0809.1881/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: %\usepackage{natbib}
4: \citestyle{aa}
5: %\bibliographystyle{apj}
6: %  \let\oldthebibliography=\thebibliography
7: %  \let\endoldthebibliography=\endthebibliography
8: %  \renewenvironment{thebibliography}[1]{%
9: %    \begin{oldthebibliography}{#1}%
10: %      \setlength{\parskip}{0ex}%
11: %      \setlength{\itemsep}{0ex}%
12: %  }%
13: %  {%
14: %    \end{oldthebibliography}%
15: %  }
16: 
17: 
18: 
19: %\setlength{\textwidth}{6.4in}
20: %\setlength{\textheight}{9.0in}
21: %\setlength{\oddsidemargin}{-15pt}
22: %\setlength{\topmargin}{-40pt}
23: 
24: 
25: \shorttitle{Progenitor of SN 2006my}\shortauthors{Leonard et al.}
26: 
27: \begin{document}
28: 
29: \title{An Upper Mass Limit on a Red Supergiant Progenitor for the Type
30:   II-Plateau Supernova SN 2006my\footnote{Some of the data presented herein
31:   were obtained at the W.M. Keck Observatory, which is operated as a scientific
32:   partnership among the California Institute of Technology, the University of
33:   California and the National Aeronautics and Space Administration. The
34:   Observatory was made possible by the generous financial support of the
35:   W.M. Keck Foundation.  Additional data were obtained from the data archive at
36:   the Space Telescope Science Institute. STScI is operated by the Association
37:   of Universities for Research in Astronomy, Inc., under NASA contract NAS
38:   5-26555. }}
39: 
40: %\vspace{2cm}
41: 
42: \author{Douglas C. Leonard\altaffilmark{1},
43: Avishay Gal-Yam\altaffilmark{2},
44: Derek B. Fox\altaffilmark{3},
45: P. B. Cameron\altaffilmark{4},
46: Erik M. Johansson\altaffilmark{5},
47: Adam L. Kraus\altaffilmark{4},
48: David Le Mignant\altaffilmark{5},
49: Marcos A. van Dam\altaffilmark{5} }
50: 
51: \altaffiltext{1}{Department of Astronomy, San Diego State University, San
52:   Diego, CA 92182-1221; leonard@sciences.sdsu.edu}
53: 
54: \altaffiltext{2}{Benoziyo Center for Astrophysics, Weizmann Institute of
55:   Science, 76100 Rehovot, Israel} 
56: 
57: \altaffiltext{3}{Department of Astronomy and Astrophysics, Pennsylvania
58: State University, 525 Davey Laboratory, University Park, PA 16802}
59: 
60: \altaffiltext{4}{Department of Astronomy, MS 105-24, California Institute of
61:   Technology, Pasadena, CA 91125}
62: 
63: \altaffiltext{5}{W. M. Keck Observatory, 65-1120 Mamalahoa Highway, Kamuela, HI
64:   96743}
65: 
66: 
67: %\vspace{1cm}
68: 
69: \begin{abstract}
70: 
71: We analyze two pre-supernova (SN) and three post-SN high-resolution images of
72: the site of the Type II-Plateau supernova SN~2006my in an effort to either
73: detect the progenitor star or to constrain its properties.  Following image
74: registration, we find that an isolated stellar object is not detected at the
75: location of SN~2006my in either of the two pre-SN images.  In the first, an
76: $I$-band image obtained with the Wide-Field and Planetary Camera 2 on board the
77: {\it Hubble Space Telescope}, the offset between the SN~2006my location and a
78: detected source (``Source 1'') is too large: $\geq 0.08^{\prime\prime}$, which
79: corresponds to a confidence level of non-association of $96\%$ from our most
80: liberal estimates of the transformation and measurement uncertainties.  In the
81: second, a similarly obtained $V$-band image, a source is detected (``Source
82: 2'') that has overlap with the SN~2006my location but is definitively an 
83: extended object.  Through artificial star tests carried out on the precise
84: location of SN~2006my in the images, we derive a $3\ \sigma$ upper bound on the
85: luminosity of a red supergiant that could have remained undetected in our
86: pre-SN images of $\log L/L_\odot = 5.10$, which translates to an upper bound on
87: such a star's initial mass of $15 {\rm\ M}_\odot$ from the STARS stellar
88: evolutionary models.  Although considered unlikely, we can not rule out the
89: possibility that {\it part} of the light comprising Source 1, which exhibits a
90: slight extension relative to other point sources in the image, or {\it part} of
91: the light contributing to the extended Source 2, may be due to the progenitor
92: of SN~2006my.  Only additional, high-resolution observations of the site taken
93: after SN~2006my has faded beyond detection can confirm or reject these
94: possibilities.
95: 
96: \end{abstract}
97: 
98: %\medskip
99: \keywords {supernovae: general --- supernovae: individual (SN 2006my) ---  stars: evolution}
100: 
101: \section{Introduction}
102: \label{sec:1}
103: 
104: The most common class of core-collapse supernovae (SNe) displays a distinct
105: plateau in its optical light curve, and is therefore dubbed Type II-Plateau
106: (II-P; see \citealt{Filippenko97} for a review of SN classifications).  This
107: type of stellar explosion has long been thought to result from the
108: core collapse and subsequent envelope ejection of isolated red supergiant (RSG)
109: stars, but it is only in recent years that direct observational evidence of the
110: progenitor-SN~II-P connection has begun to accumulate
111: \citep{Vandyk03,Smartt04,Maund05,Li06a,Eldridge07,Schawinski08}.
112: 
113: By registering pre-SN and post-SN images, usually taken at high resolution
114: using either space-based optical detectors, or ground-based infrared detectors
115: equipped with laser guide star adaptive optics systems, progenitor star
116: identifications have now been proposed for seven SNe~II-P (for a contemporary
117: review, see \citealt{Smartt08}, and references therein).  Although different in
118: detail, all seven proposed progenitors have properties consistent with those of
119: supergiant stars.  Because the field is still in its infancy --- at this point
120: none of the proposed progenitor objects has been definitively confirmed by
121: its absence in high-resolution images of the SN site taken after the SN has
122: faded beyond detection --- it is imperative to carefully examine every new
123: progenitor claim.  Here we investigate the status of the progenitor of
124: SN~2006my, for which \citet[][hereafter L07]{Li07} recently proposed the
125: identification of an RSG progenitor in pre-SN images and derived a zero-age
126: main-sequence mass for it of $M_{\rm ZAMS} = 10^{+5}_{-3} {\rm\ M}_\odot$.
127: 
128: Although discovered several months after explosion \citep{Stanishev06}, the
129: photometry and spectroscopy of SN~2006my presented by L07 clearly establish it
130: as an SN~II-P.  To identify the progenitor star, L07 register post-SN
131: ground-based optical images taken with the Canada-France-Hawaii Telescope in
132: the Sloan $r^\prime$ band under excellent seeing conditions (typical stellar
133: full width at half-maximum [FWHM] $= 0.6^{\prime\prime}$) with pre-SN {\it
134: Hubble Space Telescope} ({\it HST}) Wide-Field and Planetary Camera 2 (WFPC2)
135: images (typical stellar FWHM $= 0.15^{\prime\prime}$ on the WF2 chip, on which
136: the SN site is located), and identify a source in the pre-SN images within the
137: $1\ \sigma$ error circle estimated from the transformation uncertainty.  Here, we
138: reexamine this identification with the benefit of two new sets of
139: high-resolution, post-SN images: one taken with the wide-field channel of the
140: Near Infrared Camera 2 operated behind the laser guide star assisted adaptive
141: optics system (\citealt{Wizinowich06}) on the Keck II 10-m telescope (stellar
142: FWHM $= 0.10^{\prime\prime}$; note that SN~2006my served as its own tip-tilt
143: star for the observations), and the other taken with the {\it HST} WFPC2 camera
144: (with the SN centered on the PC chip; typical stellar FWHM $=
145: 0.08^{\prime\prime}$) as part of a study of the progenitors of core-collapse
146: supernovae (GO 10803; PI: Smartt).
147: 
148: This paper is organized as follows.  In \S~\ref{sec:2.1} we present the pre-SN
149: and post-SN images and details of the photometric measurements performed on
150: them; in \S~\ref{sec:2.2} we describe the image registration process along with
151: our conclusion that no isolated stellar source is detected at the location of
152: SN~2006my in either of the pre-SN images; in \S~\ref{sec:2.3} we estimate
153: detection limits in the pre-SN images; and in \S~\ref{sec:2.4} we derive an
154: upper mass limit on the progenitor of SN~2006my under the assumption that it
155: was a RSG.  We summarize our findings in \S~\ref{sec:3}.
156: 
157: \section{Data Analysis}
158: \label{sec:2}
159: 
160: \subsection{Pre-SN Image Photometry with HSTphot}
161: \label{sec:2.1}
162: 
163: Table~\ref{tab:1} lists information on the five sets of image data considered
164: by our study, hereafter referred to by the designation assigned to the final
165: combined image from each data set.  We first discuss our photometry of the two
166: pre-SN images, V1 and I1.  The individual frames comprising these datasets were
167: preprocessed through the standard Space Telescope Science Institute pipeline
168: using the latest calibrations as of 2008 July 23. The images were then further
169: processed following the procedure described by \citet{Leonard10}, which employs
170: the suite of programs designed specifically for the reduction of WFPC2 data
171: that are available as part of the HSTphot \citep{Dolphin00a} software package
172: (ver. 1.1.7b; our implementation includes the most recent update of 2008 July
173: 19).  The {\it hstphot} task automatically accounts for WFPC2 point-spread
174: function (PSF) variations, charge-transfer effects across the chips,
175: zeropoints, and aperture corrections.
176: 
177: We performed photometry on the final, combined images using the {\it hstphot}
178: task with option flag 10, which combines turning on local sky determination,
179: turning off empirically determined aperture corrections (using default values
180: instead), and turning on PSF residual determination. We ran {\it hstphot} on V1
181: and I1 individually with a signal-to-noise ratio (S/N) threshold of 1.0.
182: 
183: In addition to flight-system magnitudes and uncertainties, {\it hstphot}
184: returns several measurement parameters for each object detected.  For our
185: purposes, the most important of these are the ``object type'' and ``sharpness''
186: parameters.  To determine the object type, {\it hstphot} compares the
187: goodness-of-fit (i.e., the $\chi$ value; see \citealt{Dolphin00a}) of the
188: detected source's spatial profile with three different models: (1) the best
189: stellar profile (i.e., the library PSF + residuals); (2) a single ``hot'' pixel
190: without background; and (3) a completely flat profile.  If the stellar profile
191: provides the best match, the object is labeled a ``star'' and designated as
192: object Type `1', `2', or `3', depending on whether it is deemed a ``good
193: star'', a ``possible unresolved binary'', or a ``bad star' (a star centered on
194: a bad pixel), respectively.  If the single ``hot'' pixel provides the best
195: profile match, the object is labeled a ``single-pixel cosmic ray or hot pixel''
196: and designated as object Type `4'.  Finally, if a flat profile matches best,
197: the object is labeled an ``extended object'' and designated as object Type `5'.
198: \citet{Dolphin00a} notes that because this is, by design, a conservative test for
199: object type discrimination (i.e., it is a high threshold to be classified as
200: anything other than a good star), many nonstellar objects will still be
201: classified as Type 1 objects. For particular objects of interest, therefore,
202: examination of the ``sharpness'' parameter is recommended, where sharpness
203: values between -0.3 (object PSF broader than library PSF) and +0.3 (object PSF
204: narrower than library PSF) indicate confident point-source detections
205: \citep{Dolphin00a}.
206: 
207: Figures~\ref{fig:1}({\it a}) and \ref{fig:1}({\it b}) show an $\sim
208: 1^{\prime\prime}$ region surrounding the site of SN 2006my on pre-SN images I1
209: and V1 (respectively), following transformation of both images to the pixel
210: grid of image K1 (see \S~\ref{sec:2.2}). The SN 2006my site is rather complex,
211: and while common objects are evident in Figures~\ref{fig:1}({\it a}) and
212: \ref{fig:1}({\it b}), the immediate vicinity of SN~2006my appears quite
213: different in them.  In I1, an object of Type 1 (``good star'') is reported by
214: {\it hstphot} close to the SN location with a S/N of 5.6; we label this object
215: ``Source 1'' in Figure~\ref{fig:1}a.  In V1, an object of Type 5 (``extended
216: object'') is detected by {\it hstphot} with a S/N of 4.70; this extended source
217: appears to occupy a ``horseshoe-shaped'' region that includes the SN location.
218: We label this object ``Source 2'' in Figure~\ref{fig:1}b.  These two sources
219: appear to be distinct from one another in the combined images.
220: 
221: Source 1 is the object identified by L07 as the probable progenitor of
222: SN~2006my.  We confirm both the pixel location and photometry reported by L07
223: (also determined using HSTphot), with values measured by us (L07) of [x,y] =
224: [410.23, 158.59] ([410.22, 158.63]),\footnote{Note that all pixel coordinates
225: given in this paper are in the ``IRAF'' system --- the coordinate system
226: reported by using the {\it imexamine} task within IRAF or by running DAOPHOT
227: \citep{Stetson87} --- in which an integer value is assigned to a star that is
228: centered on the center of a pixel.  These pixel values are 0.5 greater in both
229: {\it x} and {\it y} than those reported by {\it hstphot}, which follows the
230: DoPHOT \citep{Schechter93} convention of assigning an integer value to a star
231: that is centered in the lower left corner of a pixel.} and flight-system
232: magnitude F814W = $24.48 \pm 0.19$ mag ($24.47 \pm 0.20$ mag), which
233: corresponds to $I = 24.46 \pm 0.19$ mag ($24.45 \pm 0.20$ mag) following
234: transformation according to the prescriptions of \citet{Holtzman95a} and
235: \citet{Dolphin00b}.  As noted by L07, although formally classified as an object
236: of Type 1 (``good star'') by {\it hstphot}, Source 1 exhibits a slight
237: east-west extension in the original WFPC2 image (see Fig.~\ref{fig:1}({\it a});
238: see also Figure 6 of L07). While L07 conclude that this object is most likely a
239: single star, we note that {\it hstphot} reports a ``sharpness'' value of
240: $-0.35$ for it, which places it beyond the limits suggested by
241: \citet{Dolphin00a} for confident point-source detections.  Source 2 is
242: unequivocally an extended source, as {\it hstphot} flags it as an object of
243: Type 5 and measures its profile to have a sharpness of $-0.58$, well beyond the
244: range for isolated star-like sources and for which {\it hstphot}'s PSF
245: star-fitting routines provide reliable photometry \citep{Dolphin02a}.  We note
246: that this object may be the ``extended source'' mentioned by L07 but not
247: further investigated due to its location near the $\sim 2\ \sigma$ error radius
248: of their derived SN pixel coordinates.  We shall return to further discuss both
249: of these ``sources of interest'' following a description of the image
250: registration process.
251: 
252: \subsection{Image Registration}
253: \label{sec:2.2}
254: 
255: To determine whether the progenitor star that exploded as SN~2006my is detected
256: in the pre-SN images, we first registered image I1 to image K1 by using the
257: IRAF tasks {\it geomap} and {\it geotran}, closely following the technique
258: described by \citet{Galyam05}.  To carry out the transformation, we used 16
259: common sources and obtained a final solution with a rms residual of 0.30 pixel
260: in {\it x} and 0.27 pixel in {\it y}, in the K1 pixel grid.  We then similarly
261: registered images V1, V2, and I2 to the transformed I1 image (using more than a
262: dozen common sources in each case), which resulted in five final images all
263: registered to the common K1 pixel grid.
264: 
265: Using the transformed images, we then measured the pixel locations of Source 1
266: (in image I1) and SN~2006my (in images K1, I2, and V2) using the centroiding
267: algorithm of the {\it imexamine} task within IRAF.\footnote{IRAF is distributed
268: by the National Optical Astronomy Observatories, which are operated by the
269: Association of Universities for Research in Astronomy, Inc., under cooperative
270: agreement with the National Science Foundation.}  The results returned by {\it
271: imexamine} depend somewhat on both the centering radius and fitting function
272: employed.  Thus, we applied a range of values --- centering radii of from 3--10
273: pixels, inclusive, using both Gaussian and Moffat profiles --- and took the
274: average as our ``best'' value and the measured scatter around the average as
275: the measurement uncertainty.  Due to the extended nature of Source 2 in image
276: V1, it was not possible to estimate a precise pixel location for it by using
277: the IRAF centroiding algorithm; instead, we ran the source-finding program
278: SExtractor \citep{Bertin96} on the image, which reported a nonstellar source
279: and pixel position at the location of Source 2.  The pixel locations,
280: measurement uncertainties, and transformation uncertainties of all objects are
281: given in Table~\ref{tab:2}.
282: 
283: Accounting for measurement and transformation uncertainties, we determine
284: offsets between Source 1 in image I1 and SN~2006my in images K1, I2, and V2, of
285: $0.083^{\prime\prime} \pm 0.017^{\prime\prime}$, $0.080^{\prime\prime} \pm
286: 0.016^{\prime\prime}$, and $0.086^{\prime\prime} \pm 0.016^{\prime\prime}$,
287: which correspond to separations significant at the $4.8\ \sigma$, $4.9\ \sigma$,
288: and $5.2\ \sigma$ levels, respectively, for a two-dimensional (i.e., both $x$
289: and $y$) Gaussian.  This indicates non-association at greater than the $99\%$
290: confidence level (see the Appendix for a full discussion of how significance of
291: source separation is determined), implying that the object detected as Source 1
292: is in all likelihood not the progenitor of SN~2006my.  Due to the extended
293: nature of Source 2, it is not possible to derive a similarly well-quantified
294: offset and uncertainty estimate between it and SN~2006my, but we note that from
295: the location reported by SExtractor, its formal separation is only
296: $0.024^{\prime\prime}$, and examination of Figure~\ref{fig:1} reveals that the
297: location of SN~2006my is indeed coincident with part of the extended region
298: identified as Source 2.
299: 
300: To serve as a check on the results obtained by transforming all images to the
301: K1 frame, we also performed a direct, ``HST-only'', registration between images
302: V2 and V1, and I2 and I1, using the {\it hstphot}-reported pixel positions
303: (with an additional correction for distortion using the solutions of
304: \citealt{Anderson03}) for all ($> 20$) objects used to determine the
305: transformation. Results of these registrations are given in Table~\ref{tab:3}.
306: To assign a measurement uncertainty on the {\it hstphot}-measured object pixel
307: positions, we consulted Figure~4 of \citet{Dolphin00a}, which presents
308: estimates of the $1\ \sigma$ astrometry error for sources detected by {\it
309: hstphot} as a function of count level.  For the count level of Source 1 ($\sim
310: 45$), a total astrometry error estimate of $\sim 0.4$ WF2 pixel, or $\sim 0.29$
311: pixels in both $x$ and $y$, is determined, while for SN~2006my (counts $>
312: 20,000$) position uncertainties of only 0.03 pixels in $x$ and $y$ are derived.
313: 
314: Accounting for measurement and transformation uncertainties, we determine the
315: separation between Source 1 in image I1 in the transformed location of
316: SN~2006my from image I2 to be $0.082^{\prime\prime} \pm 0.031^{\prime\prime}$,
317: which represents an offset of $2.6\ \sigma$ for a two-dimensional Gaussian,
318: indicating non-association at the $96\%$ confidence level (see the Appendix).
319: This separation is nearly identical to those determined in the K1 pixel frame,
320: although the significance of the offset is lower.  This is due mainly to the
321: larger uncertainty adopted for the pixel location of Source 1 in I1 than was
322: found empirically for the I1 image transformed to the K1 pixel grid.
323: Registration of I1 onto K1 resamples the data onto a finer grid and therefore
324: likely provides a better localization of Source 1, although, as noted by L07,
325: resampling the data has the risk of smoothing several (extended) sources into a
326: point source.  We also note that a direct registration of V2 to I1 (not listed
327: in Table~3) yields an offset between Source 1 and SN 2006my of
328: $0.092^{\prime\prime} \pm 0.032^{\prime\prime}$, which represents an offset of
329: $2.9\ \sigma$ for a two-dimensional Gaussian and indicates non-association at the
330: $98\%$ confidence level.  With all of these lines of evidence, we conclude that
331: a $96\%$ confidence level for non-association is a conservative estimate of the
332: significance of the offset between Source 1 and SN~2006my.
333: 
334: We note that in the original I1 WFPC2 image, Source 1 presents itself as
335: elevated flux in two pixels (i.e., pixels [411, 158] and [410, 159]; see
336: Figure~6 of L07), one of which contains our derived location of SN~2006my at
337: its far edge (Table~\ref{tab:3}).  The slight extension of Source 1 noted
338: earlier (\S~\ref{sec:2.1}) raises the possibility that this source may, in
339: fact, contain light from more than one object --- for instance, two RSG lying
340: in adjacent WF pixels, or perhaps a compact star cluster --- an idea considered
341: by L07 but deemed improbable.  This lingering possibility can only be
342: definitively removed from consideration by future high-resolution imaging after
343: the SN has faded beyond detection.  For now, we conclude that from our
344: astrometric measurements, an isolated stellar object is not detected at the
345: location of SN~2006my in either of the two pre-SN images: In image I1, the
346: offset between the SN~2006my location and the source detected by {\it hstphot}
347: is too large, and in V1 the source that overlaps with the SN~2006my site is
348: extended.  The relevant question that arises, then, is this: How faint a
349: point source at the precise location of SN~2006my would we have confidently
350: detected as a point source in the pre-SN images?
351: 
352: \subsection{Detection Limits in Pre-SN Images}
353: \label{sec:2.3}
354: 
355: To set accurate detection limits on a point source at the location of SN~2006my
356: in the pre-SN images, we proceeded as follows.  First, we used the {\it
357: showpsf}\ task within HSTphot to produce the library PSFs appropriate for point
358: sources at the precise pixel locations (accurate to 0.1 pixel) of SN~2006my in
359: the V1 and I1 images.  Using these PSFs, we then inserted artificial stars of
360: known flux (corresponding to $21.5 < V < 27.0$ and $20.5 < I < 26.0$) at the SN
361: locations in the V1 and I1 images.  Then, we ran {\it hstphot} on these images
362: in exactly the same manner as we did when seeking a progenitor star in the
363: original images.
364: 
365: To be considered a confident ``detection'' of a single star at the SN~2006my
366: location, we demanded that the object be classified by {\it hstphot} as a
367: ``good star'' (Type 1), have a sharpness between -0.3 and +0.3, and have a
368: reported pixel location no more than $1\ \sigma$ away from the nominal SN
369: location, where the uncertainty accounts for the measurement and transformation
370: uncertainties listed in Table~\ref{tab:3} as well as the astrometric
371: uncertainty appropriate for the object from Figure~4 of \citet{Dolphin00a}.
372: 
373: From this analysis, we derive detection limits of $V = 24.9 \pm 0.3$ mag and $I
374: = 24.4 \pm 0.2$ mag for point sources at the location of SN~2006my in the V1
375: and I1 images, respectively.  In both cases the limiting magnitude is set by
376: the sharpness parameter becoming less than -0.3 (i.e., source too extended to
377: be confidently considered a point source).  We note that these detection limits
378: are significantly shallower than those derived by L07, in which detection
379: limits were derived by examining the magnitudes of all $3\ \sigma$ detections in
380: the images.  Because we are specifically interested in the point-source
381: detection limits at the location of the SN, we consider our (less restrictive)
382: limits to more accurately reflect the relevant detection threshold.
383: 
384: \subsection{Properties of the Progenitor Star}
385: \label{sec:2.4}
386: 
387: To convert our detection thresholds to constraints on the initial mass of an
388: RSG progenitor star that could have exploded as SN~2006my, we employ the
389: metallicity-dependent stellar models produced with the Cambridge stellar
390: evolution code, STARS, the descendant of the code developed originally by
391: \citet{Eggleton71} and updated most recently by \citet[][see also
392: \citealt{Smartt08}, and references therein]{Eldridge04}.\footnote{The models
393: were downloaded from the code's Web site, at
394: http://www.ast.cam.ac.uk/$\sim$stars .}  The models follow stellar evolution up
395: to the initiation of core neon burning, which is likely to give an accurate
396: indication of the pre-SN luminosity; comparisons with other contemporary model
397: grids (i.e., \citealt{Heger00}; \citealt{Meynet00}) show that the endpoints for
398: stars in the $8 \rightarrow 15 {\rm\ M}_\odot$ range differ by at most $0.2$
399: dex in luminosity among the codes \citep{Smartt04}.
400: 
401: Using the metallicity and radial gradient in NGC 4651 published by
402: \citet{Pilyugin04}, L07 derive a metallicity at the SN~2006my location of
403: $\log(O/H) + 12 = 8.51 \pm 0.06$, which is subsolar according to the recent
404: analysis of \citet{Asplund04}, who found $[O/H] = 8.66 \pm 0.05$.  We thus
405: use the $Z = 0.01$ (the closest metallicity calculated) STARS stellar evolution
406: models as the basis for deriving the upper mass limit for our study.
407: Figure~\ref{fig:2} displays the final predicted luminosity for stars with
408: masses between 8 and 20 M$_\odot$ from these models.
409: 
410: We now seek to determine the lowest possible luminosity that an RSG could have
411: had and still have been confidently detected by our analysis of the pre-SN
412: images.  Since it will prove to be far more restrictive, we begin by
413: considering the I1 image, and calculate the bolometric magnitude of RSG stars
414: corresponding to our detection limit through the equation:
415: $$M_{\rm bol} = -\mu - A_V + I + (V - I)_{\rm RSG} + {\rm BC}_V,$$
416: \noindent where $\mu$ is the distance modulus of NGC 4651, $A_V$ the extinction
417: to SN~2006my, $I$ the $I$-band detection threshold, $(V - I)_{\rm RSG}$ the
418: color range of RSG stars (i.e., spectral types ${\rm K3} \rightarrow {\rm
419: M4}$), and ${\rm BC}_V$ the bolometric correction corresponding to each $(V -
420: I)_{\rm RSG}$.  As detailed by L07, distance estimates to NGC 4651 span quite a
421: wide range.  Because it is our goal here to set the most conservative lower
422: bound on our detection threshold, we adopt the long distance estimate, $\mu =
423: 31.74 \pm 0.25$ mag, found by \citet{Solanes02} by averaging seven different
424: Tully-Fisher distances to the galaxy.  For the extinction, we adopt $A_V = 0.08
425: \pm 0.02$ mag, which represents the Galactic value along the line of sight,
426: because there is no evidence for host-galaxy extinction (L07).  For the color
427: and bolometric corrections appropriate for RSG stars, we consult the values
428: reported by \citet{Elias85}, who find that the quantity $[(V - I)_{\rm RSG} +
429: {\rm BC}_V]$ lies in the remarkably tight range $0.88 \rightarrow 1.0$ for
430: supergiant stars of spectral types K3 -- M4 (i.e., RSG).  Again, because we
431: wish to set the most conservative lower detection limit, we assign $(V -
432: I)_{\rm RSG} + {\rm BC}_V = 0.88$, which is the value obtained for an M4
433: supergiant star.  Finally, we set $I = 24.4 \pm 0.2$ mag, as derived in
434: \S~\ref{sec:2.3}.
435: 
436: With these values, we derive $M_{\rm bol} = -6.54 \pm 0.32$ mag as the limiting
437: bolometric magnitude, above which any RSG would have been detected in our
438: pre-SN image.  This corresponds to a $3\ \sigma$ detection threshold of $M_{\rm
439: bol} = -7.50$ mag, which translates to $\log L/L_\odot = 4.90$.  If we adopt a
440: maximum systematic uncertainty of $0.2$ dex in the theoretical stellar model
441: endpoints, then the final $3\ \sigma$ lower bound on the luminosity of a RSG
442: that would have been confidently detected in our pre-SN image is $\log
443: L/L_\odot = 5.10$.  From Figure~\ref{fig:2} this corresponds to an upper bound
444: on the progenitor mass of $M_{\rm ZAMS} = 15 {\rm\ M}_\odot$, and we therefore
445: conclude that any RSG progenitor with an initial mass greater than this value
446: would have been detected using our analysis procedure.
447: 
448: Applying a similar analysis to the V1 image results in a detection threshold of
449: $\log L/L_\odot = 6.30$, which unfortunately does not rule out any progenitors
450: up to $200 {\rm\ M}_\odot$, the highest progenitor mass considered by the STARS
451: models.  Our most restrictive limit thus comes from the I1 image.
452: 
453: \section{Conclusions}
454: \label{sec:3}
455: 
456: We analyze two pre-SN and three new post-SN high-resolution images of the site
457: of the Type II-Plateau supernova SN~2006my in an effort to either detect the
458: progenitor star or to constrain its properties.  Our primary result is that we
459: do not detect an isolated stellar object at the location of SN~2006my in either
460: of the two pre-SN images.  From our image registration, we therefore do not
461: confirm the association found by L07 between a stellar source (Source 1) in
462: pre-SN $I$-band images (I1) and the location of SN~2006my.  Using new
463: high-resolution post-SN images, we derive an offset between SN~2006my and
464: Source 1 of $\geq 0.08^{\prime\prime}$ from the SN location, which represents a
465: confidence level of non-association of more than $96\%$ from our most liberal
466: estimates of the image transformation and measurement uncertainties.  Through
467: artificial star tests carried out at the precise location of SN~2006my in image
468: I1, we derive a $3\ \sigma$ upper bound on the mass of the progenitor of
469: SN~2006my of $M_{\rm ZAMS} = 15 {\rm\ M}_\odot$ from the STARS stellar
470: evolutionary models.
471: 
472: Although considered unlikely, we can not rule out the possibility that {\it
473: part} of the light comprising Source 1, which exhibits some extension relative
474: to other point sources in image I1, or {\it part} of the light contributing to
475: Source 2, a definitively extended source detected in pre-SN $V$-band images
476: (V1) that has overlap with the SN~2006my location, may be due to the progenitor
477: of SN~2006my.  Only additional, high-resolution observations of the site taken
478: after SN~2006my has faded beyond detection can confirm or reject these
479: possibilities.
480: 
481: 
482: \acknowledgments
483: 
484: We thank an anonymous referee for helpful comments that improved the
485: manuscript.  A.G. acknowledges support by the Benoziyo Center for Astrophysics,
486: a research grant from Peter and Patricia Gruber Awards, and the William Z. and
487: Eda Bess Novick New Scientists Fund at the Weizmann Institute.  A.L.K. is
488: supported by a NASA Origins grant to L. Hillenbrand.  This research has made
489: use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet
490: Propulsion Laboratory, California Institute of Technology, under contract with
491: NASA.  We wish to recognize and acknowledge the very significant cultural role
492: and reverence that the summit of Mauna Kea has always had within the indigenous
493: Hawaiian community.  We are most fortunate to have the opportunity to conduct
494: observations from this mountain.
495: 
496: \clearpage
497: 
498: \appendix
499: \section{APPENDIX}
500: \begin{center}
501: \bf{ON THE CONFIDENCE LEVEL OF NON-ASSOCIATION OF POINT SOURCES}
502: \end{center}
503: 
504: Determining the significance of positional offsets between objects identified
505: in two images is of paramount importance when assessing the potential
506: association between a progenitor star (in a pre-explosion image) and a
507: supernova (in a post-explosion image), where one image has been transformed
508: into the pixel frame of the other.  Six sources of uncertainty are typically
509: identified: Positional uncertainty of the putative progenitor star (in both $x$
510: and $y$), positional uncertainty of the SN (in both $x$ and $y$), and the
511: uncertainty in the transformation (also in both $x$ and $y$).  In much of the
512: past work involving SN progenitors (hereafter, the ``traditional'' approach),
513: offset significance (i.e., how many ``sigma'' away the two objects are) has
514: been derived by taking the measured radial offset between the two objects, and
515: then dividing this value by the quadrature sum of all of the uncertainties.  A
516: formal analysis, however, finds this approach to be somewhat in error.
517: 
518: Consider the two-dimensional Gaussian that describes the uncertainty in an SN's
519: measured position ($x, y$) compared with that of a putative progenitor star
520: ($x_0, y_0$):
521: 
522: $$p(x,y)\, dx\, dy\, = \frac{1}{2\pi\sigma_x\sigma_y}e^{-(x -
523:   x_0)^2/2\sigma_x^2}e^{-(y - y_0)^2/2\sigma_y^2}\,dx\,dy,$$
524: 
525: \noindent where $\sigma_{x(y)}$ represents the quadrature sum of the
526: measurement and transformation errors in $x(y)$.  Converting this to a
527: normalized form by defining $\Sigma_x \equiv (x - x_0)/\sigma_x, \Sigma_y
528: \equiv (y - y_0)/\sigma_y$, yields:
529: 
530: $$p(\Sigma_x,\Sigma_y)\,d\Sigma_x\,d\Sigma_y =
531: \frac{1}{2\pi\sigma_x\sigma_y}e^{-(\Sigma_x^2/2)}e^{-(\Sigma_y^2/2)}(\sigma_x\,d\Sigma_x)(\sigma_y\,d\Sigma_y),$$
532: 
533: \noindent which becomes the expected:
534: 
535: $$p(\Sigma_x,\Sigma_y)\,d\Sigma_x\,d\Sigma_y =
536: \frac{1}{2\pi}e^{-(\Sigma_x^2/2)}e^{-(\Sigma_y^2/2)}d\Sigma_x\,d\Sigma_y.$$
537: 
538: \noindent We convert this normalized Gaussian into radial form with a further
539: change of coordinates ($\Sigma_x \equiv \rho \cos \theta, \Sigma_y \equiv \rho
540: \sin \theta$) and a little algebra to yield:
541: 
542: $$p(\rho,\theta)\,d\rho\,d\theta = \frac{1}{2\pi}
543: e^{-\rho^2/2}\rho\,d\rho\,d\theta,$$
544: 
545: \noindent or ignoring the angular part of the distribution,
546: 
547: $$p(\rho)\,d\rho = \rho e^{-\rho^2/2}\,d\rho.$$
548: 
549: \noindent The total integrated probability for $\rho < Q$ (where $Q$ is equal
550: to, e.g., $[\Sigma_x^2 + \Sigma_y^2]^{1/2}$), then, is
551: 
552: \begin{equation}
553: p(\rho < Q) = 1 - e^{-(Q^2/2)}.
554: \label{eqn:a}
555: \end{equation}
556: 
557: A simple example serves to highlight the differences between calculating offset
558: significance in the ``traditional'' manner versus calculating it according to
559: equation~(\ref{eqn:a}).  Suppose a progenitor star candidate and a SN are
560: measured to be located at pixel locations [10,10] and [11,11] on pre-SN and
561: (transformed) post-SN images, respectively, with total uncertainties
562: (quadrature sum of all measurement and transformation uncertainties) of
563: $\sigma_x = 1$, $\sigma_y = 1$.  In the traditional approach, this would
564: represent an offset of $1.41 \pm 1.41$ pixels, or a separation significant at
565: the $1\sigma$ level, which implies a confidence level of non-association of
566: $68\%$.
567: 
568: However, if we calculate the total integrated probability for $\rho < 1.41$
569: according to equation~(\ref{eqn:a}) we find $p(\rho < 1.41) = 1 -
570: \exp[-(1.41^2/2)] = 0.63$, which implies a confidence level of non-association
571: of $63\%$.  Because we are dealing with positions and uncertainties in two
572: dimensions, it thus appears most precise to state the result as follows: ``The
573: two objects are located $1.41\sigma$ away from each other for a two-dimensional
574: Gaussian, which indicates non-association at the $63\%$ confidence level.''  We
575: therefore express all of our association significances in this paper in this
576: manner.
577: 
578: 
579: %\bibliography{/Users/leonard/misc/all_refs}
580: 
581: \begin{thebibliography}{29}
582: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
583: 
584: \bibitem[{{Anderson} \& {King}(2003)}]{Anderson03} 
585: {Anderson}, J., \& {King}, I.~R.  2003, \pasp, 115, 113 
586: 
587: \bibitem[{{Asplund} {et~al.}(2004)}]{Asplund04} 
588: {Asplund}, M., {Grevesse}, N., {Sauval}, A.~J., {Allende Prieto}, C., \&  
589: {Kiselman}, D.  2004, \aap, 417, 751 
590: 
591: \bibitem[{{Bertin} \& {Arnouts}(1996)}]{Bertin96} 
592: {Bertin}, E., \& {Arnouts}, S.  1996, \aaps, 117, 393 
593: 
594: \bibitem[{{Dolphin}(2000{\natexlab{a}})}]{Dolphin00b} 
595: {Dolphin}, A.~E.  2000{\natexlab{a}}, \pasp, 112, 1397 
596: 
597: \bibitem[{{Dolphin}(2000{\natexlab{b}})}]{Dolphin00a} 
598: ---. 2000{\natexlab{b}}, \pasp, 112, 1383 
599: 
600: \bibitem[{{Dolphin} \& {Kennicutt}(2002)}]{Dolphin02a} 
601: {Dolphin}, A.~E., \& {Kennicutt}, Jr. 2002, \aj, 123, 207 
602: 
603: \bibitem[{{Eggleton}(1971)}]{Eggleton71} 
604: {Eggleton}, P.~P.  1971, \mnras, 151, 351 
605: 
606: \bibitem[{{Eldridge} {et~al.}(2007)}]{Eldridge07} 
607: {Eldridge}, J.~J., {Mattila}, S., \& {Smartt}, S.~J.  2007, \mnras, 376, 
608: L52 
609: 
610: \bibitem[{{Eldridge} \& {Tout}(2004)}]{Eldridge04} 
611: {Eldridge}, J.~J., \& {Tout}, C.~A.  2004, \mnras, 348, 201 
612: 
613: \bibitem[{{Elias} {et~al.}(1985)}]{Elias85} 
614: {Elias}, J.~H., {Frogel}, J.~A., \& {Humphreys}, R.~M.  1985, \apjs, 57, 91 
615: 
616: \bibitem[{{Filippenko}(1997)}]{Filippenko97} 
617: {Filippenko}, A.~V.  1997, \araa, 35, 309 
618: 
619: \bibitem[{{Gal-Yam} {et~al.}(2005)}]{Galyam05} 
620: {Gal-Yam}, A., {et~al.} 2005, \apjl, 630, L29 
621: 
622: \bibitem[{{Heger} \& {Langer}(2000)}]{Heger00} 
623: {Heger}, A., \& {Langer}, N.  2000, \apj, 544, 1016 
624: 
625: \bibitem[{{Holtzman} {et~al.}(1995)}]{Holtzman95a} 
626: {Holtzman}, J.~A., {Burrows}, C.~J., {Casertano}, S., {Hester}, J.~J.,  
627: {Trauger}, J.~T., {Watson}, A.~M., \& {Worthey}, G.  1995, \pasp, 107, 1065 
628: 
629: \bibitem[{{Leonard} {et~al.}(2003)}]{Leonard10} 
630: {Leonard}, D.~C., {Kanbur}, S.~M., {Ngeow}, C.~C., \& {Tanvir}, N.~R.  
631: 2003,  \apj, 594, 247 
632: 
633: \bibitem[{{Li} {et~al.}(2006)}]{Li06a} 
634: {Li}, W., {Van Dyk}, S.~D., {Filippenko}, A.~V., {Cuillandre}, J.-C., 
635: {Jha},  S., {Bloom}, J.~S., {Riess}, A.~G., \& {Livio}, M.  2006, \apj, 
636: 641, 1060 
637: 
638: \bibitem[{{Li} {et~al.}(2007)}]{Li07} 
639: {Li}, W., {Wang}, X., {Van Dyk}, S.~D., {Cuillandre}, J.-C., {Foley}, 
640: R.~J., \&  {Filippenko}, A.~V.  2007, \apj, 661, 1013 (L07)
641: 
642: \bibitem[{{Maund} {et~al.}(2005)}]{Maund05} 
643: {Maund}, J.~R., {Smartt}, S.~J., \& {Danziger}, I.~J.  2005, \mnras, 364, 
644: L33 
645: 
646: \bibitem[{{Meynet} \& {Maeder}(2000)}]{Meynet00} 
647: {Meynet}, G., \& {Maeder}, A.  2000, \aap, 361, 101 
648: 
649: \bibitem[{{Nakano} \& {Itagaki}(2006)}]{Nakano06} 
650: {Nakano}, S., \& {Itagaki}, K.  2006, CBET, 756, 1 
651: 
652: \bibitem[{{Pilyugin} {et~al.}(2004)}]{Pilyugin04} 
653: {Pilyugin}, L.~S., {V{\'{\i}}lchez}, J.~M., \& {Contini}, T.  2004, \aap, 
654: 425,  849 
655: 
656: \bibitem[{{Schawinski} {et~al.}(2008)}]{Schawinski08} 
657: {Schawinski}, K., {et~al.} 2008, Science, 321, 223
658: 
659: \bibitem[{{Schechter} {et~al.}(1993)}]{Schechter93} 
660: {Schechter}, P.~L., {Mateo}, M., \& {Saha}, A.  1993, \pasp, 105, 1342 
661: 
662: \bibitem[{{Smartt} {et~al.}(2008)}]{Smartt08} 
663: {Smartt}, S.~J., {Eldridge}, J.~J., {Crockett}, R.~M., \& {Maund}, J.~R.  
664: 2008,  \mnras, submitted (ArXiv e-prints 0809.0403)
665: 
666: \bibitem[{{Smartt} {et~al.}(2004)}]{Smartt04} 
667: {Smartt}, S.~J., {Maund}, J.~R., {Hendry}, M.~A., {Tout}, C.~A., {Gilmore}, 
668:  G.~F., {Mattila}, S., \& {Benn}, C.~R.  2004, Science, 303, 499 
669: 
670: \bibitem[{{Solanes} {et~al.}(2002)}]{Solanes02} 
671: {Solanes}, J.~M., {Sanchis}, T., {Salvador-Sol{\'e}}, E., {Giovanelli}, R., 
672: \&  {Haynes}, M.~P.  2002, \aj, 124, 2440 
673: 
674: \bibitem[{{Stanishev} \& {Nielsen}(2005)}]{Stanishev06} 
675: {Stanishev}, V., \& {Nielsen}, T.~B.  2006, CBET, 737, 1 
676: 
677: \bibitem[{{Stetson}(1987)}]{Stetson87} 
678: {Stetson}, P.~B.  1987, \pasp, 99, 191 
679: 
680: \bibitem[{{Van Dyk} {et~al.}(2003)}]{Vandyk03} 
681: {Van Dyk}, S.~D., {Li}, W., \& {Filippenko}, A.~V.  2003, \pasp, 115, 1289 
682: 
683: \bibitem[{{Wizinowich} {et~al.}(2006)}]{Wizinowich06} 
684: {Wizinowich}, P.~L., {et~al.} 2006, \pasp, 118, 297 
685: 
686: \end{thebibliography}
687: 
688: \clearpage
689: 
690: \input{tab1.tex}
691: \input{tab2.tex}
692: \input{tab3.tex}
693: 
694: \clearpage
695: 
696: \begin{figure}
697: %\rotatebox{90}{
698: \scalebox{1.1}{
699: \plotone{f1.eps} 
700:               }
701: %\vskip -0.2in
702: \caption{Site of SN~2006my in pre-SN images I1 (panel {\it a}; all image
703:   designations are as given in Table~\ref{tab:1}) and V1 (panel {\it b}), and
704:   post-SN image I2 (panel {\it c}).  All images have been transformed and
705:   resampled to the pixel grid of the K1 image (\S~\ref{sec:2.2}).  The cyan
706:   circles indicate the approximate $5\sigma$ uncertainty (for a two-dimensional
707:   Gaussian; see the Appendix) of the position of SN~2006my relative to the
708:   transformed I1 image as measured in the transformed I2 image.  The white
709:   circles indicate the same level of uncertainty in the location of SN~2006my
710:   as measured in the K1 image relative to the transformed I1 image.  Two
711:   ``sources of interest,'' discussed in the text, are labeled Source 1 and
712:   Source 2 in panels {\it a} and {\it b}, respectively.
713: \label{fig:1} }
714: \end{figure}
715: 
716: \clearpage
717: 
718: \begin{figure}
719: %\rotatebox{90}{
720: \begin{center}
721: \scalebox{0.8}{
722: \plotone{f2.eps} 
723:               }
724: \end{center}
725: \vskip 0.1in
726: \caption{Initial mass vs. final predicted luminosity prior to explosion for
727:   $Z = 0.01$ stars evolved with the STARS stellar evolution code
728:   \citep{Eldridge04}.  The dashed line indicates the $3\sigma$ upper luminosity
729:   limit for a RSG that could have remained undetected by our analysis of pre-SN
730:   images of the site of SN~2006my.
731: \label{fig:2} }
732: \end{figure}
733: 
734: 
735: 
736: \end{document}
737: 
738: