astro-ph0611356/ms.tex
1: %- 
2: %- File    : ms.tex
3: %- ----------------
4: %- Created : Thu 13 Jul 2006 01:33:27 PM PDT
5: %- Authors : Rollin C. Thomas (RCT) - rcthomas@lbl.gov
6: %- 
7: %- $Header: /cvs/snovae/SNFactory/Analysis/SN2006D/ms.tex,v 1.13 2006/10/12 20:30:10 rthomas Exp $
8: %- 
9: %- Comments --
10: %- [1] Read preamble for commands you can use before you make more.
11: %-
12: 
13: %- Preamble 
14: 
15:    \documentclass[12pt,preprint]{aastex}
16: %  \documentclass[manuscript]{aastex}
17: %  \documentclass[preprint2]{aastex}
18: %  \documentclass[preprint2,longabstract]{aastex}
19: 
20: %- Handy RCS Label (comment out when submitting)
21: 
22: % \def\RCS$#1: #2 ${\expandafter\def\csname RCS#1\endcsname{#2}}
23: % \RCS$Revision: 1.13 $
24: % \RCS$Date: 2006/10/12 20:30:10 $
25: 
26: %- Abbreviations.
27: 
28: \newcommand{\snia}{SN~Ia}
29: \newcommand{\sneia}{SNe~Ia}
30: \newcommand{\wl}{$\lambda$}
31: \newcommand{\kms}{\ensuremath{\textrm{km~s}^{-1}}}
32: \newcommand{\snsd}{SN~2006D}
33: \newcommand{\co}{C-O} %{\ensuremath{C\!+\!O}}
34: \newcommand{\synow}{SYNOW}
35: \newcommand{\phoenix}{\texttt{PHOENIX}}
36: \newcommand{\msun}{$M_\odot$}
37: \newcommand{\bi}{\ensuremath{b_i}}
38: 
39: %- Substitutions.
40: 
41: \newcommand{\discdate}{2006~January~11.2 UTC}
42: \newcommand{\peakdate}{2006~January~21.8 UTC}
43: 
44: % \newcommand{\degree}{$^\circ$}
45: % \newcommand{\Halpha}{H\ensuremath{\alpha}}
46: % \newcommand{\Hbeta}{H\ensuremath{\beta}}
47: % \newcommand{\Hgamma}{H\ensuremath{\gamma}}
48: % \newcommand{\Hdelta}{H\ensuremath{\delta}}
49: % \newcommand{\Heps}{H\ensuremath{\epsilon}}
50: % \newcommand{\snii}{SN~II}
51: % \newcommand{\sniin}{SN~IIn}
52: % \newcommand{\sneiin}{SNe~IIn}
53: 
54: \newcommand{\rem}[1]{\textbf{[\textsl{#1}]}}
55: \newcommand{\needref}{\textbf{NEEDS REF}}
56: 
57: %- Labels appearing on the draft.
58: 
59: %\slugcomment{To appear in Astrophysical Journal Letters}
60: % \slugcomment{Revision \RCSRevision, \RCSDate}
61: 
62: \begin{document}
63: 
64: \title
65: {
66:    Nearby Supernova Factory Observations of \snsd:
67:    On Sporadic Carbon Signatures in Early Type Ia Supernova Spectra
68: }
69: 
70: %% Alphabetical - also equivalent to First author/alphabetical?
71: 
72: \author
73: {
74:    The Nearby Supernova Factory \\
75:    R.~C.~Thomas,\altaffilmark{1}
76:    G.~Aldering,\altaffilmark{1}
77:    P.~Antilogus,\altaffilmark{3}
78:    C.~Aragon,\altaffilmark{1}
79:    S.~Bailey,\altaffilmark{1}
80:    C.~Baltay,\altaffilmark{8}
81:    E.~Baron,\altaffilmark{9}
82:    A.~Bauer,\altaffilmark{8}
83:    C.~Buton,\altaffilmark{2}
84:    S.~Bongard,\altaffilmark{1,5}
85:    Y.~Copin,\altaffilmark{2}
86:    E.~Gangler,\altaffilmark{2}
87:    S.~Gilles,\altaffilmark{3}
88:    R.~Kessler,\altaffilmark{7}
89:    S.~Loken,\altaffilmark{1}
90:    P.~Nugent,\altaffilmark{1}
91:    R.~Pain,\altaffilmark{3}
92:    J.~Parrent,\altaffilmark{9}
93:    E.~P\'econtal,\altaffilmark{4}
94:    R.~Pereira,\altaffilmark{3} 
95:    S.~Perlmutter,\altaffilmark{1,6}
96:    D.~Rabinowitz,\altaffilmark{8}
97:    G.~Rigaudier,\altaffilmark{4}
98:    K.~Runge,\altaffilmark{1}
99:    R.~Scalzo,\altaffilmark{1}
100:    G.~Smadja,\altaffilmark{2}
101:    L.~Wang,\altaffilmark{1}
102:    B.~A.~Weaver\altaffilmark{1,5}
103: }
104: %  N.~Blanc,\altaffilmark{2}
105: %  M.~J.~Childress,\altaffilmark{1,6}
106: %  D.~Kocevski,\altaffilmark{1,6}
107: 
108: %- Please check your affiliation and its address.
