower extinction modelmore typical of IRDCs within the likely range of 36 lt AKs

lt 101(see Section 31) these distances can rise by up to a factor of 3despite the slightly lower temperature associated with these modelsThe photometric maximum was not in fact observed so the distanceis not well constrained in this interpretation though it would haveto be larger than the minimum distance of 31 kpc to the IRDC (seeSection 41) Nonetheless a classical nova with a prominent dustpeak could plausibly be located at a distance of several kpc iebehind the IRDC

45 The cm continuum test

The cm continuum data from ATCA provide a valuable test of theprotostellar collision and nova hypotheses Numerous radio shellsassociated with classical novae have been observed previouslyNearby novae have typical fluxes of the order of a few mJy at4ndash9 GHz continuum when observed sim2 yr after the event and theradio spectral index at those frequencies is typically approximatelyflat at that time consistent with optically thin freendashfree emission(Hjellming et al 1979 Seaquist 2008 Finzell et al 2018) Opticallythin nova shells then fade fairly rapidly at these frequencies as theshell becomes increasingly optically thin with Sν prop (t minus t0)k wherek ranges from minus3 to minus1 and t0 is the time of the explosion If VVV-WIT-01 was a classical nova we would expect the radio emissionto have faded by a factor of at least 45 and more likely by one

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855

or two orders of magnitude between the 2012 and 2019 ATCAobservations The ATCA non-detection in 2019 at a level an orderof magnitude fainter than the 2012 flux is therefore consistent withthis interpretation The flat spectral index and sim025 mJy flux levelin 2012 are also consistent with a classical nova at a distance ofseveral kpc

In the case of a collision between YSOs numerical simulationsof relatively violent collisions (those would produce a transientevent of nova-like luminosity) predict that between 1 per cent and10 per cent of the combined stellar mass would be ejected ie a massof order a few times10minus2 M to a few times10minus1 M for a collision betweenlow mass stars (Laycock amp Sills 2005 Dale amp Davies 2006 Soker ampTylenda 2006) This is supported by an empirical estimate of 1 Mejected in the case of the massive BecklinndashNeugebauer event (Ballyet al 2017) The ejection velocities would be lower than in a classicalnova of order 102 km sminus1 (ie a little less than the escape velocity)this is in line with the ejection velocities observed in V1309 ScoV4332 Sgr and V838 Mon (Tylenda amp Soker 2006 Kaminski et al2009 2015 and references therein) Classical novae typically ejecta few times 10minus4 M eg Gehrz (1999) at speeds of order 103 km sminus1Since the ejected mass from a stellar collision is likely to be twoor three orders of magnitude greater and the velocity is likely to bealmost an order of magnitude less we would expect the collisionejecta to remain optically thick in the 4-cm continuum for sim2 ordersof magnitude longer than the 2-yr time-scale that is characteristic ofnovae Freendashfree emission from optically thick ejecta tends to slowlybrighten over time as the surface area increases as is observed inthe early stages of nova shell expansion so in such an event wewould have expected the radio continuum flux to be brighter in the2019 ATCA observation than in 2012 The non-detection in 2019therefore argues fairly strongly against the collision hypothesis

46 Searches for additional very red transients

In order to further test our two main hypotheses we searched thecurrent working data base of VVVVVVX 2010ndash2018 light curvescovering the original 560 deg2 area (see Section 2) to determine howmany transient events have occurred and whether there have beenany other exceptionally red transients in star-forming regions Wedetected 59 transient events with amplitudes over 4 mag 12 of whichoccurred within the boundary of the GLIMPSE-I survey at 295 lt llt 350 minus1 lt b lt 1 where the IRDC catalogue of Peretto amp Fuller(2009) is defined None of these events were projected in IRDCsand none have SEDs as red as VVV-WIT-01 Some events that weredetected by WISE eg VVV-NOV-005 (Saito et al 2015) havecolours in the range 1 lt W1 minus W2 lt 2 consistent with emissionfrom hot dust Most of these 59 transients have been previouslyidentified as novae or nova candidates but some are new theirinfrared light curves will be the subject of a separate paper

