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: \newcommand{\kms}{km s$^{-1}$}
20: \newcommand{\kmsmpc}{{\rm km\ s\ }^{-1}{\rm\ Mpc}^{-1}}
21: \newcommand{\Kms}{{\rm km\ s}^{-1}}
22: \newcommand{\Bband}{B}
23: \newcommand{\Vband}{V}
24: \newcommand{\BminusV}{({\Bband}{\rm -}{\Vband})}
25: \newcommand{\bminusv}{[{\Bband}{\rm -}{\Vband}]}
26: \newcommand{\Bminusv}{$({\Bband}{\rm -}{\Vband})$}
27: \newcommand{\EBV}{E\bminusv}
28: \newcommand{\Ebv}{E\BminusV}
29: \newcommand{\EBv}{E\Bminusv}
30: \newcommand{\ebv}{$E\BminusV$}
31: \newcommand{\halpha}{H$\alpha$}
32: \newcommand{\fetwo}{\protect\ion{Fe}{2}\ $\lambda\lambda4924,~5018,~5169$}
33: \newcommand{\hbeta}{H$\beta$}
34: \newcommand{\hgamma}{H$\gamma$}
35: \newcommand{\hdelta}{H$\delta$}
36: \newcommand{\ubvri}{\protect\hbox{$U\!BV\!RI$}}
37: \newcommand{\bvri}{\protect\hbox{$BV\!RI$}}
38: \newcommand{\bvi}{\protect\hbox{$BV\!I$}}
39: \newcommand{\vri}{\protect\hbox{$V\!RI$}}
40: \newcommand{\ubv}{\protect\hbox{$U\!BV$}}
41: \newcommand{\ssp}{\def\baselinestretch{1.0}\large\normalsize}
42: \newcommand{\gtrsi}{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
43: \newcommand{\lesssi}{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
44: %\newcommand{\arcdeg}{\ensuremath{^{\circ}}}
45: \newcommand{\vi}{\mbox{$V\!-\!I$}}
46:
47: \slugcomment{Received: 2007 July 21; Accepted: 2007 August 10}
48:
49: \setlength{\textwidth}{6.4in}
50: \setlength{\textheight}{9.0in}
51: \setlength{\oddsidemargin}{-15pt}
52: \setlength{\topmargin}{-40pt}
53:
54:
55: \shorttitle{Constraining the Type Ia Supernova Progenitor} \shortauthors{Leonard}
56:
57:
58:
59: \received{2007 July 21}
60: \begin{document}
61:
62: \slugcomment{Submitted: 2007 July 21; Accepted: 2007 August 10}
63:
64: \title{Constraining the Type Ia Supernova Progenitor: The Search for Hydrogen
65: in Nebular Spectra\footnote{Some of the data presented herein were obtained
66: at the W.M. Keck Observatory, which is operated as a scientific partnership
67: among the California Institute of Technology, the University of California
68: and the National Aeronautics and Space Administration. The Observatory was
69: made possible by the generous financial support of the W.M. Keck Foundation.
70: Additional observations were obtained at the Gemini Observatory, which is
71: operated by the Association of Universities for Research in Astronomy, Inc.,
72: under a cooperative agreement with the NSF on behalf of the Gemini
73: partnership: the National Science Foundation (United States), the Particle
74: Physics and Astronomy Research Council (United Kingdom), the National
75: Research Council (Canada), CONICYT (Chile), the Australian Research Council
76: (Australia), CNPq (Brazil) and CONICET (Argentina).}}
77:
78: \vspace{2cm}
79:
80: \author{Douglas C. Leonard}
81: \affil{Department of Astronomy, San Diego State University, San Diego, CA 92182-1221}
82: \email{leonard@sciences.sdsu.edu}
83:
84: \vspace{1cm}
85:
86: \begin{abstract}
87:
88: Despite intense scrutiny, the progenitor system(s) that gives rise to Type Ia
89: supernovae remains unknown. The favored theory invokes a carbon-oxygen white
90: dwarf accreting hydrogen-rich material from a close companion until a
91: thermonuclear runaway ensues that incinerates the white dwarf. However,
92: simulations resulting from this single-degenerate, binary channel demand the
93: presence of low-velocity H$\alpha$ emission in spectra taken during the late
94: nebular phase, since a portion of the companion's envelope becomes entrained in
95: the ejecta. This hydrogen has never been detected, but has only rarely been
96: sought. Here we present results from a campaign to obtain deep, nebular-phase
97: spectroscopy of nearby Type Ia supernovae, and include multi-epoch observations
98: of two events: SN~2005am (slightly subluminous) and SN~2005cf (normally
99: bright). No H$\alpha$ emission is detected in the spectra of either object.
100: An upper limit of $0.01\ M_\odot$ of solar abundance material in the ejecta is
101: established from the models of \citet{Mattila05} which, when coupled with the
102: mass-stripping simulations of \citet{Marietta00} and \citet{Meng07},
103: effectively rules out progenitor systems for these supernovae with secondaries
104: close enough to the white dwarf to be experiencing Roche lobe overflow at the
105: time of explosion. Alternative explanations for the absence of H$\alpha$
106: emission, along with suggestions for future investigations necessary to
107: confidently exclude them as possibilities, are critically evaluated.
108:
109: \end{abstract}
110:
111: \medskip
112: \keywords {binaries: symbiotic --- circumstellar matter --- supernovae: general
113: --- supernovae: individual (SN 2005am, SN 2005cf) --- white dwarfs}
114:
115: \section{Introduction}
116: \label{sec:1}
117:
118: Ever since the currently favored single-degenerate, binary channel was proposed
119: as the progenitor system for Type Ia supernovae \citep{Whelan73}, all models of
120: the impact of the exploded white dwarf (WD) on the secondary star have
121: indicated that significant amounts of solar-abundance material, stripped from
122: the secondary's envelope, become entrained in the ejecta
123: \citep{Wheeler75,Fryxell81,Taam84,Chugai86,Livne92,Marietta00,Meng07}.
124: Observational evidence for this material, however, still eludes us
125: \citep{Mattila05}, and serves as one reminder among many that we still lack
126: direct observational proof for the single-degenerate scenario
127: \citep{Branch95,Livio01}.
128:
129: The most detailed theoretical investigation of the expected amount and
130: distribution of stripped material within a young Type Ia supernova (SN~Ia)
131: remnant is that of \citet{Marietta00}, who studied the problem with
132: two-dimensional numerical simulations. Four basic progenitor systems were
133: investigated, including three with secondaries (a main-sequence, subgiant, or
134: red giant star) close enough for mass-transfer to occur through Roche lobe
135: overflow (RLOF), and one containing a secondary (a red giant) donating material
136: through a strong stellar wind --- the symbiotic case. All secondaries were
137: given masses of $\sim 1\ M_\odot$ at the time of the explosion and were placed
138: either just within (the RLOF cases), or just beyond (the symbiotic case), the
139: limiting distance from the WD within which RLOF can occur (i.e., $a/R = 3$,
140: where $a$ is the orbital separation in units of the secondary star's radius,
141: $R$; see \citealt{Eggleton83}). Three additional systems, in which a
142: main-sequence secondary was placed too far away to experience RLOF (thus
143: rendering it an unlikely progenitor system) were also included to establish the
144: scaling between orbital separation and amount of stripped material.
145:
146: The numerical results of the \citet{Marietta00} study confirmed predictions
147: from earlier analytic work \citep[e.g.,][]{Wheeler75,Chugai86} that substantial
148: material is indeed stripped, with the amount ranging from a minimum of $0.15\
149: M_\odot$ for a close ($a/R = 3$) main-sequence secondary, up to nearly the
150: entire envelope ($\sim 0.5\ M_\odot$) for a similarly placed red giant.
151: Increasing the orbital separation beyond the RLOF limit in the
152: \citet{Marietta00} simulations resulted in a dramatic decrease in the amount of
153: stripped material: For $a/R = 12$, a main-sequence secondary loses only
154: $0.0018\ M_\odot$. However, since such systems lack an efficient mechanism for
155: mass transfer, they are not considered to be viable SN~Ia progenitors. A red
156: giant secondary placed at a similarly distant location, however, could
157: potentially donate mass through a strong stellar wind, making the symbiotic
158: case a possibility if large orbital separations are required
159: \citep[e.g.,][]{Munari92}.
160:
161: As discussed by \citet{Meng07}, one shortcoming of the \citet{Marietta00} study
162: is their use of standard solar-model stars for the companion, rather than
163: companions whose structures have been appropriately modified due to having
164: evolved in a binary system \citep[e.g.,][]{Eggleton73}. To investigate the
165: effect this simplification might have had on the results, \citet{Meng07} use an
166: analytic model to estimate the amount of mass expected to be stripped from a
167: variety of evolved secondaries. (Their analytic approach was first tested
168: using the unevolved secondaries used by \citealt{Marietta00}, and was
169: demonstrated to approximate the results obtained numerically.) The result is
170: that the quantity of material expected to be stripped from evolved secondaries
171: is considerably lower than that predicted for standard solar-model companions.
172: In fact, \citet{Meng07} find that the minimum value for systems experiencing
173: RLOF is diminished from $0.15\ M_\odot$ to only $0.035\ M_\odot$. The
174: reduction arises primarily from the pre-explosion mass-loss producing a more
175: compact companion star whose material is more difficult to strip than it is in
176: the unevolved case. \citet{Meng07} stress, however, that their new values are
177: really lower bounds on the amount of stripped material, since their analytic
178: approach does not consider the thermal energy imparted by the ejecta to the
179: companion envelope, which likely serves to heat and vaporize a portion of it
180: and thereby increase the amount of stripped material
181: \citep[e.g.,][]{Fryxell81,Mattila05}. Thus, $0.035\ M_\odot$ serves as a
182: conservative lower bound on the expected amount of stripped material resulting
183: from their models.
184:
185: The typical velocity of the stripped material is found in all studies to be far
186: slower than the $\sim 10,000$~\kms\ velocity that characterizes the bulk of the
187: iron-rich ejecta \citep{Chugai86,Marietta00,Meng07}. This has the effect of
188: placing it almost entirely in the central region of the supernova remnant, with
189: the majority of it predicted to be moving with a velocity of under
190: $1,000$~\kms\ \citep{Marietta00}. The stripped material is largely confined to
191: the downstream region behind the companion star, where it contaminates a solid
192: angle that ranges from $66^\circ$ for the main-sequence companion to
193: $115^\circ$ for the red-giant companion \citep{Marietta00}.
194:
195: The low velocity of the stripped material renders it undetectable when the
196: faster-moving, iron-rich ejecta are optically thick. However, detailed
197: radiative transfer calculations performed by \cite{Mattila05} predict that it
198: should become visible at late times, when the outer ejecta have thinned out and
199: become transparent enough to reveal the slower-moving gas in the central
200: regions. The most prominent expected spectral signature of the companion
201: star's stripped material is narrow H$\alpha$ emission in nebular spectra
202: \citep{Mattila05}, taken more than $\sim 250$ days after maximum light. The
203: H$\alpha$ emission should be present within $\pm 1,000$~\kms\ {\rm of\ }
204: $\lambda_0 = 6563$ \AA\ but, due to the expected asymmetry in the distribution
205: of the solar-abundance material, could present an H$\alpha$ profile ranging
206: from a very narrow spike to a broader emission line in the observed spectrum.
207: Obtaining spectra with high enough resolution to resolve fairly narrow lines is
208: thus a useful component of a targeted search for this H$\alpha$.
209:
210: To date, only a few nebular SN~Ia spectra have been obtained and H$\alpha$ has
211: never been detected, although the majority of the spectra lack the spectral
212: resolution and signal-to-noise ratio to place interesting constraints on
213: the companion. The best limits, by far, come from the recent study by
214: \cite{Mattila05}, which constrains hydrogen-rich material in SN~2001el to be
215: $\lesssi 0.03\ M_\odot$ from a low-resolution ($\sim 700$~\kms) spectrum
216: obtained 398 days past maximum light.
217:
218: In an effort to expand and improve on earlier work, both in terms of the
219: number of objects studied as well as the resolution, sensitivity, and temporal
220: coverage of the spectra, we have initiated a program to obtain deep, moderate
221: resolution ($\lesssim 150$~\kms, or $\sim 3$~\AA, at H$\alpha$), late-time
222: spectra of SNe~Ia at multiple epochs using the Keck and Gemini telescopes. The
223: first phase of this project has garnered data on two objects: SN~2005am, a
224: slightly subluminous \citep{Li06} event and SN~2005cf, an SN~Ia of normal
225: brightness \citep{Pastorello07}. We present and analyze our observations in
226: \S~\ref{sec:2}, discuss the results in \S~\ref{sec:3}, and conclude in
227: \S~\ref{sec:4}.
228:
229: \section{Observations and Analysis}
230: \label{sec:2}
231:
232: We obtained a total of five deep nebular-phase spectra: Two for SN~2005am and
233: three for SN~2005cf. Details of the observations are given in
234: Table~\ref{tab:1}. Following initial processing of the frames,\footnote{For
235: Gemini observations, the frames were processed using the tasks {\it gprepare},
236: {\it gsreduce}, {\it gsflat}, and {\it gmosaic} in the Gemini IRAF package.} we
237: extracted all one-dimensional sky-subtracted spectra optimally \citep{Horne86}
238: in the usual manner using the {\it apall} task within IRAF.\footnote{IRAF is
239: distributed by the National Optical Astronomy Observatories, which are operated
240: by the Association of Universities for Research in Astronomy, Inc., under
241: cooperative agreement with the National Science Foundation.} Each spectrum was
242: then wavelength and flux calibrated, and corrected for continuum atmospheric
243: extinction and telluric absorption bands \citep[][and references
244: therein]{Matheson00}. Careful examination of the two-dimensional spectra
245: reveals only faint night-sky emission at the nominal location of H$\alpha$, and
246: so we deem the potential for contamination at these wavelengths due to improper
247: background (``sky'') removal to be negligible.
248:
249: The final spectra are displayed in Figure~\ref{fig:1}. To search for H$\alpha$
250: emission, or to place limits on the amount of solar-abundance material present
251: in the inner ejecta, we subjected each spectrum to the following analysis
252: procedure.
253:
254: First, we attempted to place each spectrum on an absolute flux scale.
255: Adjustments to the original flux levels are necessary since most
256: observations were made with a rather narrow slit width compared with the
257: seeing (see Table~\ref{tab:1}), which makes the flux level susceptible to
258: the effects of seeing variations between the SN observations and those of
259: the flux standard star; some observations were also not obtained under
260: photometric conditions. To produce an approximate absolute flux
261: calibration, then, we computed synthetic photometry on our spectra and
262: compared it with estimates of the $V-$ or $R-$band (depending on the
263: spectral range of the spectrum) magnitudes of the SNe at the spectral
264: epochs. Since none of our spectra cover an entire passband, we extended
265: each spectrum's range by joining it with SN~Ia spectra of similar age from
266: the SUSPECT\footnote{see
267: http://bruford.nhn.ou.edu/$\sim$suspect/index1.html} database,\footnote{The
268: specific spectra used for this purpose include SN~1998bu (day 329;
269: \citealt{Cappellaro01}), SN~2003cg (day 385; \citealt{Elias-Rosa06}), and
270: SN~2002bo (day 375; \citealt{Benetti04}).} scaled to match the flux of our
271: spectra in regions of overlap. We then compared our synthetic photometry
272: with estimates of the actual brightnesses of the SNe. Since no published
273: late-time (i.e., beyond $t > 100$ days) photometry exists for either
274: SN~2005am or SN~2005cf, but high-quality early-time photometry does
275: \citep{Li06, Pastorello07}, we derived approximate $V$ and $R$ magnitudes of
276: the SNe at our spectral epochs by first extrapolating the early-time light
277: curves to day 200 through comparison with the light curves of SN~2003du
278: \citep{Stanishev07}, and then applying the late-time Type Ia decay-rates
279: reported by \citeauthor{Lair06} (\citeyear{Lair06}; $\Delta V = 1.46 {\rm\
280: mag\ 100\ day}^{-1},\ \Delta R = 1.54 {\rm\ mag\ 100\ day}^{-1}$). We then
281: adjusted each of our spectra by multiplying it by the scale factor needed to
282: bring its synthetic magnitude to the estimated actual apparent magnitude.
283: The only spectrum for which we did not apply this procedure was the day 298
284: epoch of SN~2005am, since the night was photometric, the flux of the SN from
285: one exposure to the next was very consistent (indicating non-varying seeing
286: throughout the observation sequence), and the spectrum of a nearby star
287: (HD~79289) taken immediately after the SN observations was found to have a
288: synthetic $V$ magnitude within $16\%$ of the value reported by
289: SIMBAD;\footnote{see http://simbad.u-strasbg.fr/simbad/} the scale factor
290: for this epoch was thus taken to be the value needed to place HD~79289 on a
291: correct absolute scale. The scale factors applied to all spectra are given
292: in Table~\ref{tab:2}. Given the uncertainties inherent in our technique,
293: the absolute flux calibrations are probably accurate to $\sim 25\%$, in
294: general.
295:
296: We next removed the redshift and rebinned each rest-frame spectrum to 3 \AA
297: ${\rm\ bin}^{-1}$, the approximate spectral resolution (see Table~\ref{tab:1}).
298: For SN~2005am in NGC~2811, we adopted the NASA/IPAC Extragalactic Database
299: (NED) recession velocity of $2,368$~\kms, and for SN~2005cf in MCG-01-39-3 we
300: adopted the NED recession velocity of $1,937$~\kms.
301:
302:
303: We then searched for narrow H$\alpha$ emission within $\pm 1000$~\kms\ ($\pm
304: 22$ \AA) of $\lambda_0 = 6563$~\AA\ in each rescaled, rest-frame spectrum by
305: smoothing the spectrum with a second-order Savitsky-Golay smoothing polynomial
306: \citep{Press92} of width $\sim 100$ \AA, differencing the smoothed and
307: unsmoothed spectra, and then examining the residuals for narrow emission near
308: H$\alpha$. This procedure yielded no evidence for narrow H$\alpha$ emission in
309: any spectrum. Indeed, no unexpected narrow features at any wavelength
310: (including spectral regions near [\ion{O}{1}] $\lambda\lambda 6300, 6364$ and
311: [\ion{Ca}{2}] $\lambda\lambda 7291, 7324$, which might be expected to produce
312: features similar to H$\alpha$ at these epochs from stripped solar-abundance
313: material) were found.
314:
315: To determine the greatest strength (i.e., equivalent width) of a feature that
316: could have remained undetected at the location of H$\alpha$ in each spectrum,
317: we adopted the procedure of \citet{Leonard4}, who report the $3 \sigma$ lower
318: bound on the equivalent width of an undetected feature in a spectrum to be:
319:
320: \begin{equation}
321: W_\lambda(3\sigma) = 3 \Delta\lambda\ \Delta I \sqrt{W_{line} / \Delta\lambda\
322: } \sqrt{1 / B},
323: \label{eqn:1}
324: \end{equation}
325:
326: \noindent where $\Delta \lambda$ is the width of a resolution element (in \AA),
327: $\Delta I$ is the 1$\sigma$ root-mean-square fluctuation of the flux around a
328: normalized continuum level, $W_{line}$ is the full-width at half-maximum of the
329: expected line feature (in \AA) and $B$ is the number of bins per resolution
330: element in the spectrum. The values for $\Delta \lambda$, $\Delta I$, and $B$,
331: along with the computed values of $W_\lambda(3\sigma)$, are reported in
332: Table~\ref{tab:2}. The width of the line feature, $W_{line}$, was assumed to be
333: $22$ \AA.
334:
335: While many studies have predicted the existence of narrow H$\alpha$ in nebular
336: spectra of SNe~Ia, only \citet{Mattila05} gives a quantitative estimate of its
337: expected strength. Thus, to translate our detection thresholds into estimates
338: of the maximum amount of solar-abundance material that could have given rise to
339: an H$\alpha$ line too weak to have been detected, we rely on this study alone.
340:
341: In the \citet{Mattila05} study, the code described by \citet{Kozma05} is
342: modified to artificially include varying amounts of solar abundance material in
343: the inner ($\pm 1,000$~\kms) region of the ejecta. The resulting spectra were
344: computed by \citet{Mattila05} in a time-dependent, one-dimensional, spherically
345: symmetric\footnote{Note that while the expected asymmetry in the location of
346: the stripped material would undoubtedly alter the shape of the resulting line
347: profile, \citet{Mattila05} conclude that it should not affect the total
348: strength of the line, and so the assumption of spherical symmetry is deemed to
349: be acceptable for the purposes of their study (and, hence, ours).} manner for
350: an SN~Ia 380 days after maximum light. From the parameters given by
351: \citeauthor{Mattila05} (see their Figure~6 and associated discussion), we
352: estimate that $0.05\ M_\odot$ of solar-abundance material produces a
353: luminosity in the peak of the H$\alpha$ line of $\sim 3.36 \times 10^{35} {\rm\
354: erg\ s}^{-1} {\rm\ \AA}^{-1} $.
355:
356: Translating predicted H$\alpha$ luminosity into observed H$\alpha$ strength
357: requires estimates of the distance and extinction of the SNe. For SN~2005am,
358: we adopt the NED Hubble flow ($H_0 = 73\ \kmsmpc$) distance of $36.8 \pm 2.6$
359: Mpc based on the 3k CMB rest frame as determined by \citet{Fixsen96}, and the
360: total extinction of $A_V = 0.27 \pm 0.08$ mag determined by \citet{Li06}. For
361: SN~2005cf, we use the distance and extinction values adopted by
362: \citet{Pastorello07} of $D = 31.77 \pm 4.8 {\rm\ Mpc}$ and $A_V = 0.32 \pm
363: 0.03$ mag. For $0.05\ M_\odot$ of solar-abundance material 380 days after
364: explosion, we derive expected H$\alpha$ line fluxes at the profile's peak of
365: $1.72 \times 10^{-18} $ and $2.23 \times 10^{-18} {\rm\ erg\ s}^{-1} {\rm\
366: \AA}^{-1} {\rm\ cm}^{-2}$ for SN~2005am and SN~2005cf, respectively.
367:
368: To derive the equivalent width that such features would have had in our
369: observed spectra, we approximated the H$\alpha$ emission as a Gaussian with a
370: $1\sigma$ width of $11$~\AA, and placed the resulting profile on top of the
371: estimated ``continuum'' level at $\lambda = 6563$ \AA\ (see Fig.~\ref{fig:1}).
372: We then measured the equivalent width of these theoretical profiles within the
373: region $\pm 11$~\AA\ from $\lambda_0 = 6563$ \AA\ [the same spectral range used
374: to derive $W_\lambda(3\sigma)$ earlier] in each spectrum. The calculated
375: values for $W_\lambda (0.05 M_\odot)$ are given in Table~\ref{tab:2}. Since
376: the H$\alpha$ emission is clearly a time-dependent phenomenon and the
377: \citet{Mattila05} model is specifically for day~380, our estimate of the
378: strength of the line in the day 298 and day 267 spectra of SN~2005am and
379: SN~2005cf, respectively, are quite crude. Fortunately, since the H$\alpha$
380: emission is powered primarily by gamma-ray deposition \citep{Mattila05}, and
381: the optical depth to gamma-rays in the central, hydrogen rich region should be
382: higher at earlier times, we would expect the H$\alpha$ line to be even stronger
383: at earlier times than it is on day 380 (assuming, of course, that the outer,
384: iron-rich ejecta are transparent at earlier times, and allow us to see it; see
385: \S~\ref{sec:3}). Thus, our calculated value of $W_\lambda (0.05\ M_\odot)$ for
386: epochs earlier than 380 days should serve as a conservative lower limit for
387: these spectra.
388:
389: Finally, we assume a linear scaling between the amount of stripped material and
390: the expected equivalent width of the H$\alpha$ emission line (an approximation
391: justified through examination of Figure~6 of \citealt{Mattila05}, where
392: theoretical profiles are shown for $0.01$ and $0.05\ M_\odot$) to
393: arrive at our final estimate for the upper limit on the amount of solar
394: abundance material that could have remained undetected at each spectral epoch.
395: We calculated this limit according to:
396:
397: \begin{equation}
398: {\rm M} (M_\odot) \leq \frac{W_\lambda(3\sigma)}{W_\lambda (0.05 M_\odot)}
399: \times 0.05 .
