Table Of ContentDRAFTVERSIONJANUARY6,2017
PreprinttypesetusingLATEXstyleAASTeX6v.1.0
SPIRITS:UNCOVERINGUNUSUALINFRAREDTRANSIENTSWITHSPITZER
MANSIM.KASLIWAL1,JOHNBALLY2,FRANKMASCI3,ANNMARIECODY4,HOWARDE.BOND7,10,JACOBE.JENCSON1,
SAMAPORNTINYANONT1,YICAO1,CARLOSCONTRERAS5,DEVINA.DYKHOFF6,SAMUELAMODEO6,LEEARMUS3,MARTHA
BOYER8,22,MATTEOCANTIELLO9,ROBERTL.CARLON6,ALEXANDERC.CASS6,DAVIDCOOK1,DAVIDT.CORGAN6,JOSEPH
FAELLA6,ORID.FOX10,WAYNEGREEN2,ROBERTGEHRZ6,GEORGEHELOU3,ERICHSIAO11,JOELJOHANSSON12,RUBABM.
KHAN8,RYANM.LAU1,20,NORBERTLANGER13,EMILYLEVESQUE14,PETERMILNE15,SHAZRENEMOHAMED16,21,23,NIDIA
MORRELL5,ANDYMONSON7,ANNAMOORE1,ERANO.OFEK12,DONALO’SULLIVAN1,MUDUMBAPARTHASARTHY17,ANDRES
PEREZ1,DANIELA.PERLEY18,MARKPHILLIPS5,THOMASA.PRINCE1,DINESHSHENOY6,NATHANSMITH15,JASONSURACE19,
SCHUYLERD.VANDYK3,PATRICIAWHITELOCK16,21,ROBERTWILLIAMS10
7
1
0 1DivisionofPhysics,MathematicsandAstronomy,CaliforniaInstituteofTechnology,Pasadena,CA91125,USA
2 2
CenterforAstrophysicsandSpaceAstronomy,UniversityofColorado,389UCB,Boulder,CO80309,USA
n 3InfraredProcessingandAnalysisCenter,CaliforniaInstituteofTechnology,Pasadena,CA91125,USA
a 4
NASAAmesResearchCenter,MoffettField,CA94035,USA
J
5LasCampanasObservatory,CarnegieObservatories,Casilla601,LaSerena,Chile
4
6
MinnesotaInstituteforAstrophysics,SchoolofPhysicsandAstronomy,116ChurchStreet,S.E.,UniversityofMinnesota,Minneapolis,MN55455,USA
] 7Dept.ofAstronomy&Astrophysics,PennsylvaniaStateUniversity,UniversityPark,PA16802USA
E
8
NASAGoddardSpaceFlightCenter,MC665,8800GreenbeltRoad,Greenbelt,MD20771USA
H
9
KavliInstituteforTheoreticalPhysics,UniversityofCalifornia,SantaBarbara,CA93106,USA
.
h 10
SpaceTelescopeScienceInstitute,3700SanMartinDr.,Baltimore,MD21218USA
p
11
DepartmentofPhysics,FloridaStateUniversity,77ChieftainWay,Tallahassee,FL,32306,USA
-
o 12
BenoziyoCenterforAstrophysics,WeizmannInstituteofScience,76100Rehovot,Israel
r 13
t Argelander-InstitutfrAstronomieAufdemHgel71.D-53121Bonn,Germany
s
14
a DepartmentofAstronomy,UniversityofWashingtonSeattle,Box351580,WA98195-1580USA
[ 15
StewardObservatory,UniversityofArizona,Tuscon,AZ85721,USA
16
1 SAAO,POBox9,Observatory,7935,SouthAfrica
v 17IndianInstituteofAstrophysics,Koramangala,Bangalore560034,India
1 18
DarkCosmologyCentre,NielsBohrInstitute,JulianeMariesVej30,CopenhagenAF,DK-2100,Denmark
5
19
1 EurekaScientific,Inc.2452DelmerStreetSuite100Oakland,CA94602-3017USA
1 20JetPropulsionLaboratory,CaliforniaInstituteofTechnology,4800OakGroveDrive,Pasadena,CA91109,USA
0 21
AstronomyDepartment,UniversityofCapeTown,UniversityofCapeTown,7701,Rondebosch,SouthAfrica
.
1 22DepartmentofAstronomy,UniversityofMaryland,CollegePark,MD20742USA
0 23
NationalInstituteforTheoreticalPhysics,PrivateBagX1,Matieland,7602,SouthAfrica
7
1
: ABSTRACT
v
i We present an ongoing, systematic search for extragalactic infrared transients, dubbed SPIRITS — SPitzer
X
InfraRedIntensiveTransientsSurvey. Inthefirstyear, usingSpitzer/IRAC,wesearched190nearbygalaxies
r with cadence baselines of one month and six months. We discovered over 1958 variables and 43 transients.
a
Here,wedescribethesurveydesignandhighlight14unusualinfraredtransientswithnoopticalcounterparts
to deep limits, which we refer to as SPRITEs (eSPecially Red Intermediate Luminosity Transient Events).
SPRITEs are in the infrared luminosity gap between novae and supernovae, with [4.5] absolute magnitudes
between 11 and 14 (Vega-mag) and [3.6]-[4.5] colors between 0.3mag and 1.6mag. The photometric
− −
evolution of SPRITEs is diverse, ranging from <0.1 mag yr−1 to >7mag yr−1. SPRITEs occur in star-
forming galaxies. We present an in-depth study of one of them, SPIRITS14ajc in Messier83, which shows
shock-excitedmolecularhydrogenemission. Thisshockmayhavebeentriggeredbythedynamicdecayofa
non-hierarchicalsystemofmassivestarsthatledtoeithertheformationofabinaryoraproto-stellarmerger.
Keywords:surveys,infrared,supernovae,AGBandpost-AGB,novae,cataclysmicvariables,mass-loss
1. INTRODUCTION The systematic study of explosive transients and eruptive
variablesisgrowingbyleapsandboundsespeciallywiththe
2
advent of wide-field synoptic imaging. Recently, multiple i.e. 66% of the total B-band luminosity of the cluster. Fur-
newclassesofopticaltransients(e.g.,Kasliwal2012)andnew thermore, thesegalaxiestotal9.1 1011M i.e. 71%ofthe
(cid:12)
×
radiotransients(e.g.,Thorntonetal.2013)havebeenuncov- totalstellarmassofthecluster.
ered. Yet, the dynamic infrared (IR) sky is hitherto largely Tables summarizing properties of galaxies in this sample
unexplored. in detail is presented in Kasliwal et al. 2013. In 2015 and
Whiletheopticalisapowerfulbandtoexploresupernovae 2016, weadded star-forming regionsin galaxiestoo large to
and novae, it is blind to transients and eruptive variables map with IRAC otherwise (Kasliwal et al. 2014a). In 2017
thatareeitherself-obscuredorlocatedindustyregions(e.g., and 2018, we down-sized the sample to focus on the most
molecularclouds). IRfollow-upofopticallydiscoveredtran- luminousandmostmassivegalaxies(Kasliwaletal.2016).
sients shows that IR emission dominates in supernovae with In2014,eachofthesegalaxieswasimagedthreetimesby
circumstellar interaction, particularly at late-time (Fox et al. SPIRITS, with cadence baselines of 1 month and 6 months.
2011, 2013). We are now aware of at least two new classes In 2015 and 2016, additional shorter cadence baselines of 1
of explosive transients where the bulk of the emission is in weekand3weekswereadded. Archivaldataprovidesusad-
theIR–stellarmergers(associatedwithluminousrednovae, ditional multi-year baselines. A histogram of cadence base-
e.g. V1309 Sco; Tylenda et al. 2011) and electron capture linesisavailableinKasliwaletal.2014a.
supernovae(eCSNe;associatedwithintermediateluminosity Each SPIRITS pointing is seven dithered 100s exposures
redtransients,e.g.NGC300-OT;Bondetal.2009,SN2008S; ineachIRACfilter. Thelimitingmagnitude(asdefinedfora
Prietoetal.2008). 5σ point source Vega magnitude) in each SPIRITS epoch is
SomeeffortshavebeenundertakentolookforIRtransients 20mag at [3.6] and 19.1mag at [4.5]. This gives us a [3.6]
andvariableswiththeSpitzerSpaceTelescope(Werneretal. depth of up to 8.5mag at 5Mpc and up to 11.5mag at
− −
2004;Gehrzetal.2007). AblindsearchforIRtransientsin 20Mpc.
repeatedimagingoftheBootesfieldrevealedasuperluminous
2.2. Follow-UpGround-basedobservations
supernova(Kozłowskietal.2010). Searchestargetingnearby
starformingregionsintheMilkyWayhaveshownaplethora We are undertaking concomitant ground-based surveys to
ofyoungstarvariability(Codyetal.2014;Rebulletal.2014). monitortheSPIRITSgalaxysampleinthenear-IRandtheop-
Searchesforvariable,obscuredasymptoticgiantbranchstars ticalatroughlyamonthlycadence. AttheUniversityofMin-
innearbygalaxiesisbeingundertakenbytheDUSTiNGSsur- nesota’sMtLemmonObservingFacility(MLOF),weusethe
vey(Boyeretal.2015). Searchesforobscuredsupernovaein three-channelTwoMicronAllSkySurveycameras(Milligan
starburstgalaxies,usingSpitzer(Foxetal.2012),HST(Cresci et al. 1996; Skrutskie et al. 2006) mounted on the 1.52m IR
etal.2007)andhigh-resolutionground-basedadaptiveoptics telescope(Lowetal.2007). AtLasCampanas,weundertake
imaging(e.g.,Mattilaetal.2007),haverevealedafewcandi- near-IR monitoring with the Retrocam on Dupont 100-inch
dates(Kankareetal.2008,2012). telescopeandopticalmonitoringusingtheCCDontheSwope
Motivated thus, we began a systematic search for mid-IR 40-inchtelescope.AtPalomar,weusetheSamuelOschin48-
transients in nearby galaxies with Spitzer. Here, we present inch(primarilyr-band)andPalomar60-inchtelescopes(gri-
the experiment design ( 2), the software pipelines ( 3), the bands)foropticalmonitoring.UsingtheLCOGTnetwork,we
§ §
discoveries in the first year ( 4) and a case-study ( 5). We obtainadditionalopticalmonitoringingi-bands. Inaddition,
§ §
concludewithreflectionsonapossiblewayforwardtochart follow-upofdiscoveredtransientswasundertakenbyamyr-
thedynamicIRsky( 6). iadoffacilitiesincludingKeck,Magellan,Palomar200-inch,
§
SALTandRATIR.
2. EXPERIMENTDESIGN 2.3. Follow-UpwiththeHubbleSpaceTelescope
2.1. GalaxySample,CadenceandDepth
Following non-detections from the ground, we were able
The SPIRITS experiment uses the IRAC instrument (FoV to set even deeper magnitude limits for two transients based
5’ 5’; Fazio et al. 2004) aboard the warm Spitzer telescope on a small HST Director’s Discretionary program (GO/DD-
×
tosearchforIRtransientsat3.6µm([3.6])and4.5µm([4.5]). 13935, PI H. Bond). We imaged SPIRITS14aje (in M101)
Thisisasearchtargeting190nearbygalaxiesselectedusing and SPIRITS14axa (in M81) with the Wide Field Camera 3
three criteria: (i) 37 galaxies out to 5Mpc spanning diverse (WFC3)in2014September. IntheWFC3UVISchannel,we
galaxyenvironments: early-typegalaxies, late-typegalaxies, employed an “I” filter (F814W), and in the IR channel we
dwarf galaxies and giant galaxies (ii) 116 luminous galaxies used“J”(F110W)and“H”(F160W)bandpasses. Thesites
between5Mpcand15Mpc. Oursampleofgalaxiescaptures ofbothtargetshadbeenobservedwithHST beforetheirout-
atotalof2.0 1012 L intheB-bandwithin15Mpc. Thisis bursts, makingitpossibleforustocompareournewimages
(cid:12)
×
83%ofthetotalB-bandstarlightwithinthe<15Mpcvolume. withthearchivaldata. WeregisteredtheSpitzerframeswith
(iii)The37mostluminousandmostmassivegalaxiesinthe the HST frames by measuring the positions of isolated stars
Virgo Cluster (17Mpc). These galaxies total 1.6 1012L detectedinbothimages,allowingthesitesofthetransientsto
(cid:12)
×
3
be located in the HST images to precisions of typically 0(cid:48).(cid:48)1 determined by our transient detection routines. Sky back-
(MoredetailsinBondetal. inprep). ground is measured within an annulus from 8 to 16 pixels
surrounding each source and subtracted from the total flux.
3. SOFTWAREPIPELINES Finally, fluxes are converted to magnitudes using the Warm
3.1. ImageSubtractionandSourceCatalogs Spitzer/IRACzeropointsof18.8024(channel1)and18.3174
(channel2),alongwithaperturecorrectionsof1.21and1.22,
WeconstructreferenceimagesusingarchivalSpitzerimag-
respectively, asspecifiedbytheIRACinstrumenthandbook.
ing using supermosaics or S4G stacks (Sheth et al. 2010;
Sincethesubtractionimagesarenoisy,weconservativelyset
wheresupermosaicswereunavailable)orstackingprior“bcd”
thedetectionandupperlimitthresholdat9σ.
observations in the archive (where neither supermosaics nor
S4G stacks were available). We use the “maic” products of 3.3. DatabaseandDynamicWebPortal
theSpitzerIRACpipelineasourstartingpointfordifference
We architected a postgresql database to ingest the differ-
imaging. We adapted an image differencing and transient-
ence imaging products. The search for transients is under-
source extraction code developed for the Palomar Transient
taken via a dynamic web portal. In 2014, Spitzer data was
Factory(PTFIDE1)toSpitzerimaging. Thechangesmadeto
releasedeverytwoweeks(thetime-laghasnowbeenreduced
thissoftwarewere(i)abilitytooperateonco-addsofindivid-
to only a few days). Once the data is released, the SIDE
ual IRAC exposures; (ii) masking of co-add–image regions
pipelineispromptlyrun. Teammembersareassignedgalax-
with depths <5 exposures to mitigate cosmic rays and de-
ies to look through candidate metadata and postage stamps
tector glitches; (iii) execution of the SExtractor tool (Bertin
tovisuallyvetandflaginterestingtransients,typicallywithin
and Arnouts, 1996) to extract transient candidates from the
one day of data release. These transients are then assigned
difference images (as opposed to PSF-fitting for PTF); (iv)
a name in sequential order by the database. For example,
omissionofdynamicphotometric-gainmatchingbetweenref-
in the first year, a total of 131722 candidates (on 4396 new
erenceandscience-imageco-adds,and(v)amorestreamlined
and archival subtraction images) were automatically loaded
andsimplerPSF-matchingschemebetweenimages. Updates
into our subtraction database. Of these, only 1693 sources
(iv)and(v)takeadvantageofthestablethermalenvironment
were assigned names and flagged for further inspection as
oftheSpitzertelescope.Examplesofdiscoveryimagetriplets
transients or variable stars. Additional context information
areshowninFigure1.
fromvariousground-basedandspace-basedfacilitiesissum-
One may expect difference-imaging from space to yield
marized on a source-specific webpage for reference. Inter-
fewer false positives than from a ground observatory where
esting transients are announced via Astronomers Telegrams
atmosphericconditionscontinuouslymodifythePSFbetween
(e.g.,Kasliwaletal.2014b;Jencsonetal.2015,2016a,b).
observations, however, we found this generally not to be the
case with Spitzer (in part, due tothe sparselysampled PSF). 4. FIRSTTRANSIENTDISCOVERIES
TheSpitzerdifference-imagingispronetoalargenumberof
Inthefirstyear,SPIRITSdetectedover1958variablestars
false positives when the Spitzer field-of-view had a large ro-
and 43 IR transient sources. Of these 43 transients, 21 were
tationbetweenthereferenceandscienceimageepochs. The
knownsupernovaeand4wereintheluminosityrangeofclas-
Spitzer-IRAC PSF profiles follow the fixed detector/optical
sical novae. SPIRITS supernovae have been discussed in-
diffractionpatternsandthesecouldnotbeeasilymatchedand
depth as a Type Ia sample (Johansson et al. 2014), a core-
subtracted between rotated apparitions of the same piece of
collapse sample (Tinyanont et al. 2016), and a case-study
sky. Each of our candidates are visually vetted and we re-
of the peculiar low-velocity SN2014dt (Fox et al. 2016).
quire at least two detections (in two filters or at two epochs)
Four transients had optical counterparts: SPIRITS14axm in
to weed out false positives. But, we caution that our search
NGC2403 (luminous blue variable; van Dyk et al. in prep),
is incomplete and any inferred rates of transients are likely
SPIRITS14pzinNGC4490(stellarmergercandidate;Smith
lowerlimits.
et al. 2016), SPIRITS14bme in NGC300 (high mass X-ray
binary; Lau et al. 2016) and SPIRITS14bmc in NGC300
3.2. ForcedPhotometry
(Adams et al. in prep). Here, we discuss the remaining 14
We define a source to be transient if there is no detected
events (see Table 1), which are unusual IR transients in the
quiescent point source underneath the source location in the
luminosity gap between novae and supernovae with no opti-
referenceframe(else,thesourceisavariable).Toobtainmag-
calcounterparts,hereafterreferredtoasSPRITEs(eSPecially
nitudesfortransientsources,weperformforcedaperturepho-
RedIntermediate-luminosityTransientEvents).
tometryonthesubtractedimages,assumingzerofluxpresent
ThecommonpropertiesofthisnewclassofSPRITEsare:
in the reference image. We sum the flux in an aperture with
radius4-pixels(2(cid:48).(cid:48)4)centeredattheRAandDeccoordinates 1. Peak luminosity at [4.5] brighter than 11mag and
−
fainterthan 14mag(seeTable2).
1 http://web.ipac.caltech.edu/staff/fmasci/home/miscscience/ptfide- −
v4.0.pdf 2. IRAC Color [3.6]-[4.5] between 0.3mag and 1.6mag
4
(seeTable2). body(Cardellietal.1989;Chapmanetal.2009),wefindthat
anobserved[3.6]-[4.5]colorof1magrequiresvisualextinc-
3. No optical counterpart in concomitant or follow-up
tionof30magrespectively! Thisisdifficultespeciallygiven
imagingtoatleastr<20mag(seeTable4).
the location of SPRITEs in the middle-to-outer galaxy disks
(Figure2). Moreover,ifthisiscorrect,giventhattherewere
4. Occurring in star-forming galaxies even though the
sixnewopticalsupernovaeintheSPIRITSsamplein2014,it
SPIRITS galaxy sample has a mix of all galaxy types
wouldsuggestthatopticalsurveysaremissingafourthofthe
(seeFigure2)
supernovaeduetoobscuration.
First,weplaceSPRITEsincontextoftheIRtransientphase Next, thereareafewpossibletheoreticalmodelsthatmay
space. We plot the IR luminosity evolution of SPRITEs and explainSPRITEs. Forexample,coalescenceof1–30M(cid:12) bi-
comparetowell-knownnovaeandsupernovae(seetoppanel nariesisexpectedtocreatecopiousamountsofdustinanop-
ofFigure3). SPRITEsaretooluminoustobeclassicalnovae. tically thick wind launched during the stellar merger (Soker
Thehighestabsolutemagnitudethatcanbeachievedbydusty & Tylenda 2006; Ivanova et al. 2013; Nicholls et al. 2013;
classicalnovaeisapproximately 11magat[3.6]. Thismax- Pejcha et al. 2016). Another possibility is that these are
−
imum occurs when an optically thick dust shell forms about electron-captureinducedcollapseof8–10M(cid:12) extremeAGB
50–100daysafteroutburstwhilethecentralengineisstillnear starswheretheshockbreakoutdidnotdestroyallthedustsur-
Eddingtonluminosity(e.g. NQVul,LWSer;seeGehrzetal. roundingtheprogenitor(Kochanek2011). Yetanotherpossi-
1995). Atpeakdustproduction,thesenovaecanhavea[3.6]- bilityisthatweakshocksinfailedsupernovaethatformblack
[4.5] color of 1.5mag (see Ney & Hatfield 1978; Gehrz holesmayalsonotleadtobrightopticaltransientsbutrather
≈
etal.1980). Someveryfastnovaethatdonotformdust(e.g. IR transients as the ejection of large amounts of material at
V1500 Cyg; see Gallagher & Ney 1976; Ennis et al. 1977) low velocity may condense to form dust (Piro 2013; Love-
may also reach luminosities of up to 11mag for a few grove&Woosley2013). Weconsiderthelikelihoodofeach
≈ −
daysatoutburst. ofthesemodelsbelow.
Next, we characterize the light curves of SPRITEs. The The speed of evolution is diagnostic of the origin in that
SpitzerlightcurvedataarepresentedinTable5. Webroadly a terminal, explosive event (eCSNe, obscured SNe, failed
categorize SPRITEs into two relative speed classes (see Ta- SNe)wouldlikelybelongtothefastclassandastellarmerger
ble 2): (i) six SPIRITEs evolve slowly over many year wouldlikelybelongtotheslowclass. Stellarmergerofmas-
timescaleswithspeedsslowerthan0.5magyr−1 (Figure6), sive stars may result in a slowly evolving red remnant. As
(ii)eightSPRITEsevolvefasterthan0.5magyr−1(Figure4). wasseenwithSPIRITS14pzinNGC4490(Smithetal.2016)
Ofthese,fourevolveonfewmonthtimescales,fasterthan1.5 and the transient in M101 (Blagorodnova et al. 2016), the
magyr−1andfadebelowdetectabilityinlessthanoneyear. SPRITE luminosities are consistent with late-time observa-
But,thelightcurvecomparisonislimitedasthelightcurves tionsofastellarmerger. Theyoungstellarpopulationinthe
ofSPRITEsarediverse,infrequentlysampledandtheexplo- vicinityofthesetransientsalsosupportsthepresenceofmas-
sion times are poorly constrained. Therefore, we plot IR lu- sive stars. Five slow SPRITEs could be stellar mergers (We
minosity versus IR color at all epochs (bottom panel of Fig- exclude SPIRITS14bgq as it repeats – appears, disappears
ure3).WefindthatSPRITEsoccupyauniqueregioninphase andre-appears–andcouldeitherbeabackgroundAGNoran
spacebetweennovaeandsupernovaeonthisdynamicHRdi- extremeAGBvariable). Ifthishypothesisiscorrect,therates
agramofexplosiveIRtransients. WenotethattheIRcolors appear to be higher than that estimated by Kochanek et al.
ofSPRITEsareasredasthereddestnovae,core-collapsesu- 2014.
pernovaeandILRTs. Thecorrespondingeffectiveblackbody Now for the relatively fast events, the nature of the explo-
temperaturesofSPRITEsspan350Kto1000K. sioncanbedisentangledifthereisaprogenitoridentification.
The puzzling absence of optical emission from SPRITEs Distinguishing features of the eCSNe class are: (i) the de-
challengesasupernovainterpretation. Typicalsupernovaeare tection of an IR progenitor star with an absolute magnitude
brighter than 20mag at 20Mpc for many months in the vi- brighterthan 10magandcolorredderthan1mag(Thomp-
−
sual wavebands. Thus, no prior history of optical detection son et al. 2009; Kochanek 2011), and (ii) subsequently, a
(despite the intensive monitoring of nearby galaxies by sev- monotonicdeclineofthetransientemissionbelowtheprogen-
eralsynopticsurveysandamateursintheopticalwavebands) itorluminosity(Adamsetal.2016). OnlytwoSPRITEshave
suggests that SPRITEs are unlikely to be old supernovae. If detections in archival Spitzer imaging: SPIRITS14bgq and
theIRexplosiontimeisstronglyconstrainedandthelifetime SPIRITS14bsb. But neither is a promising eCSN as SPIR-
isshorterthanayear,itimpliesthatthepeakluminositywas ITS14bgqiseruptive(notexplosive)andSPIRITS14bsbap-
notmuchhigherthanobserved(and, thatthetransientisnot pears to be part of a massive star cluster. The remaining
old). We consider the hypothesis that two fast SPRITEs — SPRITEs have upper limits deep enough (see Table 2) to
SPIRITS14axa and SPIRITS14bay — are obscured super- ruleoutmostoftheprogenitorparameterspacedelineatedby
novae. Applyingasimpleextinctionlawtoa10,000Kblack- Thompsonetal.2009.
5
IfanHSTprogenitorisdetectedandconsistentwithamas- A K-band spectrum obtained with the MOSFIRE spec-
sive star, there are two possibilities: an eruption event like trometer (McLean et al. 2012) on the Keck I 10-meter tele-
eta Carinae or a terminal explosive event like the forma- scope on 8 June 2014 found five emission lines of molecu-
tion of a stellar mass black hole. Both would be consistent larhydrogen,andneitheranycontinuumnoranyotherlines.
with a young population. One way to distinguish these two Theobservedpropertiesofthesero-vibrationaltransitionsare
wouldbetocontinuemonitoringtoseeifthefadingismono- summarized in Table 3. The velocities are consistent with
tonic(hence,terminal)orifthereisanotherepisodeoferup- theCOradialvelocityatthepositionofSPIRITS14ajc. The
tion. But none of the SPRITEs with both pre-explosion and relativelineintensitiesofthefourv=1 0ro-vibrationaltran-
−
post-explosionarchivalHSTimaginghavecandidateprogen- sitions are consistent with shock-excitation at a temperature
itor counterparts. Specifically, our HST imaging of SPIR- around 1000 to 2,000 K and similar to the relative intensi-
ITS14aje in 2014 September shows only one faint star de- tiesinHerbig-HaroobjectssuchasHH211(O’Connelletal.
tected at the edge of a 3σ positional error circle in I-band. 2005). However, theintensityratioofthev=2 1to1 0vi-
− −
Butthisstardoesnotvarybetween2003and2014,isnotde- brationaltransitioncorrespondstoahighertemperature,pos-
tected in the J-band and H-band and likely unrelated to the siblyindicatingthatfluorescentpumpingintheultravioletLy-
transient. Similarly, our HST imaging in I-band of SPIR- manandWernerbandsmayplayaroleinexcitingthehigher
ITS14axa shows one star that brightened by 0.4 mag be- vibrationalstatesofH ,orthattheshockstructureintheemis-
2
∼
tween2002and2014,butitisconsistentwithanormalfield sionregionshasamorecomplextemperaturestructure.
red giant lying close to the tip of the red-giant branch and ThesiteofSPIRITS14ajcinM83wasfortuitouslyimaged
likelyunrelated. byHST withtheWFC3camerain2012(programGO-12513,
In summary, the IR photometric data alone is not suffi- PI W. Blair). The SPIRITS14ajc light curve suggests that
cient to distinguish between various models. Ground-based theeventwasunderwayin2012,althoughunfortunatelythere
spectroscopic data has been difficult to obtain because these were no Spitzer observations in this year. Figure 6 depicts
transients are either very faint or not detected in the near- the site of SPIRITS14ajc in the HST V, I, Hα+[N II], and
IR/optical wavelengths. Next, we present a detailed case- H bandpasses. Thesourceliesinaverycrowdedstellarfield
studyononetransientforwhichweseeshock-excitedemis- with only modest extinction. In the I-filter, there is a faint
sion lines in a near-IR spectrum and hence, infer a physical star(26.0mag,M 3.8)whichliesclosetothecenterof
I
(cid:39) −
origin. the error circle and a brighter star (25.0mag, M 2.8)
I
(cid:39) −
whichliesattheeasternedgeofthe3σ positionuncertainty.
5. SPIRITS14AJC:ATRANSIENTDRIVINGASHOCK The absolute magnitudes are consistent with both stars be-
INTOAMOLECULARCLOUD
ing normal field red giants. Only the brighter of these two
SPIRITS14ajc in M83 went into outburst in 2010 and stars is marginally detected at Vand neither star is detected
has stayed at a [3.6]-band luminosity of 11 mag and a inH-band. IntheabsenceofanyotherHST imagesinthese
−
[3.6] [4.5]colorof0.7magforthepastfouryears(Figure6). filters taken at different dates, we cannot definitely rule out
−
NoquiescentsourcewasdetectedinSpitzerimagestakenbe- that either of the I-band sources is the counterpart of SPIR-
tween2006and2008.Noopticalornear-IRcounterpartisde- ITS14ajc,butthelackofadetectionatH arguesagainstthis.
tectedinground-basedfollow-upin2014(seeTable4).Based The 5σ limiting magnitude (Vega-scale) for the H-band ex-
onthecoolSEDofthetransient,theeffectiveblackbodytem- posure is about 23.5. The Hα+[N II] image (third frame)
peratureisapproximately900K. showsnothingatthesite,butnearbyisasmall,faint,bubble-
SPIRITS14ajc is located in a spiral arm of M83 with in- like emission nebula just outside the error circle. Its diam-
tense star formation (Figure 2). The 2.6 mm 12CO J=1 0 eter is about 0(cid:48).(cid:48)6, or approximately 16 pc. To characterize
−
emissionatthepositionofSPIRITS14ajc,measuredwiththe the stellar and ISM environment, we show a color rendition
Nobeyama Millimeter Array (Hirota et al. 2014), has a peak of its wider surroundings. There are several young associa-
brightness temperature of 2.8 K, a velocity integrated flux tions, some containing red supergiants, as well as a network
I(CO) = 28.6 Kelvin km s−1 in the 6(cid:48)(cid:48) by 12(cid:48)(cid:48) synthesized ofdarkdustlanes,withinafewarcsecondsofthesite. How-
beam, andiscenteredatV =572 15kms−1. TheX- ever,14ajcisnotclearlyassociatedwitheithertheveryyoung
LSR
±
factormethod(Bolattoetal.2013)toconvertI(CO)intoH clustersordarkdustfeatures.
2
column density using a Solar-vicinity value for the X-factor WeconsidertwomodelsforSPIRITS14ajc: theexplosion
(2.0 1020 cm−2 K−1) gives N(H ) 5.7 1021 cm−2 per of a supernova immediately behind or inside a dense molec-
2
× ≈ ×
beamcorrespondingtoavisual-wavelength(V-band)extinc- ular cloud (Kasliwal et al. 2005), and the production of an
tionof6magifthecloudweredistributeduniformlyoverthe eruptiveprotostellaroutflowsimilartotheexplosionthatoc-
beam. The NMA beam-size corresponds to a physical scale curredinOrionabout500yearsago(Zapataetal.2009;Bally
of 128pc 279pc at the distance of M83, making it likely etal.2011,2015).
×
that the extinction is patchy and potentially higher along the
5.1. IsSPIRITS14ajcpoweredbyasupernova?
lineofsighttoSPIRITS14ajc.
6
Inthesupernovascenario,twomechanismsmaycontribute in the sky (Allen & Burton 1993; Kaifu et al. 2000; Zapata
to the IR signal. First, the flash produced by the explosion etal.2009;Ballyetal.2011,2015).Theradialvelocityofthe
can be reprocessed into the IR portion of the spectrum by brightestpartoftheH emissionexhibitsaline-widthofless
2
dustasthelightechopropagatesthroughthedenseinterstel- than about 70 km s−1, consistent with the H line-widths of
2
larmedium. Inthisscenario,foregrounddustcompletelyex- SPIRITS14ajc. Radialvelocitiesandpropermotionsofmore
tinguishes the visual to near-IR wavelength signature of the than300kms−1 areobservedinvisualandnear-IRspectral
event. Second,ifthesupernovaissufficientlyclosetoadense lines such as [OI], [SII], and [FeII]. The momentum and ki-
cloud, the impact of the forward shock can excite H emis- netic energy is at least 160 M km s−1 and 4 1046 ergs
2 (cid:12)
×
sionastheblast-waveslamsintoadensemolecularmedium. (Snelletal.1984)to4 1047 ergs(Kwan&Scoville1976).
×
There may be two components to this emission: collisional Zapata et al. (2009) presented a CO J = 2–1 interferometric
excitation of H in the swept-up, compressed, and acceler- study and found a dynamic age of about 500 years for the
2
ated, post-shock layer, and fluorescent excitation by the UV larger OMC1 outflow. The initial explosion energy required
radiation emitted by fast shocks. Shock radiation can ex- todrivetheobservedoutflowisbetween1047to1048ergs.
citebothH locatedaheadoftheshock,andsurvivingorre- High-velocity, runaway stars are common among massive
2
formedH intheswept-up,postshocklayer. Avisualextinc- stars(Hoogerwerfetal.2000;Gualandrisetal.2004). Radio-
2
tion of more than 15mag is required to hide the visual and frequencyastrometryhasshownthattworadio-emittingstars
near-IRcontinuumofasupernova. inOMC1,the10to15M Becklin-Neugebauer(BN)object,
(cid:12)
The ionizing radiation of the supernova progenitor should and radio source I, thought to have a mass of about 20 M
(cid:12)
have created a large ionized cavity in the surrounding ISM. (Goddietal.2011),havepropermotionsof25and14kms−1
However, there is no obvious HII region at the location of respectivelyawayfromaregionlessthan500AUindiameter
SPIRITS14ajc. It is possible that foreground dust obscures fromwhichtheywereejectedabout500yearsago(Rodr´ıguez
anycavityorHIIregion. Alternatively,ahigh-velocity,‘run- etal.2005;Go´mezetal.2008;Goddietal.2011).Thekinetic
away’OB-starwouldhaveonlycarvedasmallcavity. More energyinstellarmotionsisabout4 1047ergs. Thetotalen-
×
than30%ofmassiveOBstarsareejectedfromtheirbirthsites ergyoftheOMC1event, 1048 ergs,consistsofthekinetic
∼
with velocities greater than 20 km s−1, and more than 10% energy in the ejected stars, the current kinetic energy of the
withvelocitieslargerthan100kms−1(Gies&Bolton1986). outflow,andtheenergyradiatedawaybyshocksoverthelast
Twoejectionmechanismshavebeenidentified:dynamicalin- 500years(Ballyetal.2011).
teractions such as the re-arrangement of the non-hierachical Bally&Zinnecker(2005)proposedthattheOMC1explo-
multiplestarsintoahierarchicalconfigurationsuchasacom- sion was triggered by the collision and merging of forming
pact binary and ejected members. (Hoogerwerf et al. 2000, massive stars. Dynamic friction and Bondi-Hoyle accretion
2001;Gualandrisetal.2004),andthesupernovaexplosionof ontomassiveprotostellarcoreinsideadensecluster-forming
themostmassivememberofanOB-starbinarywhichresults clumpmayhaveleadtorapidmigrationofthemostmassive
in the ejection of the surviving member at its pre-supernova protostars to the center of the clump’s potential well where
Keplerian orbital speed. Alternatively, if the runaway O star they formed a non-hierarchial system of massive stars with
wereinared-supergiantphasepriortoitsdemiseasasuper- similar interstellar separations. Such systems are unstable;
nova as it entered a molecular cloud, it would have avoided interactionsledtotheformationofahierarchicalsystemcon-
the production of an ionized cavity. It is also possible that sistingofacompactbinaryandadistantthirdmember 500
∼
the supernova was a Type Ia which happened to drift into a yearsago. Inthisscenario,thefinal3-bodyencounterwould
molecularcloud. However,sucheventsmustberaresincethe haveresultedintheformationofacompact,AU-scalebinary,
volume filling factor of dense molecular clouds in a galaxy most likely source I, and the ejection of both radio source I
tendstobelessthan1%. andBNfromtheOMC1coreBallyetal.(2011).
Goddietal.2011usedN-bodysimulationstoshowthatthe
5.2. IsSPIRITS14ajcpoweredbyamassiveprotostellar
mostlikelyinitialconfigurationofmassivestarsinOMC1that
eruption?
ledtotheobservedcurrentconfigurationofejectedstarscon-
Inthesecondmodel,theSPIRITS14ajctransientmaytrace sistsofamassivebinarystarinteractingwithanothermassive
an explosion triggered by either a protostellar collision or a star. The interaction leads to a hardening of the binary and
violent dynamical interaction of several massive protostars consequentreleaseofgravitationalpotentialenergy.
(Davies et al. 2006; Dale & Davies 2006). Such an event Binarystarsarecommonamongmassivestarsandthemass
issuggestedtohaveoccurredintheOMC1cloudcoreinthe distribution is peaked at similar component masses. A mas-
OrionAmolecularcloudlocatedbehindtheOrionNebula,the sivebinarycanformfromthebarinstabilityinamassivedisk.
nearestregionofongoingmassivestarformation(D 414pc; Thismechanismtendstoproducecircularorbitsintheplane
≈
Menten et al. 2007). The OMC1 outflow consists of a spec- ofcirumbinaryand/orcircumstellardisks.Informingclusters
tacular, wide opening-angle, arcminute-scale (0.1 to 0.3 pc) with a high density of protostars, massive stars surrounded
outflowwhichisthebrightestsourceofnear-IRH2emission by massive disks also have a high probability of capturing
7
clusters members to form binaries (Moeckel & Bally 2006, 105L .
(cid:12)
2007b). The secondary in such a capture-formed binary is TheOrionOMC1explosionisnotuniqueamongstarform-
mostlikelytobeonahigh-eccentricity,high-inclinationorbit ingregions. Zapataetal.(2013)foundevidenceforapower-
withrespecttothespinaxisofthemassivestarsanditsdisk ful explosive outflow in the DR 21 complex in Cygnus. If
(Moeckel & Bally 2007a; Cunningham et al. 2009). Thus, suchdynamicinteractionsareresponsibleforthelargenum-
binary-binary or binary-single star interactions are likely in berofrunawayOstars,andbinariesamongmassivestars,the
denseproto-clusters. eventrateofOMC1-likeeventsoughttobecomparabletothe
The kinetic energy of the outflow and ejected stars came birth-rate of massive stars. We propose that SPIRITS14ajc
fromthereleaseofgravitationalbindingenergyofacompact may trace the IR flare produced by a dynamic interaction of
binary formed by the interaction of three or more stars. The forming massive stars that either led to the formation of or
total energy liberated by the formation or hardening of a bi- hardeningofacompact,AU-scalebinary,orpossiblyaproto-
narydependsonthestellarmasses,M andM ,andtheorbit stellarmerger.
1 2
semi-majoraxis, R, asE GM M /2R. Forstarsinthe
B 1 2
massrange10to100M an≈dbinaryseparation0.5<R<10 6. CONCLUSIONANDAWAYFORWARD
(cid:12)
AU EB rangesfrom 1047 to 1051 ergs. Assumingthat radio SPIRITS has discovered a class of unusual IR transients
sourceIconsistsofapairof10M(cid:12) starsandthattheenergy calledSPRITEs. SPRITEsareintheluminositygapbetween
requiredtoejectthestars,theoutflow,andtoaccountforra- novaeandsupernovae,withmid-IR[4.5]absolutemagnitudes
diative losses is E = 1048 ergs, the semi-major axis of the between 11 to 14mag (Vega) and [3.6]-[4.5] colors be-
− −
finalbinarymustbeR GM2/2E 0.9AU. tween 0.3mag and 1.6mag. The photometric evolution of
∼ ≈
Forming massive stars accreting at rates of 10−4 to SPRITEsisdiverse,rangingfrom<0.1magyr−1 to>7mag
∼ ∼
10−3 M(cid:12) yr−1 tend to have bloated, AU-scale photospheres yr−1. SPRITEs appear to represent diverse physical origins
and resemble red supergiants (Hosokawa & Omukai 2009). and each one merits an in-depth investigation to decipher its
RadiosourceIhasaphotospherictemperatureof 4000K, nature. Next, we address challenges encountered in the first
∼
consistentwithhigh-accretionmodels(Testietal.2010). Ifat yearoftheSPIRITSsurveyandeffortstoovercomethem.
least one star before the interaction were accreting at such a AchallengeindecipheringthenatureofSPRITEswasthe
highrate,the1048ergenergyrequirementoftheOMC1event sparsesamplingoftheirmid-IRlightcurves(duetothesmall
implies that any attempt to form an AU-scale binary would number of epochs) with poor constraints on their explosion
haveledtoaprotostellarcollisionandconsequentejectionof date(relativetoarchivalimagestakenmanyyearsago). Sub-
some of the bloated star’s photosphere before ejection from sequent transients discovered in later years of the SPIRITS
thecloudcore. surveyhavebothabettercadenceontheirlightcurve(1week
In the dynamical interaction model, the disruption of cir- and 3 week cadence baselines were added) and better con-
cumstellardisksbythefinalclose-in3-bodystellarencounter, straints on their explosion date (as the 2014 SPIRITS data
combined with the recoil of the larger-scale envelope power serveasareference).
theOMC1outflowandaluminousIRflare. Ejectedmaterial Anotherchallengeinthefirstyearwasthedifficultyofim-
slams into the surrounding cloud core and lower-density en- mediate ground-based follow-up due to the delay of avail-
velope with speeds comparable to the Kepler speed at their abilityofSpitzerdataandthemismatchbetweenSpitzerand
point of origin. Ejecta velocities are expected to range from ground-basedvisibility. Bothofthesehaveimprovedinlater
10sofkms−1formaterialoriginatingtensofAUtoover500 years with Spitzer introducing early data release for time-
kms−1forejectafromwithinafewtenthsofanAUofamas- critical programs and the additional SPIRITS cadence base-
sivestar. TheX-ray,UV,andvisuallightfromshockswillbe linesbeingscheduledpreferentiallyintimewindowsthatfa-
obscured and reprocessed by the surrounding cloud into the cilitateground-basedfollow-up.
IR. Powerful shocks can produce an IR flare with luminosi- Yet another challenge was identifying possible progenitor
tiesupto1051ergsforthemostmassivestarcollisions(Bally stars due to a combination of a dense stellar population in
&Zinnecker2005). these nearby galaxies and the coarse Spitzer resolution. In
The duration of the IR-flare is given by the shock cross- ordertoaddressthis,wenowhaveanongoingHubbleSpace
ing time of the part of the clump sufficiently dense to repro- TelescopeprogramtoimagetheseIRtransientswhiletheyare
cess the shock-energy into the IR. The kinetic energy of the stillactive.
outflow is converted by surrounding dust to the mid and far- Our study of SPIRITS14ajc highlights the importance of
IR where the clump density is sufficiently large. Taking a spectroscopyinsolvingthemysteryofthephysicalnatureof
clumpradius, Rclump 0.01 0.05pcandaninitialejecta thetransient.Specifically,detectingtheexcitedmolecularhy-
∼ −
velocity of Vejecta 500 km s−1 implies a crossing time drogenlinespointedtoashockdrivenbythedynamicaldecay
∼
tcross Rclump/Vejecta 20 to 100 years. Emission of ofanon-hierarchicalsystemofmassivestars. However,spec-
∼ ∼
2 1047ergsin20years,alowerboundontheearlyradiative troscopy has also been extremely challenging as these tran-
×
losses for an Orion-like event, implies a mean luminosity of sientsaretooredforground-basedinstruments,Spitzeristoo
8
warm now for mid-IR spectroscopy and SOFIA is not sen- Cresci,G.,Mannucci,F.,DellaValle,M.,&Maiolino,R.2007,A&A,462,
sitive enough for these faint transients. Spectroscopy with 927
Cunningham,N.J.,Moeckel,N.,&Bally,J.2009,ApJ,692,943
theJamesWebbSpaceTelescope(JWST)ofslowlyevolving
Dale,J.E.,&Davies,M.B.2006,MNRAS,366,1424
SPIRITStransientswouldshedlightontheirnature. Specif-
Davies,M.B.,Bate,M.R.,Bonnell,I.A.,Bailey,V.C.,&Tout,C.A.2006,
ically, the low resolution spectrometer could determine dust MNRAS,370,2038
mass,grainchemistry,iceabundanceandenergeticstodisen- Ennis,D.,Beckwith,S.,Gatley,I.,Matthews,K.,Becklin,E.E.,Elias,J.,
tangletheproposedorigins. Neugebauer,G.,&Willner,S.P.1977,ApJ,214,478
Fazio,G.G.,etal.2004,ApJS,154,10
Regardlessoftheopenquestionsontheirorigin,thediscov-
Fox,O.,Filippenko,A.,Skrutskie,M.,Arendt,R.,Cenko,B.,&VanDyk,
eryofSPRITEsrepresentinganewclass(es)ofIRtransients
S.2012,ASearchfortheMissingSupernovaeinUltraluminous,Star
ataratecomparable(orhigher)thansupernovaeisencourag- FormingGalaxies,SpitzerProposal
ingandmotivatesasynopticIRsearch.However,undertaking Fox,O.D.,Filippenko,A.V.,Skrutskie,M.F.,Silverman,J.M.,
Ganeshalingam,M.,Cenko,S.B.,&Clubb,K.I.2013,AJ,146,2
awide-fieldIRsearchfortransientsbeyondtargetingnearby
Fox,O.D.,etal.2011,ApJ,741,7
galaxiesrequiresovercomingtheformidablechallengeposed
—.2016,AstrophysicalJournalLetters,816,L13
bythenightskybrightnessanddetectorcost. Conceptswith Gallagher,J.S.,&Ney,E.P.1976,AstrophysicalJournalLetters,204,L35
alternative semiconductors (Sullivan et al. 2014) or creative Gehrz,R.D.,Hackwell,J.A.,Grasdalen,G.I.,Ney,E.P.,Neugebauer,G.,
optics on polar sites (Moore et al. in prep) are being inves- &Sellgren,K.1980,ApJ,239,570
Gehrz,R.D.,Jones,T.J.,Matthews,K.,Neugebauer,G.,Woodward,C.E.,
tigated. If WFIRST elects a suitable survey design and pri-
Hayward,T.L.,&Greenhouse,M.A.1995,AJ,110,325
oritizes near real-time transient identification, it could be a
Gehrz,R.D.,etal.2007,ReviewofScientificInstruments,78,011302
powerfulprobetodiscoverIRtransients. Gies,D.R.,&Bolton,C.T.1986,ApJS,61,419
Insummary,theSPIRITSdiscoveryofSPRITEsbodeswell Goddi,C.,Humphreys,E.M.L.,Greenhill,L.J.,Chandler,C.J.,&
Matthews,L.D.2011,ApJ,728,15
forfuturewide-fieldexplorationsofthedynamicIRsky.
Go´mez,L.,Rodr´ıguez,L.F.,Loinard,L.,Lizano,S.,Allen,C.,Poveda,A.,
&Menten,K.M.2008,ApJ,685,333
Gualandris,A.,PortegiesZwart,S.,&Eggleton,P.P.2004,MNRAS,350,
WethankO.Pejcha,E.Lovegrove,S.Woosley,A.L.Piro, 615
E. S. Phinney, S. R. Kulkarni, L. Bildsten and E. Quataert Hirota,A.,etal.2014,PASJ,66,46
Hoogerwerf,R.,deBruijne,J.H.J.,&deZeeuw,P.T.2000,Astrophysical
forvaluablediscussions. Thisworkisbasedonobservations
JournalLetters,544,L133
madewiththeSpitzerSpaceTelescope,whichisoperatedby
—.2001,A&A,365,49
the Jet Propulsion Laboratory, California Institute of Tech- Hosokawa,T.,&Omukai,K.2009,ApJ,691,823
nology under a contract with NASA. The SPIRITS team ac- Ivanova,N.,Justham,S.,AvendanoNandez,J.L.,&Lombardi,J.C.2013,
knowledgesgeneroussupportfromtheNASASpitzergrants Science,339,433
Jacobs,B.A.,Rizzi,L.,Tully,R.B.,Shaya,E.J.,Makarov,D.I.,&
forSPIRITS.MMKthankstheNationalScienceFoundation
Makarova,L.2009,AJ,138,332
foraPIREGrantNo. 1545949fortheGROWTHproject. JJ
Jang,I.S.,Lim,S.,Park,H.S.,&Lee,M.G.2012,AstrophysicalJournal
acknowledgestheNationalScienceFoundationGraduateRe- Letters,751,L19
searchFellowshipunderGrantNo. DGE-1144469. PAWand Jencson,J.E.,etal.2015,TheAstronomer’sTelegram,7929
—.2016a,TheAstronomer’sTelegram,8688
SMaregratefultotheSouthAfricanNationalResearchFoun-
—.2016b,TheAstronomer’sTelegram,8940
dation(NRF)foraresearchgrant.RDGandtheMLOFgroup
Johansson,J.,etal.2014,ArXive-prints
weresupported,inpart,bytheUnitedStatedAirForce. Kaifu,N.,etal.2000,PASJ,52,1
Kankare,E.,etal.2008,AstrophysicalJournalLetters,689,L97
REFERENCES —.2012,AstrophysicalJournalLetters,744,L19
Kashi,A.,&Soker,N.2016,ResearchinAstronomyandAstrophysics,16,
014
Kasliwal,M.,etal.2013,SPIRITS:SPitzerInfraRedIntensiveTransients
Adams,S.M.,Kochanek,C.S.,Prieto,J.L.,Dai,X.,Shappee,B.J.,&
Stanek,K.Z.2016,MNRAS,460,1645 Survey,SpitzerProposal
Allen,D.A.,&Burton,M.G.1993,Nature,363,54 —.2014a,SPIRITS:SPitzerInfraRedIntensiveTransientsSurvey,Spitzer
Bally,J.,Cunningham,N.J.,Moeckel,N.,Burton,M.G.,Smith,N.,Frank, Proposal
A.,&Nordlund,A.2011,ApJ,727,113 Kasliwal,M.,Lau,R.,Cao,Y.,etal.2016,SpitzerProposal,13053
Bally,J.,Ginsburg,A.,Silvia,D.,&Youngblood,A.2015,A&A,579,A130 Kasliwal,M.M.2012,PublicationsoftheAstronomicalSocietyof
Bally,J.,&Zinnecker,H.2005,AJ,129,2281 Australia,29,482
Blagorodnova,N.,Kotak,R.,Polshaw,J.,etal.2016,arXiv:1607.08248 Kasliwal,M.M.,Lovelace,R.V.E.,&Houck,J.R.2005,ApJ,630,875
Bolatto,A.D.,Wolfire,M.,&Leroy,A.K.2013,ARA&A,51,207 Kasliwal,M.M.,etal.2014b,TheAstronomer’sTelegram,6644
Bond,H.E.,Bedin,L.R.,Bonanos,A.Z.,Humphreys,R.M.,Monard, Kochanek,C.S.2011,ApJ,741,37
L.A.G.B.,Prieto,J.L.,&Walter,F.M.2009,AstrophysicalJournal Kochanek,C.S.,Adams,S.M.,&Belczynski,K.2014,MNRAS,443,1319
Letters,695,L154 Kozłowski,S.,etal.2010,ApJ,722,1624
Boyer,M.L.,etal.2015,ApJS,216,10 Kwan,J.,&Scoville,N.1976,AstrophysicalJournalLetters,210,L39
Cardelli,J.A.,Clayton,G.C.,&Mathis,J.S.1989,ApJ,345,245 Lau,R.M.,etal.2016,ArXive-prints
Chapman,N.L.,Mundy,L.G.,Lai,S.-P.,&Evans,II,N.J.2009,ApJ,690, Lovegrove,E.,&Woosley,S.E.2013,ApJ,769,109
496 Low,F.J.,Rieke,G.H.,&Gehrz,R.D.2007,ARA&A,45,43
Cody,A.M.,etal.2014,AJ,147,82 Mattila,S.,etal.2007,AstrophysicalJournalLetters,659,L9
9
Table1. SPRITEs: Co-ordinatesandHostGalaxies
Name RA(J2000) DEC(J2000) HostGalaxy DistanceModulus
SPIRITS14qk 09h55m28.72s +69◦39(cid:48)58.6(cid:48)(cid:48) MESSIER82 27.73(Jacobsetal.2009)
SPIRITS14afv 12h50m49.56s +41◦05(cid:48)52.7(cid:48)(cid:48) MESSIER94 28.31(Tullyetal.2013)
SPIRITS14agd 13h05m30.87s −49◦26(cid:48)50.8(cid:48)(cid:48) NGC4945 27.90(Mould&Sakai2008)
SPIRITS14ajc 13h36m52.95s −29◦52(cid:48)16.1(cid:48)(cid:48) MESSIER83 28.41(Radburn-Smithetal.2011)
SPIRITS14ajd 13h37m05.02s −29◦48(cid:48)56.2(cid:48)(cid:48) MESSIER83 28.41(Radburn-Smithetal.2011)
SPIRITS14aje 14h02m55.51s +54◦23(cid:48)18.5(cid:48)(cid:48) MESSIER101 29.34(Tikhonovetal.2015)
SPIRITS14ajp 13h37m12.71s −29◦49(cid:48)14.9(cid:48)(cid:48) MESSIER83 28.41(Radburn-Smithetal.2011)
SPIRITS14ajr 13h36m54.81s −29◦52(cid:48)33.7(cid:48)(cid:48) MESSIER83 28.41(Radburn-Smithetal.2011)
SPIRITS14ave 03h47m03.17s +68◦09(cid:48)05.3(cid:48)(cid:48) IC342 27.58(Wuetal.2014)
SPIRITS14axa 09h56m01.52s +69◦03(cid:48)12.5(cid:48)(cid:48) MESSIER81 27.80(Jangetal.2012)
SPIRITS14axb 07h36m34.70s +65◦39(cid:48)22.4(cid:48)(cid:48) NGC2403 27.51(Radburn-Smithetal.2011)
SPIRITS14bay 12h56m43.25s +21◦42(cid:48)25.7(cid:48)(cid:48) MESSIER64 29.37(Tullyetal.2013)
SPIRITS14bgq 13h39m50.99s −31◦38(cid:48)46.0(cid:48)(cid:48) NGC5253 27.76(Mould&Sakai2008)
SPIRITS14bsb 01h35m06.72s −41◦26(cid:48)13.5(cid:48)(cid:48) NGC625 28.12(Jacobsetal.2009)
Table2. LightCurvePropertiesofSPRITEs
Name PeakTime PeakMag PeakTime PeakMag ColorPeakTime ColorPeakMag Progenitorlimit Progenitorlimit Lifespan Evolution Lifespan Evolution Speed
[3.6] [3.6] [4.5] [4.5] [3.6]-[4.5] [3.6]-[4.5] [3.6] [4.5] [3.6] [3.6] [4.5] [4.5]
MJD Mag MJD Mag MJD Mag Mag Mag yr magyr−1 yr magyr−1
SPIRITS14ave 57007.4 −11.2 57007.4 −11.8 57182.2 0.3 −10.6 −10.8 1.0 0.1 1.0 0.5 Slow
SPIRITS14afv 56722.7 −11.4 56754.9 −12.1 57250.7 1.2 −10.6 −10.9 1.4 0.9 2.1 0.5 Fast
SPIRITS14ajp 56787.2 −11.3 55290.7 −11.9 56915.5 0.7 −10.0 −10.3 5.6 0.2 4.4 <0.1 Slow
SPIRITS14axb 56827.4 −11.3 56827.4 −11.9 56827.4 0.6 −7.90 −9.37 0.1 −32.2 0.4 −6.2 Fast
SPIRITS14bay 56889.8 −11.4 56889.8 −12.6 56889.8 1.2 −11.4 −11.5 ··· ··· ··· ··· Fast
SPIRITS14ajr 56915.5 −11.3 56787.2 −12.2 56915.5 0.8 −11.1 −11.5 4.4 −0.1 5.0 0.1 Slow
SPIRITS14axa 56821.9 −11.1 56821.9 −11.6 56821.9 0.5 −9.19 −9.56 ··· ··· ··· ··· Fast
SPIRITS14agd 56764.5 −11.5 56544.9 −13.6 56787.1 0.9 −11.3 −11.7 1.1 −0.1 1.6 1.5 Fast
SPIRITS14bgq 55424.4 −10.8 55424.4 −11.6 57312.2 0.8 ··· ··· 5.6 <0.1 5.6 <0.1 Slow
SPIRITS14qk 56831.7 −10.9 54430.7 −11.2 56846.3 0.4 −10.6 −10.7 0.7 0.9 7.1 <0.1 Fast
SPIRITS14ajc 56765.1 −11.3 56765.1 −12.2 57161.2 0.5 −10.6 −11.4 5.1 <0.1 5.1 0.1 Slow
SPIRITS14aje 56742.8 −11.9 56742.8 −13.7 56771.8 1.6 −10.5 −11.6 0.4 2.4 0.1 7.3 Fast
SPIRITS14ajd 55290.7 −11.2 55290.7 −12.2 56915.5 1.0 −10.3 −10.9 4.4 0.1 5.1 0.3 Slow
SPIRITS14bsb 57078.5 −12.5 57085.0 −13.3 57446.6 1.5 ··· ··· 1.9 0.6 1.9 <0.1 Fast
Table3. SPIRITS14ajcEmissionLines
Transition Wavelength Flux GFWHM RelativeVelocity FluxRatio FluxRatio FluxRatio FluxRatio FluxRatio
(Observed) if1000K if2000K if3000K if4000K
1-0S(0) 22275.1 11.43 8.96 28 0.28 0.27 0.21 0.19 0.19
1-0S(1) 21258.9 40.13 6.74 65 1.00 1.00 1.00 1.00 1.00
2-1S(3) 20774.4 8.013 9.74 57 0.20 0.003 0.084 0.27 0.47
1-0S(2) 20376.4 10.61 6.04 53 0.26 0.27 0.37 0.42 0.44
1-0S(3) 19613.6 25.67 6.71 63 0.64 0.51 1.02 1.29 1.45
McLean,I.S.,etal.2012,inProc.SPIE,Vol.8446,Ground-basedand Ney,E.P.,&Hatfield,B.F.1978,AstrophysicalJournalLetters,219,L111
AirborneInstrumentationforAstronomyIV,84460J Nicholls,C.P.,etal.2013,MNRAS,431,L33
Menten,K.M.,Reid,M.J.,Forbrich,J.,&Brunthaler,A.2007,A&A,474, O’Connell,B.,Smith,M.D.,Froebrich,D.,Davis,C.J.,&Eislo¨ffel,J.
515
2005,A&A,431,223
Milligan,S.,Cranton,B.W.,&Skrutskie,M.F.1996,inProc.SPIE,Vol.
Pejcha,O.,Metzger,B.D.,&Tomida,K.2016,MNRAS,455,4351
2863,CurrentDevelopmentsinOpticalDesignandEngineeringVI,ed.
Piro,A.L.2013,AstrophysicalJournalLetters,768,L14
R.E.Fischer&W.J.Smith,2–13
Prieto,J.L.,Kistler,M.D.,Thompson,T.A.,etal.2008,Astrophysical
Moeckel,N.,&Bally,J.2006,ApJ,653,437
—.2007a,AstrophysicalJournalLetters,661,L183 JournalLetters,681,L9
—.2007b,ApJ,656,275 Radburn-Smith,D.J.,etal.2011,ApJS,195,18
Mould,J.,&Sakai,S.2008,AstrophysicalJournalLetters,686,L75 Rebull,L.M.,etal.2014,AJ,148,92
10
Table4. Follow-upPhotometry
Name ObservationDate Telescope Instrument Filter Photometry
SPIRITS14qk 2014stack(N=74) P48 iPTF R >23.5
SPIRITS14afv 2014stack(N=117) P48 iPTF R >20.1
2014-06-19 MLOF 2MASS J >17.4
2014-06-19 MLOF 2MASS H >17.3
2014-06-19 MLOF 2MASS Ks >16.2
SPIRITS14agd 2014stack(N=16) Swope CCD g >21.7
2014stack(N=16) Swope CCD r >21.2
2014stack(N=16) Swope CCD i >21.0
SPIRITS14ajc 2014-04-20 Swope CCD g >20.0
2014-04-20 Swope CCD r >20.0
2014-04-20 Swope CCD i >19.8
2014-05-18 duPont Retrocam J >19.8
2014-05-18 duPont Retrocam H >18.7
2014-07-02 Keck DEIMOS I >23.5
2014-06-07 Keck MOSFIRE Ks >18.7
2012-09-03 HST WFC3 H >22.0
2012-07-22 HST WFC3 I >25.5
SPIRITS14ajd 2014-04-20 Swope CCD g >20.0
2014-04-20 Swope CCD r >20.0
2014-04-20 Swope CCD i >19.8
2014-06-07 Keck MOSFIRE Ks >19.6
2014-05-24 MLOF 2MASS J >17.3
2014-05-24 MLOF 2MASS H >16.5
2014-05-24 MLOF 2MASS Ks >16.4
SPIRITS14aje 2014stack(N=157) P48 iPTF R >23.9
2014-05-01 MLOF 2MASS J >18.3
2014-05-01 MLOF 2MASS H >16.8
2014-05-01 MLOF 2MASS Ks >16.4
2014-07-02 Keck DEIMOS I >24.3
2014-06-07 Keck MOSFIRE Ks >19.4
2014-06-07 Keck MOSFIRE J >20.5
2014-09-23 HST WFC3 I >26.5
2014-09-23 HST WFC3 J >25.0
2014-09-23 HST WFC3 H >22.5
SPIRITS14ajp 2014-04-20 Swope CCD g >20.0
2014-04-20 Swope CCD r >20.0
2014-04-20 Swope CCD i >19.8
2014-06-07 Keck MOSFIRE Ks >19.2
SPIRITS14ajr 2014-04-20 Swope CCD g >20.0
2014-04-20 Swope CCD r >20.0
2014-04-20 Swope CCD i >19.8
2014-06-07 Keck MOSFIRE Ks >19.2
2014-05-24 MLOF 2MASS J >17.3
2014-05-24 MLOF 2MASS H >16.5
2014-05-24 MLOF 2MASS Ks >16.4
SPIRITS14ave 2014-09-23 LCOGT-1m SBIG i >20.0
2014-10-26 LCOGT-1m SBIG i >20.3
SPIRITS14axa 2014stack(N=117) P48 iPTF R >23.15
2014-09-26 HST WFC3 H >22.5
SPIRITS14axb 2014stack(N=18) P48 iPTF R >22.5
2014-05-01 MLOF 2MASS J >14.2
2014-05-01 MLOF 2MASS H >14.0
2014-05-01 MLOF 2MASS Ks >13.3
SPIRITS14bay 2014stack(N=66) P48 iPTF R >23.3
SPIRITS14bgq 2014stack(N=12) Swope CCD g >21.9
2014stack(N=12) Swope CCD r >20.9
2014stack(N=10) Swope CCD i >21.0
2014-04-30 MLOF 2MASS J >16.9
2014-04-30 MLOF 2MASS H >15.6
2014-04-30 MLOF 2MASS Ks >15.7
SPIRITS14bsb 2014stack(N=13) Swope CCD g >22.9
2014stack(N=13) Swope CCD r >22.6
2014stack(N=13) Swope CCD i >22.1
NOTE—ThisTableispublishedinitsentirety(includingindividualepochalobservations)inthemachine-readableformat.Upperlimitsare5σ.