With 12 events over the 2010ndash2018 interval in the GLIMPSE Isurvey area the IRDC covering fractions f noted in Section 41imply that the chance of observing at least one in an IRDC is p =1 minus (1 minus f)12 = 013 (including only confirmed IRDCs) or p =024 (including all IRDC candidates) In view of this result thenova hypothesis appears very plausible It was merely surprisingat the time to detect such a chance projection among the firstVVV transients near the start of the survey A separate search fortransients in IRDCs was conducted with the WISENEOWISE timeseries data from 2010 and 2014ndash2017 We constructed a variablestar and transient source catalogue for all sources within 2 arcmin ofthe 7139 confirmed IRDCs listed in Peretto et al (2016) This search

did not find any additional transients though some high-amplitudevariable sources were detected (Lucas et al in preparation)

47 Upper limit on the protostellar collision rate

The ATCA data and the incidence of VVV transients have led usto conclude that VVV-WIT-01 was very probably a classical novabehind an IRDC The main caveats are that our empirical knowledgeof stellar mergers is limited to a few probable events and the presentgeneration of numerical simulations is unlikely to be the final wordWe can use the non-detection of stellar mergers in VVV to place alimit on the incidence of luminous collisions between YSOs Oursearch of the 2010ndash2018 VVV data spanned sim85 yr and covered anarea of sky that encompasses sim1011 stars (assuming the Milky Waycontains at least twice that number of stars) Adopting a constant starformation rate (for simplicity) and typical stellar lifetimes in excessof 1010 yr a fraction of order 10minus4 ie 107 stars has ages le1 MyrOur detection efficiency is asymp05 given that events occurring inthe southern summer may have been missed by our search Thenon-detection in 85 yr then implies that the rate is probably le02collisions per year For 107 YSOs with ages le1 Myr likely to still bein a crowded and dynamically unstable environment a limit of 02collisions per year implies a 1 in 5 times 107 chance of a collision peryear for individual YSOs or 1 in 50 per Myr per YSO This upperlimit is at a level that begins to be useful confirming that while suchcollisions are rare there may be numerous such events within thelifetime of massive prendashmain sequence clusters of several thousandsstars such as the Orion Nebula Cluster A careful search of datasets such as VVV GLIMPSE WISE and the United KingdomInfrared Deep Sky Survey (UKIDSS) may yield more examples ofthe remnants of prendashmain sequence mergebursts to add to the threelisted by Bally amp Zinnecker (2005)

5 C O N C L U S I O N S

The VVV-WIT-01 transient event was uniquely red among tran-sients detected in the VVV data set thus far It was also unique inits projected location in an IRDC The contemporaneous timing ofobservations with VISTA and WISE in 2010 March was fortunatebut the sparse sampling of the light curve and lack of a spectrummake a definitive classification difficult Nonetheless the absenceof γ -ray and neutrino emission allows us to rule out a Galacticsupernova and the very high amplitude and steep decline in fluxmake it unlikely that this was an episodic accretion event in a YSOThe steep decline in mid-infrared flux also appears to rule out aVery Late Thermal Pulse in a post-AGB star

The hypothesis of a highly obscured classical nova with a brightdust peak appears to fit all the available data While it was initiallysurprising that such an event should be projected in an IRDC inthe first year of the survey VVV subsequently detected severaldozen transients in the 2010ndash2018 interval including 12 in theSpitzerGLIMPSE region where the Peretto amp Fuller (2009) IRDCcatalogue was defined Consequently the chance of detecting asingle such event over the course of the survey is calculated asbetween 13 per cent and 24 per cent not a very unlikely occurrence

The protostellar collisionmergeburst hypothesis is the mostinteresting one that we have considered given that the remnantsof a few such events involving prendashmain sequence stars havebeen observed in the Milky Way (Bally amp Zinnecker 2005) Therapidly fading 4-cm continuum flux detected by ATCA appears tobe inconsistent with this interpretation causing us to favour thenova hypothesis A cautionary note remains that our prediction