400: \label{eqn:2}
401: \end{equation}
402:
403: Final results are listed in Table~\ref{tab:2}, and can be summarized
404: succinctly: We restrict the amount of solar-abundance material that could have
405: evaded detected to be $< 0.01\ M_\odot$ at all spectral epochs except day 384
406: of SN~2005cf, for which the formal upper limit is $0.02\ M_\odot$.
407:
408:
409: \section{Discussion}
410: \label{sec:3}
411:
412: When coupled with the modeling results of \citet{Mattila05} and the
413: mass-stripping estimates of \citet{Marietta00} and \citet{Meng07}, our
414: restrictions on H$\alpha$ emission place strong constraints on the progenitor
415: systems that could have given rise to SN~2005am and SN~2005cf. In short, they
416: rule out all hydrogen-donating companions close enough to the WD to
417: have been experiencing RLOF at the time of explosion (i.e., $a/R
418: < 3$; see \S~\ref{sec:1}). Under the favored single-degenerate scenario, this
419: leaves only widely separated systems in which the companion (i.e., a red giant)
420: donates matter through a strong stellar wind, as a viable option. For the
421: symbiotic case, if we assume that the same scaling between the amount of
422: stripped material and the orbital separation holds for red giants as was found
423: by \cite{Marietta00} for main-sequence secondaries (see \S~\ref{sec:1}; note
424: that red giant companions were not analyzed by \citealt{Meng07}), then we
425: conclude that the mass-donating red giant must be $\gtrsim 10$ A.U. from the
426: WD to be consistent with our H$\alpha$ nondetections. For a $1.4\
427: M_\odot$ primary and a $1.0\ M_\odot$ secondary, this corresponds to an orbital
428: period of $\gtrsim 20$ yr.
429:
430: Our main conclusion --- that the most likely progenitors of SN~2005am and
431: SN~2005cf were widely separated symbiotics --- stands in evident contrast to
432: the work of \citet{Panagia06} on the search for, and subsequent nondetection
433: of, radio emission from SNe~Ia. In their study, \citet{Panagia06} set upper
434: limits on mass-loss rates of $\sim 10^{-7}\ M_\odot\ {\rm yr}^{-1}$ for the
435: progenitor systems. Since this is of the same order of magnitude as the
436: observed mass-loss rates from symbiotics \citep[e.g.,][]{Jung04,Crowley06},
437: \citet{Panagia06} specifically rule out symbiotics as potential progenitors of
438: SNe~Ia. However, the inferences drawn by \citet{Panagia06} rely on the
439: assumption that SN~Ia radio light curves behave identically to those of
440: SNe~Ib/c. Such an assumption is necessary because the correlation between wind
441: density and radio luminosity for SNe~Ia cannot be calculated from theory at
442: this point. In the absence of theoretical or observational support for such an
443: assumption the very low mass-loss rates claimed by \citet{Panagia06} cannot yet
444: be considered definitive \citep{Hughes07}. The radio non-detections may
445: therefore still permit the symbiotic scenario. Along these same lines, the
446: investigation of X-ray non-detections of SNe~Ia by \citet{Hughes07} provides
447: mass-loss limits of only $\sim 10^{-5}\ M_\odot\ {\rm yr}^{-1}$, which are not
448: stringent enough to rule out symbiotics. While a low-density circumstellar
449: environment is clearly favored by both the radio and X-ray non-detections of
450: SNe~Ia, it seems premature to rule out progenitor classes on this basis at the
451: present time.
452:
453: There are additional caveats to our conclusion that must be given. Of primary
454: significance is the fact that its robustness depends critically on the
455: validity of the theoretical modeling of just a few studies. Taken together,
456: these studies predict that a substantial amount of stripped hydrogen should
457: exist at low velocities \citep{Marietta00,Meng07} and emit a strong H$\alpha$
458: feature at late times \citep{Mattila05}. There are, however, alternative
459: explanations for the observed lack of hydrogen that must be thoroughly
460: investigated to build confidence in the inferences that can be drawn from our
461: study. We now critically evaluate four such alternatives, and point out areas
462: ripe for additional theoretical study.
463:
464: (1) {\it The hydrogen is ``hidden'' behind an opaque screen of faster-moving
465: ejecta.} Posed simply: Are the fast-moving, iron-rich ejecta truly optically
466: thin at late times? Since permitted \ion{Fe}{2} lines dominate the
467: underlying spectrum of SNe~Ia near $6563$ \AA, significant opacity in these
468: lines could provide an ``iron curtain'' that shields the hydrogen emission
469: from our view. Indeed, \citet{Kozma05} state that several Fe II lines do
470: remain optically thick in their models even at late times. On the other
471: hand, the simulations of \citet{Mattila05}, which predict the strong
472: H$\alpha$ feature, are explicitly {\it based} on the models of
473: \citet{Kozma05}, and so it seems reasonable to infer that the specific
474: \ion{Fe}{2} lines near H$\alpha$ do not provide significant opacity in the
475: \citet{Kozma05} models. Independent confirmation of the transparency of the
476: outer ejecta of SNe~Ia at late times is needed.
477:
478: (2) {\it There is far less hydrogen entrained in the ejecta than current models
479: predict.} By predicting up to a factor of four reduction in the amount of
480: stripped mass compared with the models of \citet{Marietta00},
481: \citet{Meng07} demonstrate the potentially large effect that binary
482: evolution can have on the amount of stripped material. Might other aspects
483: of binary evolution be at work to further reduce the stripped mass? A
484: possibly important mechanism not considered by \citet{Meng07} is the
485: outcome that a strong ``accretion wind'', blown from the WD during
486: the mass-accretion phase, could have on the secondary's envelope.
487: Accretion winds were originally conceived by \citet{Hachisu96} as a means
488: to stabilize the mass-transfer process and allow the system to avoid a
489: common envelope phase. Further investigation by \citet{Hachisu99} and
490: \citet{Hachisu03} revealed that the winds should also heat and ablate a
491: significant amount of material from a close secondary's envelope. From
492: these considerations, \citet{Hachisu99} estimate that the wind could strip
493: mass from such a companion at rates as high as $\sim 10^{-6}\ M_\odot\ {\rm
494: yr}^{-1}$ for a period of up to $\sim 10^{6}\ {\rm\ yr}$ prior to the
495: explosion. If such a process actually precedes an SN~Ia explosion, it
496: would clearly leave the secondary's envelope in a vastly different state
497: from the models considered to date. In fact, such secondaries could
498: potentially lose their entire hydrogen envelope and evolve to become
499: helium-donators, which would naturally account for a lack of hydrogen in
500: late-time spectra \citep{Branch95}.
501:
502: There are both theoretical and observational objections to strong WD winds
503: playing a large roll in stripping mass from the secondary star in SN~Ia
504: progenitor systems, however. On the theoretical front, \citet{Livio01}
505: questions the high efficiency of the mass stripping proposed by
506: \citet{Hachisu99}, and also points out that at high accretion rates, the
507: bulk of the WD's wind may be strongly collimated in a direction
508: perpendicular to the accretion disk rather than in the direction of the
509: companion star. Such a scenario would result in little stripping.
510: Observational doubt on accretion winds playing a major role in pre-SN~Ia
511: evolution is cast by \citet{Badenes07}, who find that such an optically
512: thick outflow from a WD's surface should excavate a large, low-density
513: cavity around the system. Such a large evacuated region is incompatible
514: with the known observational properties of SN~Ia remnants in the Galaxy,
515: the Large Magellanic Cloud, and M31. It thus appears that, while such
516: accretion winds may occur in nature \citep[e.g.,][]{Hachisu03}, they do not
517: evidently precede a majority of SN~Ia explosions. Further consideration of
518: this and other potential mass-stripping mechanisms is certainly warranted.
519:
520: (3) {\it The hydrogen gas is insufficiently powered to produce H$\alpha$
521: emission.} \citet{Mattila05} conclude that the hydrogen-rich, central
522: region of the young supernova remnant presents high optical depth to
523: gamma-ray photons that originate from the radioactive decay of isotopes
524: synthesized in the explosion. The gamma-ray trapping is then responsible
525: for powering the H$\alpha$ line emission. This mechanism is quite
526: different from the situation in the outer, lower density regions of the
527: ejecta, whose optical emission is predominantly powered by local positron
528: deposition \citep[e.g.,][]{Stritzinger07}. Since the production of a
529: strong H$\alpha$ line from even small amounts of solar-abundance material
530: relies fundamentally on the inner ejecta's ability to trap gamma rays, the
531: high opacity of the material, and the existence of sufficient numbers of
532: gamma-rays to power the emission, should be confirmed by independent
533: modeling.
534:
535: (4) {\it The single-degenerate scenario is not responsible for these SNe~Ia.}
536: A double-degenerate progenitor system, in which two WDs in a
537: binary system coalesce due to the emission of gravitational radiation and
538: ultimately explode as an SN~Ia \citep{Webbink84,Iben84}, would naturally
539: account for the absence of hydrogen in the spectra. Indeed, detection of
540: {\it any} hydrogen would deal a death blow to the double-degenerate
541: scenario. Difficulties with the double-degenerate scenario are well known
542: (see \citealt{Livio01} for a thorough review), however, and include both
543: observational (e.g., of the $\sim 120$ double-degenerate systems known,
544: none have a total mass in excess of the Chandrasekhar limit;
545: \citealt{Napiwotzki04}) and theoretical (e.g., the merger process seems
546: more likely to lead to collapse to a neutron star than to thermonuclear
547: explosion as an SN~Ia; \citealt{Saio98}, and references therein)
548: objections. Nonetheless, theoretical loopholes remain
549: \citep[e.g.,][]{Piersanti03,Yoon07}, and the observational sample of nearby
550: double-degenerate systems is not complete \citep{Nelemans05}. Since the
551: double-degenerate channel is a potential ``silver-bullet'' that can explain
552: both the lack of hydrogen in SNe~Ia as well as the occurrence of SNe~Ia in
553: both old and young star-forming systems \citep[e.g.,][]{Branch95}, it
554: requires continued theoretical and observational attention.
555:
556: \section{Conclusions}
557: \label{sec:4}
558:
559: We obtained five deep, moderate-resolution, nebular-phase spectra of two SNe~Ia
560: (SN~2005am and SN~2005cf) in order to search for narrow H$\alpha$ emission that
561: would betray the existence of material stripped from the envelope of a
562: mass-donating stellar companion to the exploding WD. No such emission is
563: detected in either object at any epoch. From the models of \citet{Mattila05},
564: we establish upper limits of $0.01\ M_\odot$ of solar abundance material in the
565: inner ejecta of both objects, which are the tightest constraints yet
566: established by such studies. Our non-detections of H$\alpha$, coupled with the
567: mass-stripping results of \citet{Marietta00} and \citet{Meng07}, rule out all
568: hydrogen-donating companions close enough to the WD to have been experiencing
569: RLOF at the time of explosion for these events. Additional theoretical work is
570: needed in several areas to buttress this conclusion, including most critically
571: verification of the transparency of the outer, more rapidly moving ejecta that
572: could potentially block H$\alpha$ photons from escaping from the inner region.
573: Bearing this caveat in mind, we propose that symbiotics are, at this time, the
574: most likely progenitor class that remains consistent with these data.
575:
576: Definitive proof of the identity of the progenitor system(s) that gives rise to
577: SNe~Ia remains elusive, and it must be admitted that our conclusion, which is
578: based on the {\it lack} of a detection, is not as satisfying as one based {\it
579: on} a detection. Should future modeling efforts prove unable to ``hide the
580: hydrogen'' for even widely separated binaries, then the continued viability of
581: the single-degenerate, hydrogen-donating progenitor will require that
582: H$\alpha$, no matter how weak, must ultimately be detected.
583:
584:
585: \acknowledgments
586:
587: I owe a debt of gratitude to Tom Matheson for valuable assistance with the
588: acquisition and reduction of the data obtained with the Gemini telescopes. I
589: acknowledge support from an NSF Astronomy and Astrophysics Postdoctoral
590: Fellowship (award AST-040147), under which part of this research was carried
591: out. This research has made use of the NASA/IPAC Extragalactic Database (NED),
592: which is operated by the Jet Propulsion Laboratory, California Institute of
593: Technology, under contract with NASA. Finally, I wish to recognize and
594: acknowledge the very significant cultural role and reverence that the summit of
595: Mauna Kea has always had within the indigenous Hawaiian community. We are most
596: fortunate to have the opportunity to conduct observations from this mountain.
597:
598:
599: %\bibliography{/Users/leonard/misc/all_refs}
600:
601: \begin{thebibliography}{50}
602: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
603:
604: \bibitem[{{Badenes} {et~al.}(2007)}]{Badenes07}
605: {Badenes}, C., {Hughes}, J.~P., {Bravo}, E., \& {Langer}, N. 2007, \apj,
606: 662, 472
607:
608: \bibitem[{{Baldwin} \& {Stone}(1984)}]{Baldwin84}
609: {Baldwin}, J.~A., \& {Stone}, R.~P.~S. 1984, \mnras, 206, 241
610:
611: \bibitem[{{Benetti} {et~al.}(2004)}]{Benetti04}
612: {Benetti}, S., {et~al.} 2004, \mnras, 348, 261
613:
614: \bibitem[{{Branch} {et~al.}(1995)}]{Branch95}
615: {Branch}, D., {Livio}, M., {Yungelson}, L.~R., {Boffi}, F.~R., \& {Baron},
616: E. 1995, \pasp, 107, 1019
617:
618: \bibitem[{{Brown} {et~al.}(2005)}]{Brown05}
619: {Brown}, P.~J., {et~al.} 2005, \apj, 635, 1192
620:
621: \bibitem[{{Cappellaro} {et~al.}(2001)}]{Cappellaro01}
622: {Cappellaro}, E., {et~al.} 2001, \apjl, 549, L215
623:
624: \bibitem[{{Chugai}(1986)}]{Chugai86}
625: {Chugai}, N.~N. 1986, Soviet Astronomy, 30, 563
626:
627: \bibitem[{{Crowley}(2006)}]{Crowley06}
628: {Crowley}, C. 2006, Ph.D. thesis, Trinity College Dublin
629:
630: \bibitem[{{Eggleton}(1973)}]{Eggleton73}
631: {Eggleton}, P.~P. 1973, \mnras, 163, 279
632:
633: \bibitem[{{Eggleton}(1983)}]{Eggleton83}
634: ---. 1983, \apj, 268, 368
635:
636: \bibitem[{{Elias-Rosa} {et~al.}(2006)}]{Elias-Rosa06}
637: {Elias-Rosa}, N., {et~al.} 2006, \mnras, 369, 1880
638:
639: \bibitem[{{Filippenko}(1982)}]{Filippenko82}
640: {Filippenko}, A.~V. 1982, \pasp, 94, 715
641:
642: \bibitem[{{Fixsen} {et~al.}(1996)}]{Fixsen96}
643: {Fixsen}, D.~J., {Cheng}, E.~S., {Gales}, J.~M., {Mather}, J.~C., {Shafer},
644: R.~A., \& {Wright}, E.~L. 1996, \apj, 473, 576
645:
646: \bibitem[{{Fryxell} \& {Arnett}(1981)}]{Fryxell81}
647: {Fryxell}, B.~A., \& {Arnett}, W.~D. 1981, \apj, 243, 994
648:
649: \bibitem[{{Hachisu} \& {Kato}(2003)}]{Hachisu03}
650: {Hachisu}, I., \& {Kato}, M. 2003, \apj, 590, 445
651:
652: \bibitem[{{Hachisu} {et~al.}(1996)}]{Hachisu96}
653: {Hachisu}, I., {Kato}, M., \& {Nomoto}, K. 1996, \apjl, 470, L97+
654:
655: \bibitem[{{Hachisu} {et~al.}(1999)}]{Hachisu99}
656: ---. 1999, \apj, 522, 487
657:
658: \bibitem[{{Hook} {et~al.}(2004)}]{Hook04}
659: {Hook}, I.~M., {J{\o}rgensen}, I., {Allington-Smith}, J.~R., {Davies},
660: R.~L., {Metcalfe}, N., {Murowinski}, R.~G., \& {Crampton}, D. 2004,
661: \pasp, 116, 425
662:
663: \bibitem[{{Horne}(1986)}]{Horne86}
664: {Horne}, K. 1986, \pasp, 98, 609
665:
666: \bibitem[{{Hughes} {et~al.}(2007)}]{Hughes07}
667: {Hughes}, J.~P., {Chugai}, N., {Chevalier}, R., {Lundqvist}, P., \&
668: {Schlegel}, E. 2007, \apj, in press
669:
670: \bibitem[{{Iben} \& {Tutukov}(1984)}]{Iben84}
671: {Iben}, Jr., \& \& {Tutukov} 1984, \apjs, 54, 335
672:
673: \bibitem[{{Jung} \& {Lee}(2004)}]{Jung04}
674: {Jung}, Y.-C., \& {Lee}, H.-W. 2004, \mnras, 355, 221
675:
676: \bibitem[{{Kozma} {et~al.}(2005)}]{Kozma05}
677: {Kozma}, C., {Fransson}, C., {Hillebrandt}, W., {Travaglio}, C.,
678: {Sollerman}, J., {Reinecke}, M., {R{\"o}pke}, F.~K., \& {Spyromilio}, J.
679: 2005, \aap, 437, 983
680:
681: \bibitem[{{Lair} {et~al.}(2006)}]{Lair06}
682: {Lair}, J.~C., {Leising}, M.~D., {Milne}, P.~A., \& {Williams}, G.~G.
683: 2006, \aj, 132, 2024
684:
685: \bibitem[{{Leonard} \& {Filippenko}(2001)}]{Leonard4}
686: {Leonard}, D.~C., \& {Filippenko}, A.~V. 2001, \pasp, 113, 920
687:
688: \bibitem[{{Li} {et~al.}(2006)}]{Li06}
689: {Li}, W., {Jha}, S., {Filippenko}, A.~V., {Bloom}, J.~S., {Pooley}, D.,
690: {Foley}, R.~J., \& {Perley}, D.~A. 2006, \pasp, 118, 37
691:
692: \bibitem[{{Livio}(2001)}]{Livio01}
693: {Livio}, M. 2001, in Supernovae and Gamma-Ray Bursts: the Greatest
694: Explosions since the Big Bang. Proceedings of the Space Telescope Science
695: Institute Symposium, May 3 - 6, 1999, ed. M. Livio, N. Panagia, \& K. Sahu
696: (Cambridge: Cambridge University Press), 334
697:
698: \bibitem[{{Livne} {et~al.}(1992)}]{Livne92}
699: {Livne}, E., {Tuchman}, Y., \& {Wheeler}, J.~C. 1992, \apj, 399, 665
700:
701: \bibitem[{{Marietta} {et~al.}(2000)}]{Marietta00}
702: {Marietta}, E., {Burrows}, A., \& {Fryxell}, B. 2000, \apjs, 128, 615
703:
704: \bibitem[{{Massey} \& {Gronwall}(1990)}]{Massey90}
705: {Massey}, P., \& {Gronwall}, C. 1990, \apj, 358, 344
706:
707: \bibitem[{{Matheson} {et~al.}(2000)}]{Matheson00}
708: {Matheson}, T., {Filippenko}, A.~V., {Ho}, L.~C., {Barth}, A.~J., \&
709: {Leonard}, D.~C. 2000, \aj, 120, 1499
710:
711: \bibitem[{{Mattila} {et~al.}(2005)}]{Mattila05}
712: {Mattila}, S., {Lundqvist}, P., {Sollerman}, J., {Kozma}, C., {Baron}, E.,
713: {Fransson}, C., {Leibundgut}, B., \& {Nomoto}, K. 2005, \aap, 443, 649
714:
715: \bibitem[{{Meng} {et~al.}(2007)}]{Meng07}
716: {Meng}, X., {Chen}, X., \& {Han}, Z. 2007, \pasj, 706, in press
717: (arXiv:0706.3581v1)
718:
719: \bibitem[{{Munari} \& {Renzini}(1992)}]{Munari92}
720: {Munari}, U., \& {Renzini}, A. 1992, \apjl, 397, L87
721:
722: \bibitem[{{Napiwotzki} {et~al.}(2004)}]{Napiwotzki04}
723: {Napiwotzki}, R., {et~al.} 2004, in ASP Conf. Ser. 318: Spectroscopically
724: and Spatially Resolving the Components of the Close Binary Stars, ed.
725: R.~W.~Hilditch, H.~Hensberge, \& K.~Pavlovski (San Francisco: ASP), 402
726:
727: \bibitem[{{Nelemans} {et~al.}(2005)}]{Nelemans05}
728: {Nelemans}, G., {et~al.} 2005, \aap, 440, 1087
729:
730: \bibitem[{{Oke} \& {Gunn}(1983)}]{Oke83}
731: {Oke}, J.~B., \& {Gunn}, J.~E. 1983, \apj, 266, 713
732:
733: \bibitem[{{Oke} {et~al.}(1995)}]{Oke95}
734: {Oke}, J.~B., {et~al.} 1995, \pasp, 107, 375
735:
736: \bibitem[{{Panagia} {et~al.}(2006)}]{Panagia06}
737: {Panagia}, N., {Van Dyk}, S.~D., {Weiler}, K.~W., {Sramek}, R.~A.,
738: {Stockdale}, C.~J., \& {Murata}, K.~P. 2006, \apj, 646, 369
739:
740: \bibitem[{{Pastorello} {et~al.}(2007)}]{Pastorello07}
741: {Pastorello}, A., {et~al.} 2007, \mnras, 376, 1301
742:
743: \bibitem[{{Piersanti} {et~al.}(2003)}]{Piersanti03}
744: {Piersanti}, L., {Gagliardi}, S., {Iben}, I.~J., \& {Tornamb{\'e}}, A.
745: 2003, \apj, 583, 885
746:
747: \bibitem[{{Press} {et~al.}(1992)}]{Press92}
748: {Press}, W.~H., {Teukolsky}, S.~A., {Vetterling}, W.~T., \& {Flannery},
749: B.~P. 1992, Numerical Recipes in C: The Art of Scientific Computing (2d
750: ed.; Cambridge: Cambridge Univ. Press)
751:
752: \bibitem[{{Saio} \& {Nomoto}(1998)}]{Saio98}
753: {Saio}, H., \& {Nomoto}, K. 1998, \apj, 500, 388
754:
755: \bibitem[{{Stanishev} {et~al.}(2007)}]{Stanishev07}
756: {Stanishev}, V., {et~al.} 2007, \aap, 469, 645
757:
758: \bibitem[{{Stritzinger} \& {Sollerman}(2007)}]{Stritzinger07}
759: {Stritzinger}, M., \& {Sollerman}, J. 2007, \aap, 470, L1
760:
761: \bibitem[{{Taam} \& {Fryxell}(1984)}]{Taam84}
762: {Taam}, R.~E., \& {Fryxell}, B.~A. 1984, \apj, 279, 166
763:
764: \bibitem[{{Webbink}(1984)}]{Webbink84}
765: {Webbink}, R.~F. 1984, \apj, 277, 355
766:
767: \bibitem[{{Wheeler} {et~al.}(1975)}]{Wheeler75}
768: {Wheeler}, J.~C., {Lecar}, M., \& {McKee}, C.~F. 1975, \apj, 200, 145
769:
770: \bibitem[{{Whelan} \& {Iben}(1973)}]{Whelan73}
771: {Whelan}, J., \& {Iben}, I.~J. 1973, \apj, 186, 1007
772:
773: \bibitem[{{Yoon} {et~al.}(2007)}]{Yoon07}
774: {Yoon}, S.-C., {Podsiadlowski}, P., \& {Rosswog}, S. 2007, \mnras, 380, 933
775:
776: \end{thebibliography}
777:
778: \clearpage
779:
780: \begin{figure}
781: \scalebox{0.9}{
782: \plotone{f1.eps}
783: }
784: \vskip -0.2in
785: \caption{Late-time spectra of two SNe~Ia, with day since $B$ maximum indicated.
786: The spectra are displayed at 3 \AA\ bin$^{-1}$, the approximate resolution at
787: H$\alpha$. The expected strengths of the H$\alpha$ line resulting from
788: $0.05\ M_\odot$ of solar-abundance material according to the day 380 models
789: of \cite{Mattila05} are shown as dot-dashed lines in the {\it insets}; note
790: that since the H$\alpha$ emission is a time-dependent phenomenon, the
791: estimated strength of these lines in the day 298 and day 267 spectra of
792: SN~2005am and SN~2005cf, respectively, are only approximate.
793: \label{fig:1} }
794: \end{figure}
795:
796: \clearpage
797:
798: \input{tab1.tex}
799: \input{tab2.tex}
800:
801: \end{document}
802:
803: