SN 2011dh: DISCOVERY OF A TYPE IIb SUPERNOVA FROM A COMPACT PROGENITOR IN THE NEARBY GALAXY M51

At the end of their lives, massive stars (M ≳ 8 M_☉) explode as core-collapse supernovae (SNe), leaving a neutron star or a black hole remnant. Several different types of SNe have been observed, including H-rich Type II-P/II-L SNe, H-rich interacting Type IIn SNe, H-poor Type IIb SNe, H-less Type Ib SNe, and H- and He-less Type Ic SNe (see Filippenko 1997 for a review).

Direct progenitor detections in pre-explosion images have established a link between low-mass red supergiants (RSGs) and SNe IIP, the most common type of core-collapse SN (see Smartt 2009 for a review). Evidence also exists linking SNe IIn with much more massive progenitors (Gal-Yam et al. 2007; Gal-Yam & Leonard 2009; Smith et al. 2011b). However, the mapping between massive star classes and types of SNe is far from complete, due to both theoretical gaps and to the scarcity of observational constraints. Numerous core-collapse SNe are discovered each year and statistical analyses of large samples are yielding interesting insights (e.g., Anderson & James 2009; Boissier & Prantzos 2009; Arcavi et al. 2010; Li et al. 2011; Smith et al. 2011a), but some of the most robust information comes from studying the few nearest events.

The Palomar Transient Factory (PTF; Rau et al. 2009; Law et al. 2009) is a wide-field variability survey. Our short-cadence strategy and real-time capability (Gal-Yam et al. 2011) enables the discovery of SNe shortly after the explosion. Some parts of the sky, such as those with photogenic nearby galaxies, are observed by amateur astronomers even more frequently. Thus, the amateur community may be the first to discover and report bright events in these nearby hosts, as is the case here.

We report the discovery of SN 2011dh (Silverman et al. 2011; also named PTF11eon), a Type IIb SN in M51, a nearby interacting spiral galaxy located at a distance of 8.03 ± 0.77 Mpc (retrieved from the NASA/IPAC Extragalactic Database, NED²²). We find rapid light curve evolution at early times, as well as relatively cool early spectra compared with those of PTF10vdl (Gal-Yam et al. 2011) and SN 1993J (Wheeler et al. 1993). Our data suggest that SN 2011dh resulted from a progenitor more compact than that of the prototypical Type IIb SN 1993J (Filippenko et al. 1993), but perhaps more extended than the progenitor of the Type IIb SN 2008ax (Pastorello et al. 2008; Chornock et al. 2011).

The SN was discovered in images taken on May 31.893 (UT times are used throughout this paper) by A. Riou in France²³ (Figure 1). Approximately 7 hr later, on 2011 June 1.191, the PTF detected the SN in images taken by the Palomar 48 in Oschin Schmidt Telescope (P48). The SN was seen at α(J2000) = $13^{\rm h}30^{\rm m}05\mbox{$.\!\!^{\mathrm s}$}08$ and δ(J2000) = +47°10'112 (astrometry good to 01, based on the Sloan Digital Sky Survey, SDSS²⁴) at a magnitude of m_g = 13.15 ± 0.09, and it was absent from a PTF image taken by the P48 on 2011 May 31.275 down to a 3σ limiting magnitude of m_g = 21.44. Subsequent imaging was undertaken at the Wise Observatory 18 in telescope on June 1.776, by T. Griga on June 1.897 and by S. L. Bailey on June 2.036. Data collected by amateur astronomers during the first hours of this event are very helpful in determining the explosion time. The data described above constrain the explosion time to be between May 31.275 and May 31.893; however, this time window can still be decreased (A. Gal-Yam et al. 2011, in preparation). SN 2011dh is the third SN to be discovered in M51 in the last 17 years, after the subluminous Type IIP SN 2005cs (e.g., Li et al. 2006b) and the Type Ic SN 1994I (e.g., Filippenko et al. 1995).

Zoom In Zoom Out Reset image size

Li & Filippenko (2011a, 2011b) identify a star at the approximate location of the SN in pre-explosion Hubble Space Telescope (HST) images. Maund et al. (2011) confirm this with adaptive optics to within 23 mas and Van Dyk et al. (2011) improve the location accuracy to 7 mas. Maund et al. (2011) further claim that this star, an F8 supergiant with log (L/L_☉) = 4.92 ± 0.20 and T_eff = 6000 ± 280 K (which implies a radius of ∼10¹³ cm), is the progenitor of SN 2011dh. Here we show that this is unlikely.

3.1. Photometry

Immediately after discovery, we initiated an extensive follow-up campaign. Optical and infrared (IR) photometry was obtained at P48, the Wise Observatory 1 m and 18 in (C18) telescopes, the Byrne Observatory at Sedgwick Reserve (BOS) 0.8 m telescope, and the Peters Automated InfraRed TELescope (PAIRITEL). Our photometry is performed on difference images by constructing a reference frame from data taken before the SN explosion and then subtracting that from each frame in which the SN is present, matching the point-spread functions (PSFs) of the two images (Gal-Yam et al. 2004; Gal-Yam et al. 2008). We calibrate the Wise data and BOS B-band data using Landolt (1992) standards. The BOS i'-band data were calibrated to SDSS. We use the average magnitudes obtained from the same filter and instrument taken consecutively on the same night. For the P48 data, the SN photometry was measured using a PSF fitting method after image subtraction, due to saturation of the SN in the images.²⁵ In each image frame, the PSF was determined from nearby field stars, and this average PSF was then fit at the position of the SN, weighting each pixel according to Poisson statistics. Saturated pixels (when present) have their weight set to zero. P48 images where the SN was brighter than 12.5 mag were not used. We calibrate our P48 light curve to the SDSS system using SDSS observations of the same field, and including both color and color-airmass terms to determine the zero point of each image (the root-mean square of these color-term fits is ∼0.015 mag with respect to published SDSS magnitudes). For the PAIRITEL data, all individual frames from the same night were stacked using the Swarp²⁶ software. Reference frames for the subtractions were taken from the Two Micron All Sky Survey (2MASS)²⁷ using the HOTPANTS²⁸ software. The subtracted images were then calibrated with respect to point sources from 2MASS. Calibration and subtraction errors were summed in quadrature.

The Ultraviolet/Optical Telescope (UVOT) on board Swift followed SN 2011dh, and preliminary results from these observations were reported by Kasliwal & Ofek (2011) and Arcavi et al. (2011). We retrieved the level-2 UVOT data for SN 2011dh from the Swift data archive. To increase the signal-to-noise ratio, we stacked the images for each individual filter on a daily basis. To remove host-galaxy contamination underlying the SN location, we subtracted pre-outburst images of M51 obtained by the UVOT. We caution that, due to the nonlinearity of the coincidence-loss correction, such a technique can lead to modest systematic uncertainties in the SN flux, particularly at high count rates. For the U filter, we employed the photometric calibration described by Li et al. (2006a) to measure the photometry of SN 2011dh in the subtracted images. For the three UV filters (UVW1, UVW2, and UVM2), the calibration from Poole et al. (2008) was adopted. Many tests have shown that the two calibrations by Li et al. (2006a) and Poole et al. (2008) are consistent with each other for bright sources.

We note that the images by A. Riou and T. Griga were taken using nonstandard photometric filters or no filters at all. We calibrate them to the R filter, and thus expect some systematic error due to filter differences. Because reference images are not available for these data, we perform aperture photometry on the SN with no image subtraction.

Our photometry is corrected for Galactic extinction using the Schlegel et al. (1998) maps via NED and applying the Cardelli et al. (1989) algorithm assuming R_V = 3.1. We do not correct for host-galaxy extinction, as it is found to be small (see below). We adopt a distance modulus of 29.52 mag and present our photometry in the Vega system (except for the P48, i'-band BOS, and UVOT data which are calibrated to the AB system).

Our photometric data are presented in the top panel of Figure 2 and in Table 1. The initial decline in the light curve on days 1–3.5, prior to the rise to peak magnitude, can be interpreted as a detection of the shock-breakout cooling tail (e.g., Chevalier 1992; Waxman et al. 2007; Chevalier & Fransson 2008; Nakar & Sari 2010). Such photometric behavior is reminiscent of that seen in the Type IIb SN 1993J (Richmond et al. 1994), but SN 2011dh exhibits a more rapid decline.

Zoom In Zoom Out Reset image size

Table 1. Photometric Observations of SN 2011dh

JD	Telescope	Instrument	Filter	Magnitude	Error
2455712.775	P48		g'	>21.440	3σ nondetection
2455713.393	A. Riou 14 in		unfiltered	12.870	0.028
2455714.400	A. Riou 10 in		unfiltered	13.310	0.053
2455714.355	T. Griga 12 in		red (Astronomik)	13.550	0.020
2455716.285	Wise 1 m	LAIWO	I	14.568	0.047

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

We also calculate a bolometric light curve (bottom panel of Figure 2) using the Swift, Wise, BOS, and PAIRITEL data between days 6 and 13 after explosion. We find that the flux contained in the J, H, and K filters is roughly constant during this period (33.78% ± 0.85%), and that the flux beyond 2 μm is negligible. We assume a constant IR flux for the interval 14–37 day after explosion, where no PAIRITEL data are available, and we also plot the possible error due to a larger (40%) or smaller (0%) IR flux at these times.

3.2. Spectroscopy

We used SYNOW (Jeffery & Branch 1990) to identify features in the spectra of SN 2011dh and find prominent H features as well as Ca ii and Fe-peak elements characteristic of SNe II (Figure 3). We see these features at velocities as high as ∼ 17,000 km s⁻¹ initially and at an earlier stage compared to other young SNe II, which tend to show a blue continuum with low-contrast Hα emission a few days after explosion (Figure 4). Unlike SN 1993J, no prominent He features are seen in the spectra of SN 2011dh, even out to day 10 after explosion. However, hints of He absorption appear on June 12 (Figure 3) and strong He lines develop in later spectra (Marion et al. 2011). We have further modeled two spectra using the radiative transfer code of Mazzali (2000) with a non-LTE treatment of H and He (Hachinger 2011). On June 12, the ejecta are optically thin outside 8300 km s⁻¹. In this zone, we find 0.024 M_☉ of H and 0.016 M_☉ of He (much more He can presumably be found below 8300 km s⁻¹).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Focusing on our earliest spectrum, the low continuum temperature (7600 K from a blackbody fit; Figure 3) and low-ionization elements indicate rapid cooling after the shock breakout. These are notably different from early-time spectra of SNe IIP (e.g., PTF10vdl; Gal-Yam et al. 2011) known to arise from RSG stars or from early-time spectra of SN 1993J (e.g., Wheeler et al. 1993; Figure 4) whose progenitor size was ∼4 × 10¹³ cm (Woosley et al. 1994). Indeed, using Equation (13) from Rabinak & Waxman (2011) with a radius of 10¹³ cm, we find that two days after explosion the color temperature should be greater than 17,000 K (as observed for RSG explosions such as PTF10vdl; Gal-Yam et al. 2011).³¹

It is highly unlikely that the radius of the progenitor of SN 2011dh was of order 10¹³ cm, as claimed by Maund et al. (2011) and Prieto & Hornoch (2011), although the latter authors do emphasize the need for a smaller progenitor than that of SN 1993J. Therefore, the object identified in pre-explosion HST images (Maund et al. 2011; Van Dyk et al. 2011) is probably not the single progenitor of SN 2011dh. Rather, the progenitor could be in a binary system or a member of a compact cluster. A relatively faint Wolf–Rayet star could have a radius as small as 10¹¹ cm, while possibly remaining below the magnitude limits imposed by the HST data (Van Dyk et al. 2011).

Chevalier & Soderberg (2010) estimate a radius of ∼10¹¹ cm for the progenitor of SN 2008ax, which shows similar spectral properties to those of SN 2011dh (Figure 4). Pastorello et al. (2008) constrain the length of the shock-breakout cooling phase of SN 2008ax to be less than a few hours, implying that the radius of the progenitor of SN 2011dh could be larger than that of SN 2008ax (yet still much smaller than that of SN 1993J). We leave detailed modeling of the shock-breakout cooling phase to a future report.

We conclude that SN 2011dh is likely an SN IIb from a compact progenitor (termed cIIb by Chevalier & Soderberg 2010).

We have presented our discovery of SN 2011dh, an SN in M51, as well as results from our early follow-up effort. This event displays similarities to the Type IIb SNe 1993J and 2008ax, but also important differences. The early-time evolution of the light curve is very rapid compared to that of SN 1993J. The relatively low temperature implied by our first spectrum indicates that SN 2011dh resulted from the explosion of a fairly compact star, which is not consistent with the object identified by Maund et al. (2011) as a single star. Rather, as suggested by Van Dyk et al. (2011), the progenitor system is possibly composed of a binary system of which the progenitor of SN 2011dh is a member.

SN 2011dh demonstrates the importance of real-time detections and of rapid and continuous follow-up observations as crucial tools for constraining both explosion and progenitor properties via shock-breakout analysis (Soderberg et al. 2008; Ofek et al. 2010). Obtaining and analyzing a statistical sample of such young SNe will serve to probe the progenitor stars of a much larger sample, beyond the ∼20 Mpc limit for direct progenitor detections, placing strong and novel constraints on the last phases of massive stellar evolution.

The Weizmann Institute PTF partnership is supported by the Israeli Science Foundation via grants to A.G. Collaborative work between A.G. and S.R.K. is supported by the US-Israel Binational Science Foundation. A.G. further acknowledges support from the EU FP7 Marie Curie program via an IRG fellowship and a Minerva grant. P.E.N. is supported by the US Department of Energy Scientific Discovery through Advanced Computing program under contract DE-FG02-06ER06-04. M.S. acknowledges support from the Royal Society; M.S. and A.G. are also grateful for a Weizmann-UK Making Connections grant. A.V.F.'s supernova group at U.C. Berkeley acknowledges generous support from Gary and Cynthia Bengier, the Richard and Rhoda Goldman Fund, US National Science Foundation grant AST-0908886, and the TABASGO Foundation. J.S.B. acknowledges support of an NSF-CDI grant 0941742 and NSF/AAG grant NSF/AST-100991. P.M., E.P., and E.S.W. acknowledge financial support from INAF through PRIN INAF 2009.

Instrumentation at Wise Observatory was funded in part by the Israel Space Agency (ISA), the Max Planck Institute for Astronomy (MPA) in Heidelberg, Germany, the German Israeli Science Foundation for Research and Development, and the Israel Science Foundation. The WHT is operated by the Isaac Newton Group in the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofisica de Canarias. The Byrne Observatory at Sedgwick (BOS) is operated by the Las Cumbres Observatory Global Telescope Network. The W. M. Keck Observatory is operated as a scientific partnership among the California Institute of Technology, the University of California, and NASA; it was made possible by the generous financial support of the W. M. Keck Foundation. PAIRITEL is operated by the Smithsonian Astrophysical Observatory (SAO) and supported by the Harvard University Milton Fund, the University of Virginia, SAO, UC Berkeley, and NASA via Swift Guest Investigator programs NNX09AQ66Q and NNX10A128G. We are grateful to the dedicated staffs at all of the observatories where we obtained data.

The National Energy Research Scientific Computing Center, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231, provided staff, computational resources, and data storage for this project.