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



regarding the radio evolution of a mergeburst rests on theoreticalsimulations in a field that is still maturing We should remain opento the possibility that the transient was a new type of event Anydetections of additional very red transients in star-forming regionswould require us to reconsider the classical nova interpretation

AC K N OW L E D G E M E N T S

We gratefully acknowledge data from the ESO Public Survey pro-gram ID 179B-2002 taken with the VISTA telescope and productsfrom the Cambridge Astronomical Survey Unit (CASU) We thankthe referee Nye Evans for a helpful and constructive report We alsothank M Vagins and K Scholberg for checking their neutrino dataduring the relevant 2009ndash2010 time period This publication makesuse of data products from the WISE satellite which is a joint projectof the University of California Los Angeles and the Jet Propul-sion LaboratoryCalifornia Institute of Technology funded by theNational Aeronautics and Space Administration (NASA) Thisresearch has made use of NASArsquos Astrophysics Data System Biblio-graphic Services and the SIMBAD data base operated at CDS Stras-bourg France PWL acknowledges support by Science and Technol-ogy Faciities Council (STFC) Consolidated Grants STR009051STM0010081 and STJ0013331 and the STFC PATT grantSTR00126X1 DM acknowledges support from the FONDECYTRegular grant no 1170121 the BASAL Center for Astrophysics andAssociated Technologies (CATA) through grant AFB170002 andthe Ministry for the Economy Development and Tourism ProgramaIniciativa Cientifica Milenio grant IC120009 awarded to the Millen-nium Institute of Astrophysics (MAS) DLK was supported by NSFgrant AST-1816492 NM acknowledges financial support from ASI-INAF contract n2017-14-H0 MH acknowledges financial supportfrom the Comite Mixto ESO-Gobierno de Chile Support for MC isprovided by Proyecto Basal AFB-170002 by the Ministry for theEconomy Development and Tourismrsquos Millennium Science Initia-tive through grant IC 120009 awarded to the Millennium Instituteof Astrophysics (MAS) and by FONDECYT project 1171273

RE FERENCES

Ackermann M et al 2013 ApJ 771 57Ageron M et al 2011 Nucl Instrum Methods Phys Res A 656 11Alonso-Garcıa J et al 2017 ApJ 849 L13Alonso-Garcıa J et al 2018 AampA 619 A4Alonso-Garcıa J Mateo M Sen B Banerjee M Catelan M Minniti D

von Braun K 2012 AJ 143 70Antonioli P et al 2004 New J Phys 6 114Audard M et al 2014 Protostars and Planets VI University of Arizona

Press Tucson AZp 387Bailer-Jones C A L Rybizki J Fouesneau M Mantelet G Andrae R

2018 AJ 156 58Bally J Zinnecker H 2005 AJ 129 2281Bally J Ginsburg A Arce H Eisner J Youngblood A Zapata L

Zinnecker H 2017 ApJ 837 60Baumgardt H Klessen R S 2011 MNRAS 413 1810Beckwith S V W Sargent A I Chini R S Guesten R 1990 AJ 99 924Benjamin R A et al 2003 PASP 115 953Bonnell I A Bate M R Zinnecker H 1998 MNRAS 298 93Burlak M A Esipov V F Komissarova G V Shenavrin V I Taranova

O G Tatarnikov A M Tatarnikova A A 2015 Balt Astron 24 109Cardelli J A Clayton G C Mathis J S 1989 ApJ 345 245Carey S J et al 2009 PASP 121 76Clayton G C et al 2013 ApJ 771 130Contreras Pena C et al 2017 MNRAS 465 3011Cross N J G et al 2012 AampA 548 A119

Cutri R M et al 2012 Technical Report Explanatory Supplement to theWISE All-Sky Data Release Products Available at httpwise2ipaccaltechedudocsreleaseallskyexpsupindexhtml

Dale J E Davies M B 2006 MNRAS 366 1424Das R K Banerjee D P K Ashok N M Mondal S 2013 Bull Astron

Soc India 41 195De K et al 2019 The Astronomerrsquos Telegram 13130 1Diehl R et al 2015 AampA 574 A72Drew J E et al 2014 MNRAS 440 2036Egan M P Shipman R F Price S D Carey S J Clark F O Cohen M

1998 ApJ 494 L199Epchtein N et al 1999 AampA 349 236Evans A et al 2006 MNRAS 373 L75Finzell T et al 2018 ApJ 852 108Fukuda S et al 2003 Nucl Instrum Methods Phys Res A 501 418Gehrz R D et al 2015 ApJ 812 132Gehrz R D 1988 ARAampA 26 377Gehrz R D 1999 Phys Rep 311 405Gehrz R D Grasdalen G L Hackwell J A Ney E P 1980a ApJ 237

855Gehrz R D Hackwell J A Grasdalen G I Ney E P Neugebauer G

Sellgren K 1980b ApJ 239 570Gieles M et al 2018 MNRAS 478 2461Gutermuth R A Megeath S T Myers P C Allen L E Pipher J L Fazio

G G 2009 ApJS 184 18Hachisu I Kato M 2006 ApJS 167 59Hachisu I Kato M 2019 ApJS 242 18Hajduk M et al 2005 Science 308 231Hildebrand R H 1983 QJRAS 24 267Hinkle K Joyce R 2002 ApampSS 279 51Hinkle K H Joyce R R 2014 ApJ 785 146Hjellming R M Wade C M Vandenberg N R Newell R T 1979 AJ

84 1619Irwin M J et al 2004 in Quinn P J Bridger A eds Proc SPIE Conf

Ser Vol 5493 Optimizing Scientific Return for Astronomy throughInformation Technologies SPIE Bellingham p 411

Johnstone D Hendricks B Herczeg G J Bruderer S 2013 ApJ 765133

Jones C Dickey J M 2012 ApJ 753 62Kaminski T Mason E Tylenda R Schmidt M R 2015 AampA 580 A34Kaminski T Schmidt M Tylenda R Konacki M Gromadzki M 2009

ApJS 182 33Kasliwal M M et al 2011 ApJ 730 134Kimeswenger S Koller J Schmeja S 2000 AampA 360 699Kochanek C S Adams S M Belczynski K 2014 MNRAS 443 1319Koenig X P Leisawitz D T 2014 ApJ 791 131Kospal A et al 2013 AampA 551 A62Krivonos R Tsygankov S Lutovinov A Revnivtsev M Churazov E

Sunyaev R 2012 AampA 545 A27Kuiper R Hosokawa T 2018 AampA 616 A101Kuiper R Klahr H Dullemond C Kley W Henning T 2010a AampA

511 A81Kuiper R Klahr H Beuther H Henning T 2010b ApJ 722 1556Laycock D Sills A 2005 ApJ 627 277Leibundgut B Suntzeff N B 2003 in Weiler K ed Lecture Notes in

Physics Vol 598 Supernovae and Gamma-Ray Bursters Springer-Verlag Berlin p 77

Lindegren L et al 2018 AampA 616 A2Lopez-Corredoira M Cabrera-Lavers A Garzon F Hammersley P L

2002 AampA 394 883Lucas P W et al 2017 MNRAS 472 2990Mannucci F et al 2003 AampA 401 519Martini P Wagner R M Tomaney A Rich R M della Valle M

Hauschildt P H 1999 AJ 118 1034Megeath S T et al 2004 ApJS 154 367Meikle W P S 2000 MNRAS 314 782Meisner A M Lang D Schlegel D J 2018 AJ 156 69Minniti D et al 2010 New A 15 433

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857

Minniti D 2016 Galactic Surveys New Results on Formation EvolutionStructure and Chemical Evolution of the Milky Way Sesto (BZ) Italyp 10

Minniti D Lucas P W Cross N Ivanov V D Dekany I Kurtev R 2012The Astronomerrsquos Telegram 4041

Miville-Deschenes M-A Murray N Lee E J 2017 ApJ 834 57Moeckel N Bonnell I A 2013 Primordial triples and collisions of massive

stars preprint (arXiv13016959)Molinari S et al 2010 AampA 518 L100Moorwood A et al 1998 The Messenger 94 7Myers P C 2009 ApJ 700 1609Odrzywolek A Plewa T 2011 AampA 529 A156Oh K et al 2018 ApJS 235 4Osborne J P 2015 J High Energy Astrophys 7 117Pastorello A et al 2019 AampA 625 L8Peretto N et al 2013 AampA 555 A112Peretto N Fuller G A 2009 AampA 505 405Peretto N Lenfestey C Fuller G A Traficante A Molinari S Thompson

M A Ward-Thompson D 2016 AampA 590 A72Railton A D Tout C A Aarseth S J 2014 PASA 31 e017Rawiraswattana K Lomax O Goodwin S P 2012 MNRAS 419 2025Reynolds S P Borkowski K J Green D A Hwang U Harrus I Petre

R 2008 ApJ 680 L41Rieke G H et al 2004 ApJS 154 25Saito R K et al 2012 AampA 537 A107Saito R K Minniti D Angeloni R Dekany I Catelan M 2015 The

Astronomerrsquos Telegram 7124Saito R K Minniti D Catelan M Angeloni R Beamin J C Palma

T Gutierrez L A Montenegro K 2016 The Astronomerrsquos Telegram8602

Schechter P L Mateo M Saha A 1993 PASP 105 1342Seaquist E R 2008 in Bode M Evans A eds Classical Novae 2nd edn

No 43 Cambridge Univ Press CambridgeShara M M 2002 in Shara M M ed ASP Conf Ser Vol 263 Stellar

Collisions Mergers and their Consequences Astron Soc Pac SanFrancisco p 1

Simon R Rathborne J M Shah R Y Jackson J M Chambers E T2006 ApJ 653 1325

Simpson R J et al 2012 MNRAS 424 2442Skrutskie M F et al 2006 AJ 131 1163Soker N Tylenda R 2006 MNRAS 373 733Strope R J Schaefer B E Henden A A 2010 AJ 140 34Sutherland W et al 2015 AampA 575 A25Tylenda R et al 2011 AampA 528 A114Tylenda R Soker N 2006 AampA 451 223Walter F M Battisti A Towers S E Bond H E Stringfellow G S

2012 PASP 124 1057Wang X Fields B D Lien A Y 2019 MNRAS 486 2910Warner B 1987 MNRAS 227 23Weiler K W Sramek R A 1988 ARAampA 26 295Williams G M Peretto N Avison A Duarte-Cabral A Fuller G A 2018

AampA 613 A11

APPENDIX A A SUBMM-BRIGHT VARIABLEY S O IN SD C G 3 3 1 0 6 2minus0 2 9 4

As noted in Section 34 the ALMA follow-up observations of VVV-WIT-01 did not detect the transient but a submmmm continuumsource was detected 45 arcsec to the east The submm flux was062 plusmn 004 mJy in band 6 (230 GHz) and 086 plusmn 006 mJy inband 7 (3365 GHz) It was not detected in band 3 (100 GHz) toa 3σ limit of 014 mJy The band 6 observations were obtainedon 2015 August 14 and September 4 and the band 7 observationswere obtained on 2015 September 25 The beam sizes were 05arcsec 03 arcsec and 025 arcsec in band 3 band 6 and band 7respectively The ALMA source is coincident (within 01 arcsec)

with the red VVV point source VVV J16105393-5155328 that isvisible about 5 arcsec east of VVV-WIT-01 in the middle panel ofFig 3 This red star has mean Ks = 1539 mag H minus Ks = 166J minus H gt 3 The VVVVVVX Ks light curve is shown in Fig A1 OurSpitzerIRAC data (see Section 31) give magnitudes I2 = 1239I1 minus I2 = 071

The infrared colours and fluxes are typical for a low mass classI or class II YSO (Megeath et al 2004 Gutermuth et al 2009) ata distance of a few kpc and in view of the projected location in anIRDC it is likely that VVV J16105393-5155328 is indeed a YSOThe Ks light curve clearly shows an infrared outburst of 09-magamplitude in the data taken on 2015 June 30 and July 13 (MJDs57203 and 57216) This further supports a YSO identification inview of the fact that high-amplitude variability is fairly commonin YSOs (Contreras Pena et al 2017) The infrared burst faded by008 mag over the 13 d between the two dates of observation Theinfrared flux was otherwise fairly constant from 2012 to 2018 andwas observed at the quiescent level on 2015 May 19 and 2016 July5 A slow 025-mag decline was seen from 2010 to 2011

The ALMA continuum detections can be attributed to emissionby dust in the cool outer regions of a circumstellar disc This wouldbe consistent with the non-detection in band 3 due to the declinein dust emissivity with increasing wavelength We can estimate thedust mass with the equation Mdust = (Fνd2)(κνBν(T)) (Hildebrand1983) Adopting T = 20 K and opacity κ = 10(ν1000)β GHzcm2 gminus1 and β = 1 (Beckwith et al 1990) yields a mass Mdust =5 times 10minus4(d3 kpc)2 M Adopting the conventional gas to dust ratioof 1001 this implies Mdisc = 005(d3 kpc)2 M Recalling that thedistance to SDC G331062minus0294 is likely to be between 31 and74 kpc this indicates that the YSO hosts a relatively massive discor perhaps an unusually warm one It is certainly possible that theinfrared outburst was ongoing at the time of the ALMA detectionsin 2015 August and September resulting in heating of the outer disc(Johnstone et al 2013)

Figure A1 VVV Ks light curve for the submm-bright sourceVVV J16105393-5155328 A jump of sim09 mag occurred in mid-2015lasting for at least 13 d The flux had returned to the pre-outburst level indata taken 12 months later The inset panel shows an expanded view of thedata during the outburst

This paper has been typeset from a TEXLATEX file prepared by the author

MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


MNRAS 492 4847ndash4857 (2020) doi101093mnrasstaa155Advance Access publication 2020 January 20

VVV-WIT-01 highly obscured classical nova or protostellar collision

P W Lucas 1lsaquo D Minniti234 A Kamble5 D L Kaplan6 N Cross7 I Dekany8

V D Ivanov9 R Kurtev210 R K Saito 11 L C Smith12 M Catelan 213

N Masetti314 I Toledo15 M Hempel3 M A Thompson1 C Contreras Pena16

J Forbrich1 M Krause 1 J Dale1 J Borissova 210 and J Emerson17

1Centre for Astrophysics University of Hertfordshire College Lane Hatfield AL10 9AB UK2Millennium Institute of Astrophysics Av Vicuna Mackenna 4860 782-0436 Macul Santiago Chile3Departamento de Ciencias Fısicas Universidad Andres Bello Fernandez Concha 700 Las Condes Santiago Chile4Vatican Observatory V00120 Vatican City State Italy5Harvard-Smithsonian Center for Astrophysics 60 Garden Street Cambridge MA 02138 USA6Center for Gravitation Cosmology and Astrophysics University of Wisconsin-Milwaukee PO Box 413 Milwaukee WI 53201 USA7Wide-Field Astronomy Unit Institute for Astronomy University of Edinburgh Royal Observatory Blackford Hill Edinburgh EH9 3HJ UK8Astronomisches Rechen-Institut Zentrum fur Astronomie der Universitat Heidelberg Monchhofstr 12-14 D-69120 Heidelberg Germany9European Southern Observatory Karl-Schwarszchild-Str 2 D-85748 Garching bei Muenchen Germany10Instituto de Fısica y Astronomıa Universidad de Valparaıso ave Gran Bretana 1111 Casilla 5030 Valparaıso Chile11Departamento de Fısica Universidade Federal de Santa Catarina Trindade 88040-900 Florianopolis SC Brazil12Institute of Astronomy University of Cambridge Madingley Road Cambridge CB3 0HA UK13Instituto de Astrofisica Pontificia Universidad Catolica de Chile Av Vicuna Mackenna 4860 7820436 Macul Santiago Chile14INAFndashOsservatorio di Astrofisica e Scienza dello Spazio via Gobetti 933 I-40129 Bologna Italy15ALMA Observatory Alonso de Cordova 3107 Vitacura Santiago Chile16School of Physics University of Exeter Stocker Road Exeter EX4 4QL UK17Astronomy Unit School of Physics and Astronomy Queen Mary University of London Mile End Road London E1 4NS UK

Accepted 2020 January 14 Received 2020 January 14 in original form 2019 November 5

ABSTRACTA search of the first Data Release of the VISTA Variables in the Via Lactea (VVV) Surveydiscovered the exceptionally red transient VVV-WIT-01 (H minus Ks = 52) It peaked beforeMarch 2010 then faded by sim95 mag over the following 2 yr The 16ndash22 μm spectral energydistribution in March 2010 was well fit by a highly obscured blackbody with T sim 1000 Kand AKs

sim 66 mag The source is projected against the Infrared Dark Cloud (IRDC) SDCG331062minus0294 The chance projection probability is small for any single event (p asymp 001ndash002) which suggests a physical association eg a collision between low mass protostarsHowever blackbody emission at T sim 1000 K is common in classical novae (especially COnovae) at the infrared peak in the light curve due to condensation of dust sim30ndash60 d after theexplosion Radio follow-up with the Australia Telescope Compact Array detected a fadingcontinuum source with properties consistent with a classical nova but probably inconsistentwith colliding protostars Considering all VVV transients that could have been projectedagainst a catalogued IRDC raises the probability of a chance association to p = 013ndash024After weighing several options it appears likely that VVV-WIT-01 was a classical nova eventlocated behind an IRDC

Key words stars formation ndash novae cataclysmic variables ndash ISM clouds ndash infrared stars

1 IN T RO D U C T I O N

Very high-amplitude near-infrared variable sources detected in theMilky Way are typically cataclysmic variables (eg Saito et al 20152016 De et al 2019) or Young Stellar Objects (YSOs) undergoingepisodic accretion variable extinction or both (eg Kospal et al

E-mail pwlucashertsacuk

2013 Contreras Pena et al 2017 Lucas et al 2017) The VVVsurvey (Minniti et al 2010 Saito et al 2012) is the first panoramicnear-infrared time domain survey of the Milky Way offering thechance to improve our knowledge of infrared variability and todiscover new types of variable star Studies of VVV transients mayassist extragalactic transient work where events such as LuminousRed Novae have uncertain physical origin (eg Kasliwal et al 2011Pastorello et al 2019) In the Milky Way the remnants of pastexplosive events can readily be detected providing examples of

Ccopy 2020 The Author(s)Published by Oxford University Press on behalf of the Royal Astronomical Society

Dow


icoupcomm













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4849








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4851



5564351 7918 0018 5592 0018 4764 0028 4295 01495696656 7908 0015 5669 0022 4753 0027 4117 01115828960 7929 0016 5685 0022 4746 0024 4066 01165961264 7922 0015 5775 0019 4747 0031 4115 01646027479 7915 0017 563 0016 4795 003 4211 01936093695 7934 0017 5603 0024 4743 0024 4025 01376159783 7922 0013 5667 0021 4761 0023 4157 01616225999 7975 0016 5672 0029 4696 0023 4137 01676292087 7945 0014 5621 0017 4745 0027 4153 0166424391 7974 0017 5546 0019 4722 0029 4113 0166424519 7939 0016 5575 0015 475 0033 4253 01656556823 7945 0015 5742 0028 4718 0023 414 00856689127 7942 0015 5627 0017 4698 0027 4196 01951257761 8078 0018 5805 0016 4806 0026 42 01661257773 8057 0021 5754 0021 4819 0026 4243 0171













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853














MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4849








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4851



5564351 7918 0018 5592 0018 4764 0028 4295 01495696656 7908 0015 5669 0022 4753 0027 4117 01115828960 7929 0016 5685 0022 4746 0024 4066 01165961264 7922 0015 5775 0019 4747 0031 4115 01646027479 7915 0017 563 0016 4795 003 4211 01936093695 7934 0017 5603 0024 4743 0024 4025 01376159783 7922 0013 5667 0021 4761 0023 4157 01616225999 7975 0016 5672 0029 4696 0023 4137 01676292087 7945 0014 5621 0017 4745 0027 4153 0166424391 7974 0017 5546 0019 4722 0029 4113 0166424519 7939 0016 5575 0015 475 0033 4253 01656556823 7945 0015 5742 0028 4718 0023 414 00856689127 7942 0015 5627 0017 4698 0027 4196 01951257761 8078 0018 5805 0016 4806 0026 42 01661257773 8057 0021 5754 0021 4819 0026 4243 0171













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853














MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4849








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4851



5564351 7918 0018 5592 0018 4764 0028 4295 01495696656 7908 0015 5669 0022 4753 0027 4117 01115828960 7929 0016 5685 0022 4746 0024 4066 01165961264 7922 0015 5775 0019 4747 0031 4115 01646027479 7915 0017 563 0016 4795 003 4211 01936093695 7934 0017 5603 0024 4743 0024 4025 01376159783 7922 0013 5667 0021 4761 0023 4157 01616225999 7975 0016 5672 0029 4696 0023 4137 01676292087 7945 0014 5621 0017 4745 0027 4153 0166424391 7974 0017 5546 0019 4722 0029 4113 0166424519 7939 0016 5575 0015 475 0033 4253 01656556823 7945 0015 5742 0028 4718 0023 414 00856689127 7942 0015 5627 0017 4698 0027 4196 01951257761 8078 0018 5805 0016 4806 0026 42 01661257773 8057 0021 5754 0021 4819 0026 4243 0171













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853














MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4851



5564351 7918 0018 5592 0018 4764 0028 4295 01495696656 7908 0015 5669 0022 4753 0027 4117 01115828960 7929 0016 5685 0022 4746 0024 4066 01165961264 7922 0015 5775 0019 4747 0031 4115 01646027479 7915 0017 563 0016 4795 003 4211 01936093695 7934 0017 5603 0024 4743 0024 4025 01376159783 7922 0013 5667 0021 4761 0023 4157 01616225999 7975 0016 5672 0029 4696 0023 4137 01676292087 7945 0014 5621 0017 4745 0027 4153 0166424391 7974 0017 5546 0019 4722 0029 4113 0166424519 7939 0016 5575 0015 475 0033 4253 01656556823 7945 0015 5742 0028 4718 0023 414 00856689127 7942 0015 5627 0017 4698 0027 4196 01951257761 8078 0018 5805 0016 4806 0026 42 01661257773 8057 0021 5754 0021 4819 0026 4243 0171













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853














MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4851



5564351 7918 0018 5592 0018 4764 0028 4295 01495696656 7908 0015 5669 0022 4753 0027 4117 01115828960 7929 0016 5685 0022 4746 0024 4066 01165961264 7922 0015 5775 0019 4747 0031 4115 01646027479 7915 0017 563 0016 4795 003 4211 01936093695 7934 0017 5603 0024 4743 0024 4025 01376159783 7922 0013 5667 0021 4761 0023 4157 01616225999 7975 0016 5672 0029 4696 0023 4137 01676292087 7945 0014 5621 0017 4745 0027 4153 0166424391 7974 0017 5546 0019 4722 0029 4113 0166424519 7939 0016 5575 0015 475 0033 4253 01656556823 7945 0015 5742 0028 4718 0023 414 00856689127 7942 0015 5627 0017 4698 0027 4196 01951257761 8078 0018 5805 0016 4806 0026 42 01661257773 8057 0021 5754 0021 4819 0026 4243 0171













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853














MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853














MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4853














MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm



















MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4855













MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm






RE FERENCES

























MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


VVV-WIT-01 4857















AampA 613 A11








MNRAS 492 4847ndash4857 (2020)

Dow


icoupcomm


doi handle monthly notices of the royal astronomical …

Documents