109: 
110: \altaffiltext{1}{Physics Division, Lawrence Berkeley National
111: Laboratory, 1 Cyclotron Road, Berkeley, CA 94720}
112: \altaffiltext{2}{Institut de Physique Nucl\'eaire de Lyon, UMR5822,
113: CNRS-IN2P3; Universit\'e Claude Bernard Lyon~1, F-69622 Villeurbanne
114: France}
115: \altaffiltext{3}{Laboratoire de Physique Nucl\'eaire et des Hautes
116: Energies IN2P3 - CNRS - Universit\'es Paris VI et Paris VII, 4 place
117: Jussieu Tour 33 - Rez de chauss\'ee 75252 Paris Cedex 05}
118: \altaffiltext{4}{Centre de Recherche Astronomique de Lyon, 9, av.
119: Charles Andr\'e, 69561 Saint Genis Laval Cedex}
120: \altaffiltext{5}{University of California, Space Sciences Laboratory,
121: Berkeley, CA 94720-7450}
122: \altaffiltext{6}{Department of Physics, University of California,
123: Berkeley, CA 94720}
124: \altaffiltext{7}{Kavli Institute for Cosmological Physics, The
125: University of Chicago, Chicago, IL 60637}
126: \altaffiltext{8}{Department of Physics, Yale University, New Haven, CT
127: 06250-8121}
128: \altaffiltext{9}{Homer L. Dodge Department of Physics and Astronomy, 440 W.
129: Brooks Street, University of Oklahoma, Norman, OK 73019}
130: 
131: \begin{abstract}
132: 
133: We present four spectra of the Type Ia supernova (\snia) 2006D extending
134: from $-7$ to $+13$ days with respect to $B$-band maximum.  The spectra
135: include the strongest signature of unburned material at photospheric
136: velocities observed in a \snia\ to date.  The earliest spectrum exhibits
137: \ion{C}{2} absorption features below 14,000 \kms, including a
138: distinctive \ion{C}{2} \wl 6580 absorption feature.  The carbon
139: signatures dissipate as the SN approaches peak brightness.  In addition
140: to discussing implications of photospheric-velocity carbon for white
141: dwarf explosion models, we outline some factors that may influence the
142: frequency of its detection before and around peak brightness.  Two
143: effects are explored in this regard, including depopulation of the
144: \ion{C}{2} optical levels by non-LTE effects, and line-of-sight effects
145: resulting from a clumpy distribution of unburned material with low
146: volume-filling factor.
147: 
148: \end{abstract}
149: 
150: \keywords
151: {
152:    supernovae: general --- 
153:    supernovae: individual (\snsd)
154: }
155: 
156: \section{Introduction}
157: 
158: Type Ia supernovae (\sneia) make valuable standard candles because of
159: their intrinsic brightness and the homogeneity of their light curves.
160: Observations of high-redshift \sneia\ are responsible for the recent
161: revelation that the rate of expansion of the Universe is accelerating
162: \citep{perl98,garnavich98,riess98,perl99}.  A push to better calibrate
163: \sneia\ as distance indicators and control systematics in future
164: high-precision cosmology experiments has helped motivate the search for
165: deeper physical insight into \snia\ progenitors and their explosion
166: mechanism.
167: 
168: %F \begin{figure*}[ht]
169: %F    \centering
170: %F    \includegraphics[scale=1.00,clip=true]{time_series}
171: %F    \caption{SNIFS spectroscopy of \snsd.  Phases are expressed relative
172: %F    to a $B$-band peak brightness date of \peakdate.  Dotted vertical
173: %F    lines mark rest wavelengths of \ion{C}{2} lines \wl\wl 4267, 4745,
174: %F    6580, and 7234.  Dark bands indicate blueshifts between 10,000 and
175: %F    14,000 \kms\ with respect to these lines, typical of the velocity at
176: %F    the photosphere at these phases.  The inset is a zoom of the region
177: %F    around the 6320~\AA\ notch in the $-7$ day spectrum.}
178: %F    \label{fig:time_series}
179: %F \end{figure*}
180: 
181: Certain recent multidimensional hydrodynamical \snia\ explosion models
182: involve the thermonuclear disruption of a Chandrasekhar-mass \co\ white
183: dwarf (WD) by \emph{deflagration}, the propagation of a subsonic flame
184: front through the star \citep[e.g.,][]{gamezo2003,ropke2005}.  A generic
185: characteristic of these new models is the presence of unprocessed WD
186: material below the canonical 14,000 \kms\ cutoff seen in W7, a tuned
187: spherically-symmetric deflagration model \citep{nomoto1984}.  Thus,
188: carbon detection at such low velocities would generally provide
189: observational support for the newer multidimensional deflagration
190: models.  Few \snia\ spectra actually exhibit a robust low-velocity
191: carbon signature; to date the lowest measured are the \ion{C}{2} \wl
192: 6580 and \wl 7234 features in the spectra of SN~1998aq, which may extend
193: deeper than 11,000 \kms\ \citep{branch2003}.  The lack of ubiquitous
194: carbon signatures in the spectra of \sneia, and the observed deficit of
195: kinetic energy and synthesized nickel indicate that these models are not
196: viable, though increasingly sophisticated models (e.g., with greater
197: resolution and more detailed nucleosynthesis) may prove otherwise.  Most
198: recently, \citet{marion2006} used near-infrared spectroscopy to conclude
199: that nuclear burning in at least three normal \sneia\ is complete below
200: 18,000 \kms, a conclusion supporting models where an initial
201: deflagration transitions to a supersonic \emph{detonation}
202: \citep{khokhlov1991}.
