Probing the Final-stage Progenitor Evolution for Type IIP Supernova 2017eaw in NGC 6946
Liming Rui, Xiaofeng Wang, Jun Mo, Danfeng Xiang, Jujia Zhang, Justyn R. Maund, Avishy Gal-Yam, Lifan Wang, Tianmeng Zhang
aa r X i v : . [ a s t r o - ph . H E ] F e b MNRAS , 1–11 (XXX) Preprint 19 February 2019 Compiled using MNRAS L A TEX style file v3.0
Probing the Final-stage Progenitor Evolution for TypeIIP Supernova 2017eaw in NGC 6946
Liming Rui, Xiaofeng Wang, Jun Mo, Danfeng Xiang, Jujia Zhang, , Justyn R. Maund, Avishy Gal-Yam, Lifan Wang, Tianmeng Zhang Physics Department and Tsinghua Center for Astrophysics, Tsinghua University, Beijing, 100084, China Yunnan Observatories, Chinese Academy of Sciences, Kunming 650011, China Key Laboratory for the Structure and Evolution of Celestial Objects, Chinese Academy of Sciences, Kunming 650216, China Department of Physics and Astronomy, University of Sheffield, Hicks Building, Hounsfield Road, Sheffield S3 7RH, UK Department of Particle Physics and Astrophysics, Weizmann Institute of Sciencem Rehovot 76100, Israel Physics and Astronomy Department, Texas A&M University, College Station, TX 77843, USA Key Laboratory of Optical Astronomy, National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012,China
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We presented a detailed analysis of progenitor properties of type IIP supernova2017eaw in NGC 6946, based on the pre-explosion images and early-time obser-vations obtained immediately after the explosion. An unusually red star, withM F W = − F W − m F W =2.9 ± Hubble Space Telescope (HST) images takenin 2016. The observed spectral energy distribution of this star, covering the wave-length of 0.6-2.0 µ m , matches that of an M4-type red supergiant (RSG) witha temperature of about 3550 K. These results suggest that SN 2017eaw has aRSG progenitor with an initial mass of 12 ± ⊙ . The absolute F814W-bandmagnitude of this progenitor star is found to evolve from − − α emission feature blueshifted by ∼
160 km s − . This narrow componentdisappeared in the spectrum taken two days later, suggesting the presence of acircumstellar material (CSM) shell (i.e., at a distance of < × cm). Com-bining the inferred distance with the expansion velocity of the CSM, we suggestthat the progenitor of SN 2017eaw should have experienced violent mass loss atabout 1-2 years prior to explosion, perhaps invoked by pulsational envelop ejec-tion. This mechanism may help explain its luminosity decline in 2016 as well asthe lack of detections of RSGs with initial mass in the range of 17 M ⊙ < M < ⊙ as progenitors of SNe IIP. Key words: stars: evolution – supernovae: general – supernovae: individual: SN2017eaw – galaxies: individual: NGC 6946
Massive stars with initial masses ranging from 8 to 25M ⊙ usually end their lives as type II core-collapse su-pernovae (Heger et al. 2003). Among them, type IIP su-pernovae (SNe IIP) are the dominant subclass, whichare characterized by prominent Balmer lines in the spec-tra and a plateau phase lasting for about 100 days inthe light curves (Barbon et al. 1979; Filippenko 1997;Branch & Wheeler 2017). These hydrogen-rich SNe areidentified to arise from core-collapse of red supergiants(RSGs) with initial mass lying between 8 − ⊙ (Li et al. 2007; Maund et al. 2008; Smartt 2009, and refer-ence therein). However, the observation of RSGs in MilkyWay, Large Magellanic Cloud(LMC) and Small Magel-lanic Cloud(SMC) indicate a mass range of RSGs from 9 − ⊙ (Levesque et al. 2005, 2006, 2007). The lack ofRSGs with mass in the range of 17 −
25 M ⊙ as progenitorsof SNe IIP challenges current theory of stellar evolution.One possible explanation is that some massive RSGs ex-perience prominent mass loss during its final stage evolu-tion towards the explosion, due to the stellar wind, binaryinteraction, or pulsational eruption. The accumulated cir-cumstellar materials (CSM) may obscure the RSGs andlead to the underestimation of the progenitor luminos-ity and its initial mass. The condensed dust shell willleave an imprint on the SN spectra via photoionizationor interaction with the SN ejecta (e.g. Yaron et al. 2017;Hosseinzadeh et al. 2018).Signatures for the presence of nearby CSMs havebeen reported for SNe IIP/IIL/IIb. Early spectra of type c (cid:13) XXX The Authors
Rui LM et al.
IIb SN 2013cu (Gal-Yam et al. 2014) and type IIP SN2013fs (Yaron et al. 2017) reveal flash-ionized emissionlines from dense circumstellar wind. The narrow emis-sion lines are also detected in the early-time spectra takenin few hours or days after explosion, i.e., SN 2016bkv(Hosseinzadeh et al. 2018), SN 2006bp (Quimby et al.2007) and some young type IIP/IIL SNe reported byKhazov et al. (2016). The narrow interacting featurescould be generated by recombination of the CSMs ionizedby ultraviolet (UV) radiation emitted during the shockbreakout or the shock-cooling phase from the supernovae.However, the mechanism driving the mass loss of RSGs,i.e., whether the mass loss is stable or unstable, is notfully understood. Dense CSM surrounding a typical typeIIP supernova is found to have a expansion velocity thatis apparently higher than typical wind velocity of RSGs(Yaron et al. 2017) (i.e., 130 km s − versus 10 km s − ),suggestive of eruptive mass loss during the late phase evo-lution of RSGs. Moreover, Yoon & Cantiello (2010) foundthat strong pulsation induced by partial ionization of hy-drogen in the envelope is expected to enhance the massloss during the RSG phase.SN 2017eaw provides another rare opportunity tostudy the late-time evolution of the red supergiant withrelatively low initial mass. This SN was discovered byPatrick Wiggins on 2017 May 14.238 (UT dates) inNGC 6946 at a distance of 5.5 Mpc, with an unfilteredCCD magnitude of 12.8 mag (Wiggins 2017), and it wasimmediately classified as a young type IIP supernova(Xiang et al. 2017b). Note that NGC 6946 is a well-knownnearby Scd galaxy that has recorded nine SNe before SN2017eaw, which makes it one of the most prolific SN fac-tories known in the local universe. Follow-up observationsof SN 2017eaw were thus initiated less than 0.6 d af-ter the discovery, including multi-color ultraviolet/opticalphotometry and rapid high-cadence spectroscopy. TheUBVRI-band photometric evolution covering the first 200days has been presented by Tsvetkov et al. (2017), whichshows a long plateau characterized by a normal typeIIP SN. Kilpatrick & Foley (2018) studied the progeni-tor properties of SN 2017eaw, suggesting that it has a redsupergiant progenitor with a 13 M ⊙ . Based on the his-torical Spitzer data, their studies also suggested that theprogenitor star might have experienced dust-shell forma-tion a few years before the explosion.In this paper, we also attempt to constrain the pro-genitor properties of type IIP supernova 2017eaw usingphotometric and spectroscopic observations obtained im-mediately after explosion and the archival images fromHST and Spitzer. The description of early-time observa-tion and brief data reduction process are presented in Sec-tion 2. In Section 3, we describe the identification of theprogenitor candidate from HST and Spitzer in detail. Wediscuss the progenitor properties and explore its possiblefinal-stage evolution before explosion in Section 4. Ourconclusions are given in Section 5. The UBVRI photometry observations of SN 2017eaw weretriggered immediately after the SN discovery using theTsinghua-NAOC 80-cm telescope(TNT) (Huang et al.2012) at Xinglong Observatory of NAOC (Located inHebei, China), with the first observations obtained att ∼ uvw2, uvm2, uvw1 ) and three opticalfilters (U, B, V), starting at t ∼ , this includes bias subtraction and flatfield correction. The instrumental magnitudes were mea-sured with point-spread function (PSF) photometry andphotometric calibration was determined with 10 stars inthe field of SN 2017eaw observed on photometric night.We adopted aperture photometry to reduce all Swift im-ages using HEASoft(the High Energy Astrophysics Soft-ware) and detailed data descriptions are addressed in Ruiet al. (2019; in prep).Figure 1 displays the evolution of the UV and opticalphotometry of SN 2017eaw up to ∼
45 days after the dis-covery. The unfiltered light curve derived from the obser-vations by Patrick Wiggins is overplotted. The first nonde-tection can be traced back to 2017 May 12.20 UT, with anunfiltered CCD magnitude upper limit of about 19.0 mag.The explosion time (or first light) can be estimated as JD2457886.72 ± UBV RI bands, while the UV dataalready started to decline from the first data points. Wefit the TNT R -band light curve during the first week afterexplosion with a f ∝ ( t − t ) n function, and find that thefirst (unfiltered) discovery point is brighter than the fittedcurve by about 0.2 mag. After normalizing the unfilteredmagnitudes to the R-band values in the plateau phase,this difference is about 0.16 mag, which could be relatedto the emission of shock breakout cooling tail. Such post-shock cooling decline before reaching the optical peakhas been reported for SN 2014cx (Huang et al. 2016),SN 2016X (Huang et al. 2018), SN 2006bp (Quimby et al.2007), KSN 2011d (Garnavich et al. 2016). We caution,however, that the evidence for shock cooling detection inSN 2017eaw should be less significant due to that the dif-ference between the unfiltered and R-band magnitudeschange with time as a result of evolution of the spectralenergy distribution of SNe IIP in the early phase.By fitting a low-order polynomial to the data aroundmaximum light, the supernova reached the U - and B -bandpeak of 12.39 mag and 13.16 mag at t ∼ +5.0 days andt ∼ +5.1 days from the discovery respectively. We note thatthe rise time in B band is almost the shortest among alarge sample of SNe II (Gonz´alez-Gait´an et al. 2015).For the spectrum of SN 2017eaw displayed in Fig.2, the first two were obtained with BFOSC and the thirdone was obtained with OMR on the Xinglong 2.16-m tele-scope. All spectra were reduced using standard IRAF rou-tines. Flux of these spectra were calibrated using nearbyspectrophotometric standard stars with similar airmass.Atmospheric extinction was corrected basing on the ex-tinction curve of Xinglong observatory.The earliest spectrum, obtained at ∼ IRAF is distributed by the National Optical Astronomy Ob-servatories, which are operated by the Association of Universi-ties fo Research in Astronomy, Inc., under cooperative agree-ment with the National Science Foundation (NSF). MNRAS000
45 days after the dis-covery. The unfiltered light curve derived from the obser-vations by Patrick Wiggins is overplotted. The first nonde-tection can be traced back to 2017 May 12.20 UT, with anunfiltered CCD magnitude upper limit of about 19.0 mag.The explosion time (or first light) can be estimated as JD2457886.72 ± UBV RI bands, while the UV dataalready started to decline from the first data points. Wefit the TNT R -band light curve during the first week afterexplosion with a f ∝ ( t − t ) n function, and find that thefirst (unfiltered) discovery point is brighter than the fittedcurve by about 0.2 mag. After normalizing the unfilteredmagnitudes to the R-band values in the plateau phase,this difference is about 0.16 mag, which could be relatedto the emission of shock breakout cooling tail. Such post-shock cooling decline before reaching the optical peakhas been reported for SN 2014cx (Huang et al. 2016),SN 2016X (Huang et al. 2018), SN 2006bp (Quimby et al.2007), KSN 2011d (Garnavich et al. 2016). We caution,however, that the evidence for shock cooling detection inSN 2017eaw should be less significant due to that the dif-ference between the unfiltered and R-band magnitudeschange with time as a result of evolution of the spectralenergy distribution of SNe IIP in the early phase.By fitting a low-order polynomial to the data aroundmaximum light, the supernova reached the U - and B -bandpeak of 12.39 mag and 13.16 mag at t ∼ +5.0 days andt ∼ +5.1 days from the discovery respectively. We note thatthe rise time in B band is almost the shortest among alarge sample of SNe II (Gonz´alez-Gait´an et al. 2015).For the spectrum of SN 2017eaw displayed in Fig.2, the first two were obtained with BFOSC and the thirdone was obtained with OMR on the Xinglong 2.16-m tele-scope. All spectra were reduced using standard IRAF rou-tines. Flux of these spectra were calibrated using nearbyspectrophotometric standard stars with similar airmass.Atmospheric extinction was corrected basing on the ex-tinction curve of Xinglong observatory.The earliest spectrum, obtained at ∼ IRAF is distributed by the National Optical Astronomy Ob-servatories, which are operated by the Association of Universi-ties fo Research in Astronomy, Inc., under cooperative agree-ment with the National Science Foundation (NSF). MNRAS000 , 1–11 (XXX) Days Since Explosion A pp a r e n t m a g n i t u d e Rcuvw2+4.5uvm2+2.75uvw1+2.5uvot u+2.5uvot b+0.5uvot v0.25TNT U+1.75TNT B+0.5TNT V0.25TNT RTNT I-0.25
Figure 1.
Optical and Ultraviolet follow-up photometry of SN 2017eaw obtained with the Tsinghua-NAOC 0.8-m telescope (TNT)and the Swift UVOT. The dark dashed lines show the best fit to the R-band data in a form of f ∝ ( t − t ) n , and the first open starrepresents the last non-detection(unfiltered) from Patrick Wiggins. (Xiang et al. 2017a). These Balmer lines become domi-nant features in the following spectra, similar to thoseseen in SNe IIP(Fig. 2). Broad absorption feature of He I λ + = +
8d is similar to other well-studiedtype IIP supernova(especially SN 2004et), while showingweaker and broader profiles of Balmer and He I lines com-pared to SN 1999em. Noticeable narrow interstellar Na ID absorption features can be seen at around 5892.5˚A inthe spectra, with an equivalent width (EW) of 1.6 ± V =1.24 ± ± V =1.83 ± V =0.94 (Schlafly & Finkbeiner 2011) towardsSN 2017eaw, the host-galaxy extinction is estimated asA V =0.89 ± Pre-explosion images of SN 2017eaw are available fromMikulski Archive for Space Telescopes (MAST) and the
MNRAS , 1–11 (XXX)
Rui LM et al. l o g ( F l u x ) + c o n s t a n t H γ H β H e I N a I D H α Figure 2.
The spectra of SN 2017eaw in the early phase, and comparison with those of other well-studied SNe IIP around themaximum.
Table 1.
Log of the pre- and post-explosion HST images covering the explosion site of SN 2017eawInstrument Filter Obs date Exposure time Prospsal ID PI Drizzled image a ACS/WFC F814W 2016-10-26 + + × + + × × × × × × × a The retrieved drizzled images used in geometric transformation and finding of the progenitor of SN 2017eaw.
Hubble Legacy Archive (HLA). By searching the pub-licly available archive images, we found there are F128N-and F110W-band (WFC3/IR) images that were takenon 9 February 2016 (PI Leroy), F164N- and F160W-band images taken on 24 October 2016 (PI Long), andF606W- and F814W-band (ACS/WFC) images taken on26 October 2016 (PI Williams). Besides many imagestaken in 2016, there are also F658N- and F814W-band(ACS/WFC) images taken on 29 July 2004 (PI Ho). Ta-ble 1 list the observation log of these images.To measure the SN position on these HST images, weused I -band image of SN 2017eaw taken by the Tsinghua-NAOC 0.8-m telescope (TNT) on 17 May, which hasa typical full-width at half maximum (FWHM) of . ′′ .The position of SN 2017eaw was determined by averagingthe results from six different methods (centroid, Gaussianand ofilter center algorithm of IRAF task phot, IRAF task imexamine, daofind, and SExtractor) as ( . ± . , . ± . ) . The uncertainty of this position isestimated as the standard deviation of the mean value.Several stars commonly seen on the HST images and the I -band TNT image are used to establish the transformationfunction (a 2nd-order polynomial) converting the coordi-nates of the I -band TNT image to that of the HST images.The resulting positions and uncertainties are listed in Ta-ble 2. The uncertainties include errors in SN position andgeometric transformation. The reference stars used in po-sition transformation varies in different HST wavebands,and the exact number is listed as N stars in Table 2.Figure 3 shows the region centering on the SN sitewith the transformed position marked. A point-like sourcecan be detected near the position of the SN on the F128N-, F110W- and F814W-band images, while no detection onthe F658N-band image. Difference between the positions MNRAS000
Hubble Legacy Archive (HLA). By searching the pub-licly available archive images, we found there are F128N-and F110W-band (WFC3/IR) images that were takenon 9 February 2016 (PI Leroy), F164N- and F160W-band images taken on 24 October 2016 (PI Long), andF606W- and F814W-band (ACS/WFC) images taken on26 October 2016 (PI Williams). Besides many imagestaken in 2016, there are also F658N- and F814W-band(ACS/WFC) images taken on 29 July 2004 (PI Ho). Ta-ble 1 list the observation log of these images.To measure the SN position on these HST images, weused I -band image of SN 2017eaw taken by the Tsinghua-NAOC 0.8-m telescope (TNT) on 17 May, which hasa typical full-width at half maximum (FWHM) of . ′′ .The position of SN 2017eaw was determined by averagingthe results from six different methods (centroid, Gaussianand ofilter center algorithm of IRAF task phot, IRAF task imexamine, daofind, and SExtractor) as ( . ± . , . ± . ) . The uncertainty of this position isestimated as the standard deviation of the mean value.Several stars commonly seen on the HST images and the I -band TNT image are used to establish the transformationfunction (a 2nd-order polynomial) converting the coordi-nates of the I -band TNT image to that of the HST images.The resulting positions and uncertainties are listed in Ta-ble 2. The uncertainties include errors in SN position andgeometric transformation. The reference stars used in po-sition transformation varies in different HST wavebands,and the exact number is listed as N stars in Table 2.Figure 3 shows the region centering on the SN sitewith the transformed position marked. A point-like sourcecan be detected near the position of the SN on the F128N-, F110W- and F814W-band images, while no detection onthe F658N-band image. Difference between the positions MNRAS000 , 1–11 (XXX) Figure 3.
The zoomed-in region of the progenitor site on the HST pre-explosion images. All images are aligned. The white crossmarks the measured center of the possible SN progenitor. The center of the ellipses shows the transformed SN positions, with thesize of indicating the uncertainty. Red ellipses show the SN position determined from the post-explosion HST image, while the greenones represent the ones from the post-explosion TNT image. F110W(a): F110W image of icyqf2010 drz.fits; F110W(b): F110Wimage of icyqe2040 drz.fits.
Table 2.
Transformed SN positions on pre-explosion HST images using post-explosion TNT imageimage F128N F110W(a) ∗ F110W(b) ∗∗ F814W(2004) F658N s pre / ( arcsec · pixel − ) a N b stars
25 24 26 9 14 s post / ( arcsec · pixel − ) c σ total / mas d ( x , y ) e SN (558.28, 650.24) (569.27, 656.08) (539.49, 633.24) (3360.32, 3978.51) (877.62, 267.01) ( x , y ) ∗ f (558.23, 650.44) (569.28, 656.50) (539.65, 633.47) (3360.48, 3978.36) ··· image F160W F164N F606W F814W(2016) s pre / ( arcsec · pixel − ) a N b stars
18 18 15 16 s post / ( arcsec · pixel − ) c σ total / mas d ( x , y ) e SN (935.82, 600.28) (935.91, 600.43) (2809.77, 989.06) (2809.59, 990.81) ( x , y ) ∗ f (936.04, 600.64) (936.22, 600.78) (2809.33, 989.40) (2809.23, 990.51) ∗ F110W image of icyqf2010 drz.fits. ∗∗ F110W image of icyqe2040 drz.fits. a Pixel scale of the pre-explosion image. b Number of stars used in the geometric transformation. c Pixel scale of the post-explosion image. d Total uncertainty of the transformed SN position on the pre-explosion images, with the values separated by the slashrepresenting uncertainties along X-axis and Y-axis, respectively. e The transformed SN position of pre-explosion image. f The position of center of the nearby star.MNRAS , 1–11 (XXX)
Rui LM et al. of SN and this point-like source is generally less than theuncertainty of the transformed positions.On 29 May 2017, an F814W-band image of the SNwas taken again by the HST with WFC3/UVIS (PI VanDyk), which became publicly available immediately. Weretrieved the image and measured the SN position, withthe X-pixel being at 547.90 ± ± package on the bias-subtracted, flat-corrected, so-called FLT FITS imagesthat are retrieved from the MAST. Photometry is si-multaneously measured using all ACS/WFC images andWFC3/IR images, following the routines that includethe selection of a deep drizzled image as reference im-age, masking of bad pixels, calculations of sky back-ground, alignment of images to the reference, and adop-tion of the recommended photometry parameters as de-scribed in the User’s Guide of DOLPHOT. Magnitudesand their uncertainties of the progenitor candidate areextracted from the output of DOLPHOT. The Magni-tudes are PSF magnitudes, and all in Vega system. Anupper limit of magnitude is given for the non-detectionF658N image. The measured magnitudes of the SN pro-genitor in different wavebands and phases are reported inTable 3, with F606W-band and F814W-band magnitudesbeing 26.419 mag and 22.845 mag. The magnitudes ofthe SN progenitor were also independently measured byKilpatrick & Foley (2018) using the published HST im-ages. In F606W, F814W, F160W bands, our results areconsistent with theirs within 0.05 mag, while in F110Wand F164N their results are fainter than ours by 0.4 magand 0.1 mag, respectively.
The position calibration step was skipped for the Spitzerimages, as the retrieved WCS information is well consis-tent with that from the GSC-2.3 standard catalogue. Wefind the counterpart of the progenitor candidate at the ex-pected coordinates on . µ m and . µ m images (Figure4). On the . µ m , . µ m and µ m images, there are noclear point-like sources at the SN site. Photometry of theprogenitor on Spitzer images is directly taken from (Khan2017a), and the limiting magnitudes of non-detection im-ages are estimated using the same method. It should bepointed out the flux at the SN site of these images couldbe seriously contaminated by nearby sources that are notrelated to the progenitor of the SN(Khan 2017b) due tothe low angular resolution of the Spitzer images. For ex-ample, the photometric aperture is adopted as 2.4” for the3.6um/4.5um images(Khan 2017a), which corresponds to48 pixels on the HST ACS images and 19 pixels on theWFC3/IR images, respectively. Adopting this large aper-ture for the HST images would result in an overestimate ofthe F814W-band magnitude by about 1.0 mag (i.e., ∼ http://americano.dolphinsim.com/dolphot/ mag). Thus, we caution the use of the Spitzer magnitudesto constrain the progenitor properties of SN 2017eaw. SN 2017eaw provides a good opportunity to study theprogenitor of SNe IIP, as the pre-explosion Hubble SpaceTelescope (HST) images are available in both optical andnear-infrared bands. Inspecting the HST WFC3 imagestaken in 2016 reveals that a point source can be identi-fied at the SN position in the F606W-, F814W-, F110W-,F128N-, F160W-, and F164N-bands (see Fig. 3 and Sec.3.1 for details of the progenitor identification). This pointsource tends to become progressively brighter at longerwavelengths, consistent with a very red star (see Table 1and Fig.3). In the F814W band, the SN progenitor is mea-sured to have a magnitude of 22.845 ± − ± µ =28.67 ± V = 1.83 ± Spitzer images at . µ m and . µ m wave-lengths (see Fig.4). The point source has an absolute mag-nitude of − ± µ m] and a colorof 0.27 ± µ m]-[4.5 µ m], which are con-sistent with the typical values of RSGs(Szczygie l et al.2010). Note, however, that the nearby red stars could con-taminate the Spitzer magnitudes due to the low angularresolution of the Spitzer images.It is interesting that the region including the site ofSN 2017eaw was also observed in 2004 by the HST Ad-vanced Camera for Surveys (ACS). These observationswere made in the F658N and F814W bands, where theprogenitor can be clearly detected in F814W with m F W = 22.558 ± ∼ K to 4200 K to fit the observedspectral energy distribution (SED). This sample consistsof 74 Galactic RSGs(Levesque et al. 2005), 7 RSGs inthe Large Magellanic Cloud (LMC), and 4 RSGs in theSmall Magellanic Cloud (SMC)(Levesque et al. 2007). Asecond-order polynomial is used to fit the BV RIJH -bandmagnitudes of the observed RSGs and then derive theirSED curves. These curves are then compared with thereddening-corrected SED of the progenitor. The compar-ison shows that the progenitor matches well with an M4-type red supergiant with an effective temperature of about3550 ±
100 K (see the upper panel of Fig.5). In compari-son, Kilpatrick & Foley (2018) derived the effective tem-perature of 3350 + − K by fitting the SED of progeni-tor with stellar SED and dust models. Based on the de-rived absolute K-band magnitude (i.e., − ± MNRAS000
100 K (see the upper panel of Fig.5). In compari-son, Kilpatrick & Foley (2018) derived the effective tem-perature of 3350 + − K by fitting the SED of progeni-tor with stellar SED and dust models. Based on the de-rived absolute K-band magnitude (i.e., − ± MNRAS000 , 1–11 (XXX) Table 3.
Photometry of the Progenitor Star Identified for SN 2017eaw.Telescope Instrument Filter Exposure time(s) m a HST ACS/WFC F606W 530+520+690 × × > × × × × × . µ m . µ m . µ m > . µ m > µ m > a The values listed in the parentheses represent the 1- σ uncertainty of magnitude, in unit of 0.001 mag. The upper limit listed forsome observations represents the 5- σ detection limit. Figure 4.
Subsections of pre-explosion Spitzer images centering on the site of the SN. All images are aligned. The white cross hairsin the . µ m and . µ m images mask the progenitor candidate. nosity of the progenitor is derived as − ± L bol / L ⊙ ) = 4.88 ± ⊙ to 14 M ⊙ and a radius of about 700 R ⊙ .The radius of progenitor can be estimated from theevolution of photospheric temperature in the early stageof the explosion, which is primarily determined by theradius, opacity, and density profile of the progenitor(Rabinak & Waxman 2011). This early-time temperatureevolution can be derived using the ultraviolet (UV) andoptical photometry obtained within the first week afterthe explosion. Adopting the typical density profile as f ρ = 0.13 and Thomson scattering opacity as 0.34 cm g − ,the progenitor radius of SN 2017eaw can be determined as636 ± ⊙ (Fig.6). Moreover, the rise time of the lightcurve can be also used to estimate the radius of the pro-genitor of type IIP supernova through the relation be- tween the progenitor radius and the rise time of the lightcurve, i.e., log R[R ⊙ ] = (1.225 ± rise [day] +(1.692 ± rise = 6.81 ± ±
180 R ⊙ . The above two estimates of the progenitorradius agree within 1- σ error.We also try to fit the observed SED of the progenitorstar from the prediscovery HST and Spitzer images withsynthesized spectra calculated using the stellar evolutioncode MARCS . However, the best-fit model spectrum hasan effective temperature of only about 2500 K (A similarvalue of 2600 K was also suggested by Kilpatrick & Foley(2018)) with A V =1.83 ± ⊙ for theprogenitor, which is inconsistent with the estimate eitherfrom the rise time or temperature evolution of the shock http:marcs.astro.uu.se/ MNRAS , 1–11 (XXX)
Rui LM et al. A pp a r e n t M a g n i t u d e HD 14469T = 3575Kspectral type: M3-4 Iχ = 2.16 RSG3550±100KProgenitor eff /K)3.84.04.24.44.64.85.05.25.4 l o g ( L / L ⊙ ) R ⊙ R ⊙ R ⊙ R ⊙ ⊙ ⊙ ⊙ ⊙ Figure 5.
Spectral energy distribution (SED) of the progeni-tor star and its location in Hertzsprung-Russell diagram. Upperpanel: the SED of the progenitor star compared to that of theGalactic Red Supergiants (RSGs). The red points represent theSED of the progenitor of SN 2017eaw obtained in 2016, whilethe black dots represent the ones of red supergiants from MilkyWay. The black dashed line is yielded by applying a second-order polynomial fit to the values of Galactic RSGs. The grayshadow region represents the distribution inferred from theGalactic RSGs with the temperature ranging from 3450 Kto 3650 K. Lower panel: the solid lines represent the stellarevolution tracks for 9 −
20 M ⊙ , based on the Geneva model(Ekstr¨om et al. 2012). The measurements of the progenitors ofSN 2005cs (Maund et al. 2005) and SN 2013ej (Fraser et al.2014) are over-plotted for comparison. The dashed straightlines indicate the lines of equal stellar radii. cooling in the early phase. Despite an overall reasonablefit, the measurements in two Spitzer bands may sufferlarge contaminations of the nearby red sources due tolarge aperture photometry and/or the presence of circum-stellar dust. In the latter case, the light of the progeni-tor star is scattered/absorbed by the surrounding dustand re-emitted at longer wavelengths (see also the anal-ysis by (Kilpatrick & Foley 2018)). We also derive theF606W-band magnitude in 2004 by assuming that thedimming behavior in F814W-band is totally due to theCSM dust. Taking into account this speculated F606W-band magnitude and the observed F814W-band magni-tude obtained in 2004, the SED of the progenitor inferredin 2004 matches with the theoretical stellar spectra modelcharacterized by an effective temperature of 3550K ± ± ⊙ (see Fig. 7). As an alternative,the dimming behavior in F814W band could be also dueto the intrinsic variation of a RSG star, but this explana- Days since explosion T e ff ( K ) RSG 636±155 R ⊙ RSG 850R ⊙ RSG 400R ⊙ BSG 8⊙R ⊙ SN 1987ASN 2017eaw
Figure 6.
Constraining the progenitor radius of SN 2017eawbased on the cooling of the photospheric temperature in theearly phase after the explosion. The red dots are the tempera-tures inferred from the blackbody fit to the observed UV andoptical photometry; while the red curve represents the best-fitof the radius. The green and black curves show the upper andlower limits of the estimates of the radius, respectively. Theblue stars show the temperature evolution of SN 1987A, andthe blue dashed line shows the corresponding best-fit radiususing the theoretical model (Rabinak & Waxman 2011).
Effective wavelength(Å) (19.5(19.0(18.5(18.0(17.5(17.0(16.5 l g ( F l u x ( e r g c m − s − )) Figure 7.
The gray squares and red circles are the spectral en-ergy distribution of the progenitor observed in 2004 and 2016respectively. The best fitting result of the observed SED in 2016with theoretical MARCS stellar spectra indicates the progeni-tor to be a RSG with effective temperature of 2500K. The opensquare is the estimated F606W-band result in 2004. And thegray shadow region is the distribution of the SED inferred fromthe MARCS stellar spectra model with a radius of 636 ± ⊙ . tion is inconsistent with either the observation taken atone year to a few days before the explosion or the theoret-ical prediction for its final stage evolution (see discussionsin Sec 4.3). Some recent studies reveal that the ion-flashed featurescan be detected in very early spectra of SN 2013fs(Yaron et al. 2017), and SN 2016bkv (Nakaoka et al.2018), which proves the existence of CSM around SNeIIP. The detection of prominent asymmetric H α emissionfeature in the t= + α line profile seen in SN 2017eaw can be MNRAS000
The gray squares and red circles are the spectral en-ergy distribution of the progenitor observed in 2004 and 2016respectively. The best fitting result of the observed SED in 2016with theoretical MARCS stellar spectra indicates the progeni-tor to be a RSG with effective temperature of 2500K. The opensquare is the estimated F606W-band result in 2004. And thegray shadow region is the distribution of the SED inferred fromthe MARCS stellar spectra model with a radius of 636 ± ⊙ . tion is inconsistent with either the observation taken atone year to a few days before the explosion or the theoret-ical prediction for its final stage evolution (see discussionsin Sec 4.3). Some recent studies reveal that the ion-flashed featurescan be detected in very early spectra of SN 2013fs(Yaron et al. 2017), and SN 2016bkv (Nakaoka et al.2018), which proves the existence of CSM around SNeIIP. The detection of prominent asymmetric H α emissionfeature in the t= + α line profile seen in SN 2017eaw can be MNRAS000 , 1–11 (XXX) decomposed into two components, with the broad compo-nent formed in the SN ejecta and the weak narrow peakformed due to the ionization of the surrounding CSM.The weak emission feature is measured to have a wave-length of 6559.50 ± − for the CSM, as shown inFig.9. This measurement of velocity is validated by care-fully inspecting the position of sky emission lines [O I] λ λ ∼
100 km − ) (Yaron et al. 2017).Similar asymmetric H α line profile can also be seenin the earliest spectra of SN 2006bp and iPTF13dkk(Khazov et al. 2016), but only narrow emission featuresare shortly detected in the flash spectra of SN 2013fs, SN2016bkv, and SN 2018zd. In the very early phase, thespectra of SN 2013fs and SN 2006bp also show prominentHe II lines, while this feature is blended with C III/N IIIin iPTF13dkk, SN 2016bkv, and SN 2017eaw. Such an ob-served diversity of the flashed spectroscopy suggests thatthe CSM/dust environments around their progenitor starsare quite diverse.The duration of narrow emission features also variesfor different SNe II (see Fig.8). For SN 2017eaw, the weaknarrow H α emission lines seen at t ∼ +1.4 days disappearedin the t ∼ +3.4 spectrum, suggesting that the obscuringdust shell should be thin and close to the progenitor star.Such a quick variation can be explained by that the CSMwas initially ionized by the supernova photons, and thenit was captured and destroyed by the fast-expanding SNejecta. Assuming that the outermost layer of the SN ejectahas a velocity of 15,000-20,000 km s − as inferred fromthe blue-shifted absorption of H α and it swept the CSMwithin two days, and we estimate that the CSM has a dis-tance < × cm from the supernova. With this dis-tance, one can speculate that the progenitor star ejectedthe CS materials at 1-2 years before the explosion withthe derived stellar wind velocity of about 160 km s − . To understand the dimming behavior of the progeni-tor star in F814W-band over the period from 2004 to2016, we examined in Figure 11 the magnitude varia-tion of the stars within 200 pc from the progenitor. On atimescale of a decade, most of the field stars are foundto have light variations less than 0.3 mag when con-sidering the larger photometric uncertainty at the faintend (i.e., m F W > F W -m F W color as measured from the 2016 HST images. Af-ter corrections for the reddening, the progenitor is foundto have a m F W -m F W color of 2.94 ± ⊙ , in particular when consid-ering the effect of CSM dust (see Fig.11). On the otherhand, instability mass ejection is also needed to repro-duce the coolest (or reddest) variables with masses lowerthan 10-15 M ⊙ according to the census of luminous stellarvariability in M51 (Conroy et al. 2018). In the supergiantinstability regime, a pulsational superwind could lead toa dramatically enhanced mass loss at late phases, whichmay account for an ejection of the H-envelope of RSGstars just before their explosions (Yoon & Cantiello 2010;Conroy et al. 2018). From the above analysis, we conclude that the normaltype IIP supernova 2017eaw has an M4-type red super-giant star with an initial mass 12 ± ⊙ , which furtherconfirms the trend that most SNe IIP arise from core-collapse explosions of RSGs. However, the multi-epochHST images obtained before the SN detection reveal thatthe RSG progenitor became faint by 30% one year beforeit exploded. Although this dimming phenomenon couldbe due to the intrinsic light variation of the RSG star,detailed analysis of the neighbouring stars and early timespectra suggest that it is more likely due to the obscura-tion of circumstellar dust shell newly-formed at a distanceof < × cm from the star. The fast-moving cir-cumstellar materials (at a velocity of ∼
160 km s − ) indi-cate that they were ejected from the star in a violent man-ner within a few years prior to the explosion. Thus, theobservations of SN 2017eaw indicates that some low-massRSGs could also experience a strong mass loss, perhapsinvoked by pulsational instability, during its final-stageevolution towards explosion. ACKNOWLEDGEMENTS
We thank Patrick Wiggins for useful unfiltered images.This work is supported by the National Natural Sci-ence Foundation of China (NSFC grants 11325313 and11633002), and the National Program on Key Researchand Development Project (grant no. 2016YFA0400803).J.-J. Zhang is supported by NSFC (grants 11403096,11773067), the Youth Innovation Promotion Associationof the CAS(grants 2018081), the Western Light YouthProject, and the Key Research Program of the CAS(Grant NO. KJZD-EW-M06). T.-M. Zhang is supportedby the NSFC (grants 11203034). This work was also par-tially Supported by the Open Project Program of the KeyLaboratory of Optical Astronomy, National AstronomicalObservatories, Chinese Academy of Sciences. The researchof JRM is supported through a Royal Society UniversityResearch Fellowship.
MNRAS , 1–11 (XXX) Rui LM et al. l o g ( F l u x / F BB ) + c o n s t HHeC
N IIIO VI O V
SN 2017eawiPTF13dkkSN2006bpSN 2016bkvSN 2013fsSN 2018zd+1.4day+1.2day+2day+4.1day+6.2hour+4daySN 2013fs +2day 6563+3.4day+6day+8day+7day+8day+1.4day+1.2day+2day+4day+4.1day+6.2hour
Rest Wavelength(Å)
Figure 8.
Early time spectroscopy of SN 2017eaw and some comparison SNe II that show clear signatures of circumstellar materials.Left panel: the early spectroscopy of SN 2017eaw, SN 2006bp (Quimby et al. 2007), SN 2013fs (Yaron et al. 2017), iPTF13bkk(Khazov et al. 2016), SN 2016bkv (Nakaoka et al. 2018), and SN 2018zd (from our own database), covering the wavelength from3500˚A to 7500˚A. All the spectra were obtained within 4 days after explosion and have been normalized to the continuums andcorrected for the host-galaxy redshift. The positions of narrow emission lines of hydrogen, helium, and carbon are marked withdashed lines. Right panel: early-time spectral evolution centering at H α line profile, with the ion-flashed/circumstellar featuresbeing initially detected and subsequently having disappeared. The red flux deficiency and blue excess asymmetry feature of H α inthe early spectra of SN 2017eaw also disappears in the later spectra (at t ∼ +3.4day) making the line profile symmetric, which is asignature of dust. −0.06−0.04−0.020.000.020.040.060.080.100.12N o r m a li z e d F l u x (+1.4 day) −0.10−0.050.000.050.100.15 (+3.4 day) H α Lo entzianGaussianLo entzian+Gaussian−20000 −15000 −10000 −5000 0 5000 10000
Velocity(km s −1 ) −3σ0.03σ R e s i d u a l (CSM) −20000 −15000 −10000 −5000 0 5000 10000 Velocity(km s −1 ) −3σ0.03σ Figure 9.
The H α line profile in the earliest two spectra of SN 2017eaw, displayed in a velocity space.Left panel: the H α line profilein t=+1.4 day spectrum. The red curve represents the best-fit result using a combination of Lorentzian and Gaussian functions,with the blue and magenta curves showing the absorption and emission components of the P-Cygni profile, respectively. Rightpanel: the same case as in the left panel but for the t=+3.4 day spectrum. The lower panels show the residual of the observed lineprofile relative to the best-fit profile; and the narrow emission component in the left panel, due to the CSM, can be fitted by aLorentzian function with a center wavelength of 6559.50 ± REFERENCES
Barbon R., Ciatti F., Rosino L., 1979, A&A, 72, 287Barbon R., Benetti S., Rosino L., Cappellaro E., Turatto M.,1990, A&A, 237, 79Branch D., Wheeler J. C., 2017, Supernova Explosions,doi:10.1007/978-3-662-55054-0. Cheng Y.-C., Chen T.-W., Prentice S., 2017, The Astronomer’sTelegram, 10374Conroy C., et al., 2018, preprint, ( arXiv:1804.05860 )Ekstr¨om S., et al., 2012, A&A, 537, A146Faran T., et al., 2014, MNRAS, 442, 844Filippenko A. V., 1997, ARA&A, 35, 309Fraser M., et al., 2014, MNRAS, 439, L56MNRAS000
Barbon R., Ciatti F., Rosino L., 1979, A&A, 72, 287Barbon R., Benetti S., Rosino L., Cappellaro E., Turatto M.,1990, A&A, 237, 79Branch D., Wheeler J. C., 2017, Supernova Explosions,doi:10.1007/978-3-662-55054-0. Cheng Y.-C., Chen T.-W., Prentice S., 2017, The Astronomer’sTelegram, 10374Conroy C., et al., 2018, preprint, ( arXiv:1804.05860 )Ekstr¨om S., et al., 2012, A&A, 537, A146Faran T., et al., 2014, MNRAS, 442, 844Filippenko A. V., 1997, ARA&A, 35, 309Fraser M., et al., 2014, MNRAS, 439, L56MNRAS000 , 1–11 (XXX) −1000 −500 0 500 1000 (1.4d)Δλ=-0.0Å OI λ5577Å −1000 −500 0 500 1000 (1.4d)Δλ=-1.4Å
OI λ6300Å −1000 −500 0 500 1000 (3.4d)Δλ=-1.1Å −1000 −500 0 500 1000 (3.4d)Δλ=-4.3Å R e l a t i v e F l u x Relative Velocity(km s −1 ) Figure 10. [OI λλ λλ + . day spectrum, while the lower panel indicates the casefor the t ∼ A TEX file preparedby the author.MNRAS , 1–11 (XXX) Rui LM et al. m F814W, 2016 (mag) −1.2−1.0−0.8−0.6−0.4−0.20.00.20.4 m F W , − m F W , ( m a g ) B C D A(progenitor)E F GField ObjectsProge itor −0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 m F606W, 2016 − m
F814W, 2016 (mag) −9.0−8.5−8.0−7.5−7.0−6.5−6.0−5.5−5.0 M F W ( m a g ) B CD A(proge itor)E F G AGB?
Figure 11.
Magnitude and color variations for the progenitor star of SN 2017eaw and the neighboring stars at a distance of 200pc, measured from the HST prediscovery images. Left panel: the F814W-band magnitude variation over the period from 2004to 2016. The blue dashed lines indicate the 3- σ limit of the photometric error. The progenitor of SN 2017eaw is represented bysymbol ”A”, and those stars showing larger variations are also labeled with specific symbols. Right panel: the F606W-F814W colordistribution of progenitor and field stars, measured from the HST images taken in 2016. The red open circle represents the colorthat the progenitor would have had in 2004 assuming that the observed dimming behavior is totally due to the obscuration byaccumulated dust. The orange solid and dashed lines show the instability region of supergiants for stars in the mass range of 15M ⊙ < M <
40 M ⊙ and 10 M ⊙ < M <
15 M ⊙ , respectively. These regions are obtained by applying the pulsation growth rate tothe MIST stellar tracks(Conroy et al. 2018). The objects locating in the dashed circle are likely luminous asymptotic giant branch(AGB) stars according to their luminosity and red colors. MNRAS000