203: 
204: We present Nearby Supernova Factory \citep[SNfactory,][]{aldering2002}
205: spectroscopy of the \snia\ 2006D, which includes an unambiguous
206: photospheric-velocity \ion{C}{2} signature.  The data were obtained
207: using the Supernova Integral Field Spectrograph
208: \citep[SNIFS,][]{aldering2002,lantz2004} on the University of Hawaii
209: 2.2-meter telescope on Mauna Kea.  Our focus here is on the early
210: photospheric-phase spectra of \snsd; an analysis involving all of our
211: data will appear elsewhere.
212: 
213: \section{Data \& Analysis}
214: 
215: The Brazilian Supernova Search \citep{colesanti2006} discovered \snsd\
216: in MCG -01-33-34 \citep[$z = 0.00853$,][]{davoust2004} on \discdate.
217: Using a SNIFS spectrum obtained January~14.6~UTC, we classified \snsd\
218: as a \snia\ one week prior to maximum brightness \citep{atel}.  This and
219: three subsequent spectra appear in Figure~\ref{fig:time_series} and are
220: summarized in Table~\ref{tab:spectra}.  The spectra were reduced using
221: our dedicated data-reduction procedure, similar to that presented in
222: \S~4 of \citet{bacon2001} and \S~2 of \citet{aldering2006}.  No
223: correction for interstellar reddening has been applied.  The
224: intermediate-mass element (IME) spectral features typical of premaximum
225: \sneia\ are labelled on the first spectrum for reference.
226: A preliminary light curve fit derived from SNIFS acquisition
227: images suggests a $B$-band peak date of \peakdate.  The light curve is
228: more narrow than is typical for a spectroscopically normal \snia\ ---
229: the derived time-scale ``stretch'' parameter $s \sim 0.75$ is
230: significantly smaller than in the normal case ($s \equiv 1$).  This
231: value is similar to that derived from the light curves of SNe~1986G and
232: 1992bo \citep[0.74 and 0.73 respectively,][]{guy2005}, but not as
233: extreme as in the sub-luminous SNe~1991bg or 1999by \citep[$s \sim
234: 0.6$;][]{ruiz2004, garnavich2004,jha2006}.
235: 
236: The dotted vertical lines in Figure~\ref{fig:time_series} indicate the
237: rest-frame wavelengths of the four \ion{C}{2} lines that would be the
238: strongest under the assumption of local thermodynamic equilibrium (LTE)
239: at 10,000~K.  The shaded regions to the blue of each line cover Doppler
240: shifts between 14,000 and 10,000 \kms, representing velocities typical
241: of the photosphere before and around peak brightness.  A striking
242: feature of the $-7$ day spectrum is the $\vee$-shaped absorption
243: centered at 6320~\AA\ (rest frame), which we attribute to \ion{C}{2} \wl
244: 6580 at a Doppler shift of 12,000 \kms.  In the first spectrum, two much
245: weaker notches fall within the bands of the shaded regions corresponding
246: to \ion{C}{2} \wl\wl 4745 and 7234.  These two features and the
247: 6580~\AA\ notch weaken by day $+13$.
248: 
249: Another interesting feature in our spectra is the notch centered at
250: 4100~\AA, just to the red of the small absorption typically attributed
251: to \ion{Si}{2} \wl\wl 4128,4131.  If the 4100~\AA\ absorption is due to
252: \ion{C}{2} \wl 4267, then it extends from 14,000 \kms\ down to 8,000
253: \kms, slower than the red edge of the \wl 6580 absorption by 2,000 \kms\
254: (of course, such a low velocity edge could be the result of
255: line-blending).  While the other \ion{C}{2} features dissipate by day
256: $+13$, the 4100~\AA\ notch persists and the absorption along its red
257: edge strengthens somewhat.  In later spectra not shown here, this
258: feature steadily becomes washed out by iron peak lines in this part of
259: the spectrum.  If this feature does arise from \ion{C}{2} \wl 4267, then
260: it does so nonthermally --- in LTE at 10,000 K it would be approximately
261: 30 times weaker than \ion{C}{2} \wl 6580; we return to this point
262: shortly.
263: 
264: %F \begin{figure}[h]
265: %F    \centering
266: %F    \includegraphics[width=1.00\textwidth,clip=true]{spec_comp_early}
267: %F    \caption{Comparison of the $-7$ day spectrum of \snsd\ to premaximum
268: %F    spectra of SNe~1998aq, 1986G, and 1999by.}
269: %F    \label{fig:spec_comp_early}
270: %F \end{figure}
271: 
272: In Figure~\ref{fig:spec_comp_early}, we plot our earliest spectrum
273: together with premaximum spectra of SNe~1998aq, 1986G, and 1999by.  The
274: spectrum of SN~1998aq is an excellent match to that of \snsd,
275: particularly redward of 4400~\AA, aside from the stronger \ion{C}{2} \wl
276: 6580 absorption in the latter.  Blueward of 4400~\AA, the agreement is
277: not as good; the 4100~\AA\ notch in the spectrum of \snsd\ has no
278: counterpart in that of SN~1998aq save for a minor wiggle \citep[in fact,
279: this wiggle has been previously observed in the spectra of \sneia, see
280: Figure 2 of][]{branch2006}.  The spectra of SNe~1986G and 1999by exhibit
281: larger \ion{Si}{2} 5800~\AA\ to \ion{Si}{2} 6150~\AA\ absorption ratios
282: than does \snsd, indicative of lower temperature than in more
283: spectroscopically normal \sneia\ \citep{nugent1995}.  The region from
284: 4000 to 4400~\AA\ in these two spectra is dominated by \ion{Ti}{2}, also
285: characteristic of lower temperature.  The differences between the
286: spectra of \snsd\ and SN~1998aq in this region, and the possibility
287: based on its light curve that \snsd\ is slightly sub-luminous, suggest
288: that \ion{Ti}{2} may play a role in the formation of its spectrum.
289: 
290: %F \begin{figure}[h]
291: %F    \centering
292: %F    \includegraphics[width=1.00\textwidth,clip=true]{synow_06_014_088_001}
293: %F    \caption{Synthetic spectra overlaid on the $-7$ day spectrum of
294: %F    \snsd.  Inset plots show the coincidence between the synthetic
295: %F    \ion{C}{2} spectrum and the data.}
296: %F    \label{fig:synow_m10}
297: %F \end{figure}
298: 
299: More detailed SN spectral line identification requires an accounting for
300: line blending, so we compare the $-7$ day spectrum of \snsd\ with spectra
301: generated by the highly-parameterized SN spectrum synthesis code \synow\
302: \citep[e.g.,][]{branch2003}.  \synow\ uses a simplified model of a SN
303: atmosphere consisting of a sharply-defined continuum-emitting core
304: surrounded by a line-forming region.  Line formation is treated under
305: the Sobolev approximation, with line opacity parameterized as a function
306: of ejection velocity and with relative strengths for a given ion set by
307: assuming Boltzmann-factor population of the levels.  A synthetic
308: spectrum is overlaid on the observed $-7$ day spectrum in
309: Figure~\ref{fig:synow_m10}.
310: 
311: The dashed fit reproduces most of the main absorption features of the
312: spectrum, including those of IME ions \ion{Ca}{2}, \ion{Si}{2}, and
313: \ion{S}{2}.  A velocity at the photosphere of 11,000~\kms\ produces
314: satisfactory results.  The blend at 4700~\AA\ consists of lines of
315: \ion{Fe}{2} and \ion{Si}{2}, while the blend from 4100~\AA\ to 4400~\AA\
316: consists of lines from \ion{Fe}{3}, \ion{Mg}{2}, and \ion{Si}{3}, and
317: indeed some \ion{Ti}{2} (which also fits the absorption at 3600~\AA\ as
318: well).  This \ion{Ti}{2} detection is consistent with cooler,
319: fast-declining
320: \sneia.
321: 
322: The sharpness of the \wl 6320 feature, and the lack of a
323: discernable accompanying emission feature to its red motivates us to
324: distribute \ion{C}{2} Sobolev line opacity in a layer ``detached'' from
325: the SYNOW photosphere \citep{jeffery1990}.  An absorption feature
326: without a corresponding P~Cygni emission may also be indicative of a
327: clump along the line of sight partially obscuring the photosphere
328: \citep{thomas2002}.  A thin shell of Sobolev line opacity immediately
329: above the photosphere would result in a flat absorption trough instead
330: of a sharper downward spike.  In the synthetic spectrum, \ion{C}{2}
331: opacity is detached from the photosphere to a velocity of 12,000 \kms,
332: producing a good match to the \ion{C}{2} \wl\wl 6350 and 7234 features.
333: The dotted synthetic spectra in the insets of Figure~\ref{fig:synow_m10}
334: are from a pure \ion{C}{2} spectrum, overlaid on the neighborhoods of
335: the observed \ion{C}{2} \wl\wl 4745, 6580, and 7234 features.
336: 
337: We can account for the 4100~\AA\ feature with a separate higher velocity
338: \ion{Ti}{2} component (included in the fit with $v > 17,000$ \kms) but
339: ultimately reject this hypothesis due to the lack of any sign of
340: high-velocity components to explain other spectral features.  In
341: particular, \ion{Ca}{2} opacity should trace the same high-velocity
342: processed material, and we detect no such signature.
343: 
344: In the absence of any obvious alternative to \ion{Ti}{2} for the
345: 4100~\AA\ notch, we suggest that the feature could come from \ion{C}{2}
346: \wl 4267, stimulated by some non-LTE process.  To investigate this
347: hypothesis, we examined the non-LTE departure coefficients ($b_i \equiv
348: n_i / n_i^{LTE}$, where $n_i$ is the population number of the atomic
349: level, $i$) of the lower levels of the optical \ion{C}{2} lines in a
350: \phoenix\ calculation of the model W7 \citep{baron2006}.  In the
351: \co-rich zone, departure coefficients in optical lines are found to be
352: significantly smaller than unity ($10^{-7}$ to $10^{-3}$).  Furthermore,
353: the ratio of the \wl 4267 lower-level \bi\ to that of \wl 6580 is
354: observed to range between 1 and 60.  This situation arises due
355: to more efficient recombination through the ultraviolet resonance line
356: \ion{C}{2} \wl 687, so that recombination into the optical levels is
357: suppressed and they take on level populations significantly smaller than
358: LTE.  The size of this ratio could account for the persistence of
359: \ion{C}{2} \wl 4267 while the other \ion{C}{2} lines dissipate.  Further
360: detailed models are required to discover whether this effect is indeed
361: generic and holds in other models besides W7.  This effect may also
362: resolve the long-standing difficulty of making a reliable identification
363: of \ion{C}{2} in the spectra of \sneia.
364: 
365: \section{Discussion}
366: 
367: That some \snia\ spectra possess \ion{C}{2} signatures, particularly
368: before maximum light, is now quite clear.  Past tentative
369: identifications of \ion{C}{2} \wl 6580 in such cases as SN~1990N
370: \citep{mazzali2001} and SN~1999ac \citep{garavini2005}, and previously
371: unidentified notches seen in SN~1994D and SN~1996X pointed out by
372: \citet{branch2003} now seem more meaningful.  These detections and their
373: circumstances prompt consideration of several issues relevant to the
374: study of \snia\ explosion models.
375: 
376: %F \begin{figure}[ht]
377: %F    \centering
378: %F    \includegraphics[width=0.4\textwidth,clip=true]{c_comp}
379: %F    \caption{Comparison of \ion{C}{2} \wl 6580 feature across several
380: %F    premaximum \snia\ spectra.  The spectra are resampled into 10~\AA\
381: %F    bins.}
382: %F    \label{fig:c_comp}
383: %F \end{figure}
384: 
385: The primary issue is the amount of carbon detected.  To place a lower
386: limit on the carbon mass (particularly in light of the small \bi's
387: observed in the \phoenix\ model), we compute LTE electron level
388: populations and corresponding \ion{C}{2} \wl 6580 Sobolev optical depth
389: on a temperature-density grid, assuming a \co-rich composition.  The
390: absolute minimum density giving rise to optical depth $\tau \sim 1$ at
391: 10 days after outburst is $5.9 \times 10^{-15}$~g~cm$^{-3}$.  Treating
392: this as an average density results in a lower limit of 0.007 \msun\ of
393: carbon (0.014 \msun\ of unburned material) over the velocity range
394: 10,000 to 14,000 \kms\ where the absorption feature is detected.  This
395: value is smaller than the 0.085 \msun\ of unburned material present at
396: the same velocity interval in the multidimensional model \emph{b150\_3d}
397: of \citet{ropke2006}.  
398: 
399: If the behavior of the \ion{C}{2} \bi's seen in the \co-rich zone of the
400: W7 \phoenix\ model is generic, then the amount of unburned material
401: detected could be \emph{significantly higher} than this estimate.  The
402: \bi's need not be so miniscule as observed in the \phoenix\ model ---
403: here only a factor of 6 is needed to produce general agreement with
404: model \emph{b150\_3d}.  Extremely small departure coefficients drive the
405: mass up but the mass is obviously limited by the mass-energy budget of
406: the progenitor, fixed by the Chandrasekhar mass.  One clear way to make
407: the detection consistent with this constraint is for the unburned
408: material to be distributed in clumps if the \bi's are indeed very small.
409: 
410: The presence of unburned material at the observed velocities may appear
411: more consistent with published multidimensional deflagration models than
412: with delayed-detonation models, but the situation is not quite so
413: simple.  For example, \citet{hoflich2002} demonstrate that sub-luminous
414: delayed-detonation models can fit the light curve and infrared spectra
415: of the sub-luminous SN~1999by, and account for observed \ion{C}{1}
416: features observed with a small amount of carbon extending below 14,000
417: \kms.  Interestingly, the best-fit model for SN~1999by is from a grid of
418: delayed-detonation models with variable transition density, and
419: corresponds to the lowest luminosity/lowest transition density in the
420: grid.  The trend in those models seems to indicate that less luminous
421: models, with lower transition densities, have a correspondingly
422: decreased carbon cutoff velocity.  Thus, if the more luminous \snsd\
423: were to fall on that particular grid of models, it would need to
424: correspond to a higher transition density and have a higher carbon
425: cutoff velocity than SN~1999by, at odds with our observations.  More
426: speculatively, \citet{gamezo2004} outline a scenario in which a small
427: amount of unburned material in pockets could be left by a
428: multidimensional delayed detonation --- as the detonation wave processes
429: material left between plumes generated during the deflagration phase,
430: abrupt twists and turns could cut it off from fresh WD fuel.
431: Elucidating the details (the mass and spatial distribution of such
432: pockets) will rely on progress in future multidimensional
433: delayed-detonation models.
434: 
435: It is interesting to consider the significance of the sporadic and
436: diverse nature of the detections of the \ion{C}{2} \wl 6580 feature
437: itself.  The premaximum rest-frame 5800-6600~\AA\ spectra of \snsd\ and
438: five other \sneia\ plotted in Figure~\ref{fig:c_comp} exhibit a variety
439: of strengths and velocity ranges in \ion{C}{2} \wl 6580.  The
440: theoretical multidimensional models themselves may suggest an
441: explanation for this phenomenon beyond any non-LTE effect --- the
442: distribution of unburned material at these velocities could be quite
443: asymmetric, confined to clumps of small filling factor compared to the
444: photosphere.  Thus, differences from SN to SN would naturally be
445: explained as a function of the distribution of the clumps with respect
446: to the observer's line of sight.  This argument is consistent with the
447: complementary finding by \citet{thomas2002}, where the relatively
448: consistent shape and depth of \ion{Si}{2} \wl 6355 absorption at maximum
449: light indicates that the distribution of IME's in \sneia\ cannot be
450: highly asymmetric.  Of course, differences in the details of temperature
451: structure and ionization from one SN to another may be in play, though
452: we would expect temperature effects to run along a continuum correlated
453: to, say, radioactive nickel mass.  If low-velocity carbon detections
454: are sporadic and less correlated with luminosity \citep[note that the
455: light curve of SN~1998aq is characterized as ``typical,''][]{branch2003},
456: one would expect the geometrical argument to be at least as plausible.
457: An underlying mechanism for either explanation (or both) will depend on
458: further observations of the earliest \snia\ spectra that probe the outer
459: layers of these events, and on improved detailed modeling of these
460: spectra.
461: 
462: \section{Conclusions}
463: 
464: We have presented the strongest evidence to date of unburned ejecta at
465: low velocity in the early optical spectra of a \snia.  The case of \snsd\
466: reiterates the importance of early-time spectroscopy (and
467: spectropolarimetry) for \snia\ research.  Complimentary observations in
468: the nebular phase (one year after explosion) may present further
469: opportunities to probe the distribution and amount of unburned material
470: at low velocity in \snsd;  while their predictions are not strictly
471: applicable at the velocities considered in this work, the
472: multidimensional spectrum synthesis models of \citet{kozma2005} suggest
473: that unburned material at velocities below 10,000 \kms\ roughly
474: consistent with our mass estimates may be detectable at late times as
475: forbidden [\ion{O}{1}] and [\ion{C}{1}] emission.  
476: 
477: \acknowledgments
478: 
479: The authors thank the anonymous referee for many helpful comments.
480: We are grateful to the technical and scientific staff of the University
481: of Hawaii 2.2-meter telescope for their assistance in obtaining these
482: data.  The authors wish to recognize and acknowledge the very
483: significant cultural role and reverence that the summit of Mauna Kea has
484: always had within the indigenous Hawaiian community.  We are most
485: fortunate to have the opportunity to conduct observations from this
486: mountain.  We also thank F.~R\"opke for kindly providing the
487: angle-averaged mass distribution of model \emph{b150\_3d}.  Our analysis
488: included data from the University of Oklahoma Online SUpernova SPECTrum
489: (SUSPECT) Archive.  This work was supported in part by the Director,
490: Office of Science, Office of High Energy and Nuclear Physics, of the
491: U.S. Department of Energy under Contract No. DE-FG02-92ER40704, by a
492: grant from the Gordon \& Betty Moore Foundation, by  National Science
493: Foundation Grant Number AST-0407297, and in France by support from
494: CNRS/IN2P3, CNRS/INSU and PNC.  This research used resources of the
495: National Energy Research Scientific Computing Center, which is supported
496: by the Office of Science of the U.S.  Department of Energy under
497: Contract No. DE-AC02-05CH11231.  We also acknowledge support from the
498: U.S. Department of Energy Scientific Discovery through Advanced
499: Computing program under Contract No.  DE-FG02-06ER06-04.
500: 
501: {\it Facilities:} 
502: \facility{UH:2.2m (SNIFS)}
503: 
504: 
505: \clearpage
506: 
507: \begin{thebibliography}{}
508: 
509: \bibitem[Aldering et al.(2002)]{aldering2002} 
510: Aldering, G., et al.\ 2002, \procspie, 4836, 61
511: 
512: \bibitem[Aldering et al.(2006)]{aldering2006}
513: Aldering, G., et al.\ 2006, \apj, 650, 510
514: 
515: \bibitem[Bacon et al.(2001)]{bacon2001} 
516: Bacon, R., et al.\ 2001, \mnras, 326, 23 
517: 
518: \bibitem[Baron et al.(2006)]{baron2006}
519: Baron, E., Bongard, S., Branch., D.\ \& Hauschildt, P.\ 2006, \apj, 645,
520: 480
521: 
522: \bibitem[Branch et al.(2003)]{branch2003} 
523: Branch, D., et al.\ 2003, \aj, 126, 1489
524: % Garnavich, P., Matheson, T., Baron, E., Thomas, R.~C.,
525: % Hatano, K., Challis, P., Jha, S., \& Kirshner, R.~P.\ 
526: 
527: \bibitem[Branch et al.(2006)]{branch2006}
528: Branch, D., et al.\ 2006, \pasp, 118, 560
529: 
530: \bibitem[Colesanti et al.(2006)]{colesanti2006} 
531: % Colesanti, C., Jacques, C., Pimental, E., \& Napoleao, T.\ 2006,
532:  Colesanti, C., et al.\ 2006,
533: \iaucirc, 8658, 2
534: 
535: \bibitem[Davoust \& Contini(2004)]{davoust2004} 
536: Davoust, E. \& Contini, T.\ 2004, \aap, 416, 515
537: 
538: \bibitem[Gamezo et al.(2003)]{gamezo2003}
539: Gamezo, V., Khokhlov, A.~M., Oran, E.~S., Chtchelkanova, A., \&
540: Rosenberg, R.\ 2003, Science, 299, 77
541: 
542: \bibitem[Gamezo, Khokhlov \& Oran(2004)]{gamezo2004}
543: Gamezo, V.~N., Khokhlov, A.~M., \& Oran, E.~S.\ 2004, \prl, 92, 1102
544: 
545: \bibitem[Garavini et al.(2005)]{garavini2005}
546: Garavini, G., et. al.\ 2005, \aj, 130, 2287
547: 
548: \bibitem[Garnavich et al.(1998)]{garnavich98} 
549: Garnavich, P.~M., et al.\ 1998, \apjl, 493, L53 
550: 
551: \bibitem[Garnavich et al.(2004)]{garnavich2004}
552: Garnavich, P.~M., et al.\ 2004, \apj, 613, 1120
553: 
554: \bibitem[Guy et al.(2005)]{guy2005}
555: Guy, J., Astier, P., Nobili, S., Regnault, N., \& Pain, R.\ 2005, \aap,
556: 443, 781
557: 
558: % \bibitem[Goldhaber et al.(2001)]{goldhaber2001}
559: % Goldhaber, G., et al.\ 2001, \apj, 558, 359
560: 
561: % \bibitem[Hauschildt \& Baron(1999)]{hauschildt1999}
562: % Hauschildt, P.~H.\ \& Baron, E.\ 1999, J.\ Comp.\ Applied Math., 109, 41
563: 
564: \bibitem[H\"oflich et al.(2002)]{hoflich2002}
565: H\"oflich, et al.\ 2002, \apj, 568, 791
566: 
567: \bibitem[Jeffery \& Branch(1990)]{jeffery1990}
568: Jeffery, D.~J.\ \& Branch, D.\ 1990, Jerusalem Winter School for 
569: Theoretical Physics, Vol.\ 6, ed.\ J.~C.~Wheeler \& S.~Weinberg
570: (Singapore: World Scientific), 149
571: 
572: \bibitem[Jha(2006)]{jha2006}
573: Jha, S., et al.\ 2006, \aj, 132, 189
574: 
575: \bibitem[Khokhlov(1991)]{khokhlov1991}
576: Khokhlov, A.\ 1991, \apj, 245, 114
577: 
578: \bibitem[Kozma et al.(2005)]{kozma2005}
579: Kozma, C., et al.\ 2005, \aap, 437, 983
580: 
581: \bibitem[Lantz et al.(2004)]{lantz2004} 
582: Lantz, B., et al.\ 2004, \procspie, 5249, 146 
583: 
584: \bibitem[Marion et al.(2006)]{marion2006}
585: Marion, G.~H., et al.\ 2006, \apj, 645, 1392
586: 
587: % \bibitem[Maza(1994)]{maza1994}
588: % Maza, J., et al.\ 1994, \apjl, 424, 107
589: 
590: \bibitem[Mazzali(2001)]{mazzali2001}
591: Mazzali, P.~A.\ 2001, \mnras, 321, 341
592: 
593: \bibitem[Nomoto, Thielemann \& Yokoi(1984)]{nomoto1984}
594: Nomoto, K., Thielemman, F.-K., \& Yokoi, K.\ 1984, \apj, 286, 644
595: 
596: \bibitem[Nugent et al.(1995)]{nugent1995}
597: Nugent, P., et al.\ 1995, \apjl, 455, 147
598: 
599: \bibitem[Perlmutter et al.(1998)]{perl98} 
600: Perlmutter, S., et al.\ 1998, \nat, 391, 51 
601: 
602: \bibitem[Perlmutter et al.(1999)]{perl99} 
603: Perlmutter, S., et al.\ 1999, \apj, 517, 565
604: 
605: \bibitem[Riess et al.(1998)]{riess98} 
606: Riess, A.~G., et al.\ 1998, \aj, 116, 1009 
607: 
608: \bibitem[R\"opke \& Hillebrandt(2005)]{ropke2005}
609: R\"opke, F.~K.\ \& Hillebrandt, W.\ 2005, \aap, 431, 635
610: 
611: \bibitem[R\"opke et al.(2006)]{ropke2006}
612: R\"opke, F.~K., Hillebrandt, W., Niemeyer, J.~C., \& Woosley, S.~E.\
613: 2006, \aap, 448, 1
614: 
615: \bibitem[Ruiz-Lapuente(2004)]{ruiz2004}
616: Ruiz-Lapuente, P.\ 2004, \apj, 612, 357
617: 
618: \bibitem[SNfactory(2006)]{atel}
619: The Nearby Supernova Factory 2006, ATel, 689, 1
620: 
621: \bibitem[Thomas et al.(2002)]{thomas2002}
622: Thomas, R.~C., Kasen, D., Branch, D., Baron, E.\ 2002, \apj, 567, 1037
623: 
624: \end{thebibliography}
625: 
626: \clearpage
627: 
628: \begin{deluxetable}{llccccl}
629:    \tabletypesize{\scriptsize}
630:    \tablecaption{Journal of Spectroscopy of \snsd\label{tab:spectra}}
631:    \tablewidth{0pt}
632:    \tablehead
633:    {
634:       \colhead{Julian Day}               &
635:       \colhead{UTC Date}                 &
636:       \colhead{Exposure}                 &
637:       \colhead{Airmass}                  &
638:       \multicolumn{2}{c}{Median S/N\tablenotemark{a}} &
639:       \colhead{Conditions}               \\
640:       \colhead{}                         &
641:       \colhead{(2006)}                   &
642:       \colhead{(s)}                      &
643:       \colhead{}                         &
644:       \colhead{blue}                     &
645:       \colhead{red}                      &
646:       \colhead{}      
647:    }
648:    \startdata
649:    2453750.10 & Jan.\ 14.60 &  900 & 1.20 & 76  & 85  & non-photometric \\
650:    2453752.12 & Jan.\ 16.62 & 1000 & 1.16 & 72  & 69  & cloudy          \\
651:    2453756.07 & Jan.\ 20.57 & 1800 & 1.22 & 205 & 136 & near Moon       \\
652:    2453770.11 & Feb.\ 03.61 & 1000 & 1.16 & 72  & 93  & photometric     \\
653:    \enddata
654:    \tablenotetext{a}{
655:    Blue channel: 3500-5050 \AA\ with 2.4 \AA\ bins. Red channel:
656:    5000-9000 \AA\ with 3.0 \AA\ bins.}
657: \end{deluxetable}
658: 
659: \clearpage
660: 
661: \begin{figure}
662:    \centering
663:    \includegraphics[scale=1.00,clip=true]{f1.eps} % {time_series}
664:    \caption{SNIFS spectroscopy of \snsd.  Phases are expressed relative
665:    to a $B$-band peak brightness date of \peakdate.  Dotted vertical
666:    lines mark rest wavelengths of \ion{C}{2} lines \wl\wl 4267, 4745,
667:    6580, and 7234.  Dark bands indicate blueshifts between 10,000 and
668:    14,000 \kms\ with respect to these lines, typical of the velocity at
669:    the photosphere at these phases.  The inset is a zoom of the region
670:    around the 6320~\AA\ notch in the $-7$ day spectrum.}
671:    \label{fig:time_series}
672: \end{figure}
673: 
674: \clearpage
675: 
676: \begin{figure}
677:    \centering
678:    \includegraphics[width=1.00\textwidth,clip=true]{f2.eps} % {spec_comp_early}
679:    \caption{Comparison of the $-7$ day spectrum of \snsd\ to premaximum
680:    spectra of SNe~1998aq, 1986G, and 1999by.}
681:    \label{fig:spec_comp_early}
682: \end{figure}
683: 
684: \clearpage
685: 
686: \begin{figure}
687:    \centering
688:    \includegraphics[width=1.00\textwidth,clip=true]{f3.eps} % {synow_06_014_088_001}
689:    \caption{Synthetic spectra overlaid on the $-7$ day spectrum of
690:    \snsd.  Inset plots show the coincidence between the synthetic
691:    \ion{C}{2} spectrum and the data.}
692:    \label{fig:synow_m10}
693: \end{figure}
694: 
695: \clearpage
696: 
697: \begin{figure}
698:    \centering
699:    \includegraphics[width=0.4\textwidth,clip=true]{f4.eps} % {c_comp}
700:    \caption{Comparison of \ion{C}{2} \wl 6580 feature across several
701:    premaximum \snia\ spectra.  The spectra are resampled into 10~\AA\
702:    bins.}
703:    \label{fig:c_comp}
704: \end{figure}
705: 
706: \end{document}
707: