0712.2823/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\documentclass[preprint2]{aastex}
4: \documentclass[apj]{emulateapj}
5: \usepackage{apjfonts}
6: 
7: \newcommand{\etal}{et~al.}
8: \newcommand{\eg}{e.g., }
9: \newcommand{\ie}{i.e., }
10: \newcommand{\Msun}{M_{\odot}}
11: \newcommand{\kms}{km~s$^{-1}$}
12: \newcommand{\ergs}{erg~s$^{-1}$}
13: \newcommand{\Fefs}{$^{56}$Fe}
14: \newcommand{\Cofs}{$^{56}$Co}
15: \newcommand{\Nifs}{$^{56}$Ni}
16: \newcommand{\Mms}{M_{\rm MS}}
17: \newcommand{\Mej}{M_{\rm ej}}
18: \newcommand{\KE}{E_{\rm K}}
19: \newcommand{\vph}{v_{\rm ph}}
20: \newcommand{\Tph}{T_{\rm ph}}
21: \def\gsim{\mathrel{\rlap{\lower 4pt \hbox{\hskip 1pt $\sim$}}\raise 1pt
22: \hbox {$>$}}}
23: \def\lsim{\mathrel{\rlap{\lower 4pt \hbox{\hskip 1pt $\sim$}}\raise 1pt
24: \hbox {$<$}}}
25: \def\ion#1#2{{\rm #1}~{\sc #2}}
26: 
27: \shorttitle{The Outermost Ejecta of Type Ia Supernovae}
28: \shortauthors{Tanaka et al.}
29: 
30: \begin{document}
31: 
32: \title{The Outermost Ejecta of Type I\lowercase{a} Supernovae}
33: \author{
34: Masaomi Tanaka\altaffilmark{1},
35: Paolo A. Mazzali\altaffilmark{2,3,4},
36: Stefano Benetti\altaffilmark{5},
37: Ken'ichi Nomoto\altaffilmark{1,4,6},
38: Nancy Elias-Rosa\altaffilmark{2},
39: Rubina Kotak\altaffilmark{7},
40: Giuliano Pignata\altaffilmark{8,9},
41: Vallery Stanishev\altaffilmark{10}, and
42: Stephan Hachinger\altaffilmark{2}
43: }
44: 
45: \altaffiltext{1}{Department of Astronomy, Graduate School of Science, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-0003, Japan; mtanaka@astron.s.u-tokyo.ac.jp}
46: \altaffiltext{2}{Max-Planck-Institut f\"{u}r Astrophysik, Karl-Schwarzschild-Str. 1, D-85741 Garching bei M\"{u}nchen, Germany}
47: \altaffiltext{3}{Istituto Nazionale di Astrofisica - Osservatorio Astronomico di Triste, Via Tiepolo 11, I-34131 Triste, Italy}
48: \altaffiltext{4}{Research Center for the Early universe, School of Science, University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan}
49: \altaffiltext{5}{Istituto Nazionale di Astrofisica - Osservatorio Astronomico di Padova, vicolo dell'Osservatorio 5 - 35122 Padova, Italy}
50: \altaffiltext{6}{Institute for the Physics and Mathematics of the Universe, 
51: University of Tokyo, Kashiwa, Chiba 277-8582, Japan}
52: \altaffiltext{7}{Astrophysics Research Centre, School of Mathematics and Physics, Queen's University Belfast, BT7 1NN, UK}
53: \altaffiltext{8}{Departamento de Astronom\'ia y Astrof\'isica, Pontificia Universidad Cat\'olica de Chile, Casilla 306, Santiago 22, Chile}
54: \altaffiltext{9}{Departamento de Astronom\'ia, Universidad de Chile, Casilla 36-D, Santiago, Chile}
55: \altaffiltext{10}{Department of Physics, Stockholm University, AlbaNova University Center, SE-10691 Stockholm, Sweden}
56: 
57: 
58: \begin{abstract}
59: The properties of the highest velocity ejecta of normal Type Ia supernovae 
60: (SNe Ia) are studied via models of very early optical spectra of 6 SNe.
61: At epochs earlier than 1 week before maximum, SNe with a rapidly evolving 
62: \ion{Si}{ii} $\lambda$6355 line velocity (HVG) have a larger photospheric 
63: velocity than SNe with a slowly evolving \ion{Si}{ii} $\lambda$6355 line 
64: velocity (LVG). 
65: Since the two groups have comparable luminosities, 
66: the temperature at the photosphere is higher in LVG SNe. 
67: This explains the different overall spectral appearance of HVG and LVG SNe.
68: However, the variation of the \ion{Ca}{ii} and \ion{Si}{ii} absorptions at the 
69: highest velocities ($v \gsim 20,000$ \kms) suggests that additional factors, 
70: such as asphericity or different abundances in the progenitor white dwarf, 
71: affect the outermost layers.
72: The \ion{C}{ii} $\lambda$6578 line is marginally detected in 3 LVG SNe,  
73: suggesting that LVG undergo less intense burning.
74: The carbon mass fraction is small, only less than 0.01 near the photosphere, 
75: so that he mass of unburned C is only $\lsim 0.01 \Msun$.
76: Radioactive \Nifs\ and stable Fe are detected in both LVG and HVG SNe.
77: Different Fe-group abundances in the outer layers may be one of the reasons 
78: for spectral diversity among SNe Ia at the earliest times.
79: The diversity among SNe Ia at the earliest phases could also indicate an 
80: intrinsic dispersion in the width-luminosity relation of the light curve.
81: \end{abstract}
82: 
83: \keywords{supernovae: general --- supernovae: individual 
84: (SN 2001el, SN 2002bo, SN 2002dj, SN 2002er, SN 2003cg, SN 2003du)
85: --- radiative transfer ---  line: profiles}
86: 
87: 
88: \section{Introduction}
89: \label{sec:intro}
90: 
91: Type Ia supernovae (SNe Ia) are believed to be thermonuclear explosions 
92: of CO white dwarfs (WDs).
93: Because of their high luminosity together with a tight
94: relation between the maximum brightness and decline rate of the light curve
95: (LC), SNe Ia are one of the most accurate distance indicators 
96: (\eg Phillips 1993; Riess et al. 1998; Perlmutter et al. 1999; 
97: Knop et al. 2003; Riess et al. 2004; 
98: Astier et al. 2006; Wood-Vasey et al. 2007).
99: 
100: In addition to the homogeneity of the LC, 
101: optical spectra of SNe Ia are also rather homogeneous. 
102: The ratio of two \ion{Si}{ii} lines in spectra at maximum
103: correlates well with the SN luminosity (Nugent et al. 1995).
104: The width of the Fe emission line in the nebular spectra 
105: also correlates with the LC decline rate (Mazzali et al. 1998).
106: These works show that the spectral properties both at maximum brightness 
107: and at the nebular phases are determined primarily by the maximum luminosity.
108: 
109: However, recent studies using large samples revealed large diversity in the 
110: early spectra of SNe Ia, in particular in the line velocities (see \eg Hatano
111: et al. 2000; Benetti et al. 2005, hereafter B05; Branch et al. 2006). B05
112: divided SNe Ia into three groups according to the post-maximum evolution of the
113: \ion{Si}{ii} $\lambda$6355 line velocity.  High Velocity Gradient (HVG) SNe
114: have a large \ion{Si}{ii} line velocity that decreases rapidly after maximum.
115: On the other hand, Low Velocity Gradient (LVG) SNe have a lower  velocity with
116: a slower post-maximum evolution (see Fig. 1 in B05).  HVG and LVG SNe have
117: similar maximum luminosities, implying that line velocity is independent of
118: luminosity or LC shape (see Fig. 3 in Hachinger, Mazzali \& Benetti 2006).
119: FAINT SNe are characterized by a lower luminosity and a low \ion{Si}{ii} line
120: velocity, whose evolution is comparable to that of LVG SNe. 
121: 
122: The origin of the spectral diversity is not yet fully understood, reflecting
123: uncertainties on the explosion mechanism. Despite general consensus that 
124: burning starts as a subsonic deflagration (Nomoto, Sugimoto \& Neo 1976;
125: Nomoto, Thielemann \& Yokoi 1984), it is still unclear whether or not a
126: transition to a supersonic detonation (Khokhlov 1991b) occurs. 
127: In recent years, the explosion mechanism has been studied by 
128: detailed numerical simulations free from spherically symmetry 
129: (\eg Reinecke,
130: Hillebrandt, \& Niemeyer 2002; Gamezo et al. 2003; Plewa, Calder \& Lamb 2004;
131: R\"opke \& Hillebrandt 2005; Gamezo, Khokhlov \& Oran 2005). 
132: 
133: Early phase spectra are an unique tool to investigate the state of the outer
134: layers of the ejecta, \ie the final fate of the burning front.  Furthermore, as
135: diversity in line velocities is largest at pre-maximum phases, modeling spectra
136: at these epochs is likely to yield clues as to the origins of this diversity.
137: Although several studies focussed on spectral diversity at maximum light (e.g.
138: Bongard et al. 2006; Branch et al. 2006; Hachinger et al. 2006), only few have
139: addressed pre-maximum spectra (e.g. Mazzali et al. 1993, Fisher et al. 1997;
140: Mazzali 2001; Branch et al. 2007).  This is mostly due to the lack of very
141: early data, which stems from the difficulty in discovering SNe well before
142: maximum.  In order to remedy this situation, the European Supernova
143: Collaboration (ESC \footnote{http://www.mpa-garching.mpg.de/\~{}rtn/}) has
144: endeavored to gather very early spectra of nearby SNe Ia.
145: 
146: One of the most interesting results of very early SN Ia spectroscopy is the
147: possible presence of unburned carbon in the outer layers. 
148: This may be a critical signature of the outer extent of burning. 
149: The \ion{C}{ii} $\lambda$6578 line has been detected in SN 1998aq 
150: (at $\sim$ 11,000 \kms, Branch et al. 2003) and in SN 1999ac  
151: (at $\sim$ 16,000 \kms, Garavini et al. 2005).  
152: The clearest \ion{C}{ii} features were seen in a pre-maximum spectrum of 
153: SN 2006D (at $\sim$ 12,000 \kms, Thomas et al. 2006)
154: \footnote{In SN 2006gz, the \ion{C}{ii} line is clearly seen at 
155: $\sim$ 15,000 \kms, although the SN is not normal and
156: possibly has a super-Chandrasekhar mass progenitor (Hicken et al. 2007),
157: as suggested for SN 2003fg (Howell et al. 2006)}.
158: On the other hand, \ion{C}{i} in near-infrared (NIR) spectra has never 
159: been firmly detected (Marion et al. 2006).
160: 
161: High-velocity absorption features (HVFs, $v \gsim 20,000$ \kms) in the 
162: \ion{Ca}{ii} IR triplet (\eg Hatano et al. 1999; Mazzali et al. 2005a;  
163: Thomas et al. 2004; Quimby et al. 2006) may be a feature of all SNe Ia at the 
164: earliest epochs (Mazzali et al. 2005b), and may indicate aspherical ejecta, as 
165: suggested by the high polarization level 
166: (Wang et al. 2003; Kasen et al. 2003).  
167: They may result from the explosion itself (Thomas et al. 2004; Mattila et al. 
168: 2005) or from the interaction of the ejecta with circumstellar medium 
169: (CSM; Gerardy et al. 2004; Mazzali et al. 2005a). 
170: The variation and distribution of the strength of \ion{Ca}{ii} HVFs may be 
171: explained by line-of-sight effects if HVFs result from aspherical structures 
172: like a few blobs or a thick disk at $v \sim 18-25,000$ \kms\ 
173: (Tanaka et al. 2006).
174: 
175: In this paper, we use SNe Ia pre-maximum optical spectra to study the 
176: physical properties of the ejecta,
177: modeling the spectra using a numerical code.
178: Details of the method are presented in \S \ref{sec:method}.
179: The results of the modeling are shown in \S \ref{sec:results}.
180: In \S \ref{sec:discussion}, our results are summarized
181: and the implications are discussed. 
182: Finally, conclusions are in \S \ref{sec:conclusions}.
183: 
184: 
185: 
186: 
187: \section{METHOD OF ANALYSIS}
188: \label{sec:method}
189: 
190: 
191: Spectra of 6 nearby SNe Ia observed earlier than 1 week before maximum 
192: are studied (Table \ref{tab:prop}).
193: The data were obtained by the ESC campaign.
194: The ESC has collected photometric and spectroscopic data of a number of 
195: nearby SNe Ia with very good temporal and wavelength coverage.
196: SNe with very early spectra include
197: SN 2001el (LVG; Mattila et al. 2005), 
198: SN 2002bo (HVG; Benetti et al. 2004), 
199: SN 2002dj (HVG; Pignata et al. 2005), 
200: SN 2002er (HVG; Pignata et al. 2004; Kotak et al. 2004),
201: SN 2003cg (LVG; Elias-Rosa et al. 2006),
202: and SN 2003du (LVG; Stanishev et al. 2007).
203: Our sample does not include FAINT SNe.
204: 
205: To investigate the physical conditions of the ejecta, such as temperatures,
206: ionization profiles, and element abundance distribution, 
207: we use a Monte Carlo spectrum
208: synthesis code (Mazzali \& Lucy 1993), which was used for several SNe  (\eg
209: Mazzali et al. 1993).  The code requires as input the bolometric luminosity
210: ($L$), photospheric velocity ($\vph$), and a model of the density and abundance
211: distribution in the SN ejecta.
212: 
213: The code assumes a sharply defined spherical photosphere.
214: The position of the photosphere is expressed in terms of a velocity 
215: thanks to the homologous expansion ($v \propto r$).
216: The temperature structure of the optically thin atmosphere is solved in 
217: radiative equilibrium tracing a large number of energy packets as they 
218: propagate in the SN ejecta.
219: 
220: Starting from a trial temperature distribution (radiation temperature, $T_R$),
221: the population of the excited level ($n_j$; $j=1$ for the ground state) is 
222: computed as
223: \begin{equation}
224: \frac{n_j}{n_1} = W \frac{g_j}{g_1} e^{-\epsilon_j /k_B T_R},
225: \label{eq:level}
226: \end{equation}
227: where $g_j$ and $\epsilon_j$ are the statistical weight  
228: and the excitation energy from the ground level, respectively.
229: Here W is the so called dilution factor, which is defined as
230: \begin{equation}
231: J = WB(T_R).
232: \end{equation}
233: In Eq\. (1), W is set to unity for metastable levels (Lucy 1999).
234: The ionization regime is computed using a modified nebular approximation 
235: (Abbott \& Lucy 1985, Mazzali \& Lucy 1993):
236: \begin{equation}
237: \frac{N_{i+1}N_e}{N_i} 
238: = \eta W \left( \frac{T_e}{T_R} \right) 
239:  \left( \frac{N_{i+1}N_e}{N_i}  \right)^{*}_{T_R},
240: \label{eq:ion}
241: \end{equation}
242: where $N_e$ is the electron density and $T_e$ the electron temperature.
243: The starred term on the right hand side is the value computed using the Saha 
244: equation with the temperature $T_R$; the partition functions entering the Saha 
245: equation are, however, computed with Eq. (\ref{eq:level}).
246: In Eq\. (\ref{eq:ion}), $\eta$ is defined as
247: \begin{equation}
248: \eta = \delta \zeta + W(1-\zeta),
249: \end{equation}
250: where $\zeta$ is the fraction of recombinations going directly to the ground 
251: state. Mazzali \& Lucy (1993) introduced a correction factor 
252: $\delta$ for an optically thick continuum at shorter wavelengths 
253: than the \ion{Ca}{ii} ionization edge 
254: (for a detailed definition, see Eqs. (15) and (20) in Mazzali \& Lucy 1993).
255: The second term in Eq\. (4) represents ionization from excited levels
256: (Mazzali \& Lucy 1993).
257: 
258: The Monte Carlo radiation packets are followed through the atmosphere, 
259: where they can undergo electron scattering or interact with spectral lines.
260: For line scattering, Sobolev approximation,
261: which is a sound approximation in a rapidly expanding medium, is applied.
262: From the level populations, a Sobolev line opacity is obtained: 
263: \begin{equation}
264: \tau_{lu} = \frac{hct}{4\pi} (B_{lu}n_l - B_{ul}n_u).
265: \end{equation}
266: Here $t$ is the time since the explosion and $B_{lu}$ and $B_{ul}$ 
267: are Einstein $B$-coefficients.
268: The effect of photon branching is considered
269: (Lucy 1999; Mazzali 2000; see also Pinto \& Eastman 2000).
270: 
271: The Monte Carlo experiment gives a flux in each radial point, and 
272: then a frequency moment
273: \begin{equation}
274: \bar{\nu} = \frac{\int \nu J_{\nu} d\nu}{\int J_{\nu} d\nu}
275: \end{equation}
276: is computed.
277: Hence the temperature structure can be determined via
278: \begin{equation}
279: \bar{x} = \frac{h\bar{\nu}}{k_B T_R}.
280: \end{equation}
281: Here $\bar{x}$ represents for mean energy of blackbody radiation
282: ($\bar{x} = 3.832$).
283: The electron temperature is crudely assumed to be $T_e = 0.9 T_R$ to simulate 
284: approximately\footnote{This assumption does not affect the spectral fitting, 
285: and thus the results in this paper, as ionization is determined mostly by the 
286: exponential part in Eq. (\ref{eq:ion}) computed with $T_R$, which is a direct 
287: result of the Monte Carlo simulation.} 
288: the situation where the electron temperature is 
289: largely controlled by radiation (Klein \& Castor 1978; Abbott \& Lucy 1985).
290: Level populations, ionizations and opacities are then updated 
291: using the temperature structure derived from the Monte Carlo simulation. 
292: This process is iterated until convergence.
293: 
294: 
295: \begin{deluxetable*}{lrccccccc} 
296: \tablewidth{0pt}
297: \tablecaption{Properties of the SNe and the spectra}
298: \tablehead{
299: Name (group)& 
300: Epoch \tablenotemark{a} &
301: $\Delta m_{15}$ \tablenotemark{b} &
302: $E(B-V)$ \tablenotemark{c} &
303: $\mu$ \tablenotemark{d} &
304: $v_{\rm Si}$ \tablenotemark{e} & 
305: $v_{\rm Ca, ph}$ \tablenotemark{f} & 
306: $v_{\rm Ca, hv}$ \tablenotemark{g} & 
307: References 
308: }
309: \startdata
310: SN 2002bo (HVG \tablenotemark{h}) & $ -8.0~$ & 1.17 & 0.38 & 31.67 &15.0 & 14.9 & 22.1 & 1 \\
311: SN 2002dj (HVG) 		  & $-11.0~$ & 1.12 & 0.10 & 32.98 &16.8 & 17.7 & 27.6 & 2 \\
312: SN 2002er (HVG) 		  & $ -7.4~$ & 1.33 & 0.36 & 32.90 &13.0 & 15.6 & 23.1 & 3, 4 \\
313: SN 2001el (LVG \tablenotemark{i}) & $ -9.0~$ & 1.15 & 0.22 & 31.26 &14.0 & 17.1 & 23.8 & 5 \\
314: SN 2003cg (LVG) 		  & $ -7.6~$ & 1.25 & 1.36 & 31.28 &12.0 & 12.7 & 22.0 & 6 \\
315: SN 2003du (LVG) 		  & $-11.0~$ & 1.02 & 0.01 & 32.79 &12.0 & 15.5 & 22.5 & 7 \\ 
316: \enddata
317: \tablenotetext{a}{Days since B maximum} 
318: \tablenotetext{b}{Difference of B magnitude between maximum and 15 days later}
319: \tablenotetext{c}{Color excess 
320: [sum of that caused by our Galaxy and the SN host galaxy;
321: $R_V = 3.1$ is adopted except for the host galaxy of SN 2003cg 
322: ($R_V = 1.8, E(B-V) = 1.33 $; Elias-Rosa et al. 2006)]}  
323: \tablenotetext{d}{Distance modulus (see literature)}
324: \tablenotetext{e}{Doppler velocity of \ion{Si}{ii} $\lambda$6355 absorption minimum ($10^3$ \kms)} 
325: \tablenotetext{f}{Doppler velocity of the photospheric component of Ca IR triplet ($10^3$ \kms). see Mazzali et al. (2005b)} 
326: \tablenotetext{g}{Doppler velocity of the high-velocity component of Ca IR triplet ($10^3$ \kms). see Mazzali et al. (2005b)} 
327: \tablenotetext{h}{High Velocity Gradient (a group of SNe Ia 
328: whose Si line velocity declines rapidly 
329: (see \S \ref{sec:intro}; B05), shown in blue throughout the paper)}
330: \tablenotetext{i}{{L}ow Velocity Gradient (a group of SNe Ia 
331: whose Si line velocity declines slowly 
332: (see \S \ref{sec:intro}; B05), shown in green throughout the paper)}
333: \tablerefs{(1) Benetti et al. (2004);
334: (2) Pignata et al. 2005; (3) Pignata et al. (2004); 
335: (4) Kotak et al. (2005); (5) Mattila et al. (2005);
336: (6) Elias-Rosa et al. (2006); (7) Stanishev et al. (2007)}
337: \label{tab:prop}
338: \end{deluxetable*}
339: 
340: We use the W7 deflagration model (Nomoto et al. 1984) 
341: as a standard density profile.
342: Alternatively, a density structure of a delayed detonation model is used 
343: (model WDD2 in Iwamoto et al. 1999, hereafter, DD model).
344: The density structures of these models are most different in 
345: layers above $v = 20,000$ \kms.
346: The density in the DD model at $v \sim 25,000$ \kms\ is higher than that of 
347: W7 by a factor of $\sim 10$ (see Fig. 11 in Baron et al. 2006).
348: Thus the ``density enhancement'' referred to in Mazzali et al. (2005a) is 
349: equivalent to the structure of the DD model if the densities are increased 
350: by a factor of $\lsim 10$.
351: The kinetic energy of the DD model ($1.4 \times 10^{51}$ ergs) is
352: slightly larger than that of W7 ($1.3 \times 10^{51}$ ergs).
353: 
354: Element abundance distributions (\ie mass fractions of elements) 
355: are parameterized for better reproduction of the
356: observed spectra. Initially, a homogeneous abundance distribution is assumed,
357: averaging the abundance distribution of W7 above the photospheric velocity, and
358: a general fit of the spectrum is obtained adjusting the abundances. In this
359: process, the input parameters $L$ and $\vph$ are also optimized. Then a
360: stratified abundance distribution is adopted and optimized to reproduce the
361: details. In this paper, we employ two different-abundance zones except for SN
362: 2003cg (3 zones are used; see \S \ref{sec:03cgSi}).
363: 
364: The abundances in the outermost ejecta are further checked against the spectra 
365: at maximum.
366: Although the photosphere recedes with time as the ejecta expand, 
367: the abundances at high velocities can still 
368: affect the synthetic spectra at later epochs (Stehle et al. 2005).
369: Therefore we modelled the spectra at maximum using the abundance ratios
370: derived from the pre-maximum spectra for the region at 
371: $v > \vph ({\rm early})$, where $\vph ({\rm early})$ is the 
372: photospheric velocity of the pre-maximum spectra.
373: If the result is inconsistent (\eg incorrect high-velocity absorptions
374: resulting from the pre-maximum spectra photospheric component),
375: the pre-maximum spectra are re-investigated until consistence is achieved.
376: 
377: 
378: 
379: \section{RESULTS}
380: \label{sec:results}
381: 
382: \begin{deluxetable}{lrccc} 
383: \tablewidth{0pt}
384: \tablecaption{Parameters of the synthetic spectra}
385: \tablehead{
386: Name (group) & 
387: Epoch & 
388: log($L$) \tablenotemark{a} & 
389: $\vph$ \tablenotemark{b} &
390: $T_{\rm ph}$ \tablenotemark{c} 
391: }
392: \startdata
393: SN 2002bo (HVG) & $ -8.0$  & 42.78 & 12,900 & 10,900 \\
394: SN 2002dj (HVG) & $-11.0$  & 42.44 & 14,000 & 10,200 \\
395: SN 2002er (HVG) & $ -7.4$  & 42.98 &  9,500 & 17,300 \\
396: SN 2001el (LVG) & $ -9.0$  & 42.87 & 10,500 & 17,100 \\
397: SN 2003cg (LVG) & $ -7.6$  & 42.84 &  9,000 & 17,000 \\
398: SN 2003du (LVG) & $-11.0$  & 42.71 & 10,000 & 18,600 \\ 
399: \enddata
400: \tablenotetext{a}{Bolometric luminosity (\ergs) in log} 
401: \tablenotetext{b}{Photospheric velocity (\kms)} 
402: \tablenotetext{c}{Photospheric temperature (K) computed by the code 
403: taking the back-scattering effect into account}
404: \label{tab:param}
405: \end{deluxetable}
406: 
407: \begin{deluxetable*}{lrcccccccc} 
408: \tablewidth{0pt}
409: \tablecaption{Mass fractions used for the synthetic spectra}
410: \tablehead{
411: Name (group) &
412: Epoch & 
413: $X$(C)\tablenotemark{a} &
414: $X$(Si)\tablenotemark{b} &
415: $X$(S)\tablenotemark{c} &
416: $X$(Ca)\tablenotemark{d} &
417: $X_{\rm HV}^{\rm W7}$(Ca)\tablenotemark{e} &
418: $X_{\rm HV}^{\rm DD}$(Ca) \tablenotemark{f} &
419: $X$(Fe)$_0$ \tablenotemark{g} &
420: $X$(\Nifs)$_0$ \tablenotemark{h} 
421: }
422: \startdata
423: SN 2002bo (HVG)  & $-8.0$  & --- & 0.51 & 0.10 & 0.02 &
424: 0.20/0.02 & 0.02/0.006 & 0.020  & 0.045\\
425: 
426: SN 2002dj (HVG) & $-11.0$ & --- & 0.50 & 0.09 & 0.02 &
427: $>$0.80/0.08 & 0.02/0.005 & 0.005 & 0.0052\\
428: 
429: SN 2002er (HVG) & $-7.4$  &  --- & 0.76  & 0.12 &  0.05 &
430: 0.40/0.05 & 0.05/0.007 & 0.012-0.013  & $<$ 0.025\\
431: 
432: SN 2001el (LVG) & $-9.0$  & 0.003 & 0.42 & 0.05 & 0.02 &
433: $>$ 0.80/0.10 & 0.10/0.02 & 0.048 & 0.19 \\
434: 
435: SN 2003cg (LVG) & $-7.6$  & $<$ 0.002 & 0.40 (0.08\tablenotemark{i}) & 0.080 & 0.02 & 
436: 0.03/0.005 & 0.001/0.0001 & 0.0033-0.0036 & $<$ 0.0088 \\
437: 
438: SN 2003du (LVG) & $-11.0$ & 0.002  & 0.29  & 0.060 & 0.016 &
439: 0.70/0.11 & 0.10/0.015 &  0.003 & 0.005  \\
440: \enddata
441: \tablenotetext{a}{Mass fraction of C at the photosphere} 
442: \tablenotetext{b}{Mass fraction of Si at the photosphere} 
443: \tablenotetext{c}{Mass fraction of S at the photosphere} 
444: \tablenotetext{d}{Mass fraction of Ca at the photosphere} 
445: \tablenotetext{e}{Mass fraction of Ca in the high-velocity layers 
446: with the density structure of W7 model without/with CSM interaction} 
447: \tablenotetext{f}{Mass fraction of Ca in the high-velocity layers
448: with the density structure of the DD model without/with CSM interaction}
449: \tablenotetext{g}{Mass fraction of {\it stable} Fe 
450: (Fe except for \Fefs\ from \Cofs\ decay) at the photosphere}
451: \tablenotetext{h}{Mass fraction of \Nifs\ at the photosphere 
452: (at the time of the explosion)}
453: \tablenotetext{i}{Mass fraction of Si in the intermediate shell at 
454: $v=17,000 - 20,500$ \kms}
455: \label{tab:abun}
456: \end{deluxetable*}
457: 
458: 
459: In this section, we present the results of the modeling for each SN.
460: The properties of the SNe and the spectra are summarized 
461: in Table \ref{tab:prop}.
462: In addition to the parameters $L$ and $v_{\rm ph}$, and the abundances 
463: in the ejecta, the distance modulus ($\mu$) 
464: and the reddening [in terms of $E(B-V)$ 
465: and $R_V$] are also required to reproduce the observed spectra. 
466: Thus the luminosity and the epoch, which are often neglected by rescaling 
467: the flux arbitrarily, are important because $\vph$ is constrained not only by 
468: the line velocities but also by the relation 
469: $L \propto v_{\rm ph}^2 t^2 T_{eff}^4$.
470: Here $t$ is the time since the explosion and $T_{eff}$ the effective 
471: temperature. 
472: 
473: All spectra have been calibrated against the photometry. The epoch is estimated
474: assuming a rise time of 19 days for all objects for simplicity. 
475: We take $\mu$ and $E(B-V)$ from the literature (see Table \ref{tab:prop}).
476: Using all the quantities above, 
477: the photospheric temperature $T_{\rm ph}$ is computed in the code considering 
478: the back-scattering effect, and thus tends to be higher than 
479: $T_{eff}$ (Mazzali \& Lucy 1993; Mazzali 2000).
480: 
481: In order to facilitate the comparison of intrinsic colors, all spectra are
482: displayed after correcting for reddening.  The values of $L$ and $\vph$ derived
483: from modeling are shown in Table \ref{tab:param} along with the computed
484: photospheric temperature ($T_{\rm ph}$), while the abundances of the elements
485: at the photosphere that most affect the synthetic spectra 
486: are shown in Table \ref{tab:abun}.
487: The abundances in the outer shell are not strongly constrained
488: because of the lack of strong lines.
489: Only \ion{Ca}{ii} has high velocity lines (see following sections), and
490: the Ca mass fraction in the outer shell is shown in the table ($X_{\rm HV}$(Ca)).
491: The iron mass fraction $X$(Fe) indicates the sum of stable Fe and \Fefs\
492: produced by \Nifs\ decay at the time of the observation. If only stable Fe is
493: considered, this is explicitely denoted as $X$(Fe)$_0$, where the index ''0``
494: means ''at explosion``. Furthermore, we use $X$(\Nifs)$_0$ for the initial 
495: mass fraction of \Nifs\ at the time of the explosion.
496: 
497: 
498: 
499: 
500: \subsection{SN 2002bo (Day $-8$, HVG)}
501: 
502: \begin{figure}
503: \begin{center}
504: %\plotone{f01.eps}
505: \includegraphics[scale=0.85]{f01.eps} 
506: \figcaption{
507: Observed spectrum of the HVG SN 2002bo at day $-8$ (black solid) 
508: and a model with W7-CSM density/abundance structure (blue dashed).
509: The left inset shows the \ion{Ca}{ii} H\&K feature (mean $\lambda\sim3950$),
510: the right inset the \ion{Ca}{ii} IR feature (mean $\lambda\sim8567$). 
511: The W7-CSM spectrum is compared to W7/homogeneous abundance (thin cyan) and DD
512: (dotted red) models, both without CSM interaction. In both insets, the 
513: wavelength scale is replaced by the corresponding Doppler velocies.
514: \label{fig:02bo}}
515: \end{center}
516: \end{figure}
517: 
518: 
519: 
520: \subsubsection{General Properties}
521: 
522: The well-studied SN 2002bo (Stehle et al. 2005) is a member of the HVG group.
523: The spectrum at $-8$ days is shown in Figure \ref{fig:02bo}, together with our 
524: synthetic spectrum. Good agreement is obtained with $\log L$ (\ergs) $ = 42.78$ 
525: and $\vph = 12,900$ \kms, giving a $\Tph = 10,900$\,K.
526: The flux redward of $\lambda \sim 6000$ \AA\ is overestimated 
527: (see also Stehle et al. 2005).
528: This may be due to a deviation from the blackbody approximation. 
529: A modified input radiation field, where the blue region ($\lambda \lsim 5000$
530: \AA) has greater weight, can reproduce the spectrum better, with only a small 
531: difference in the photospheric temperature and ionization states compared to 
532: the original case.
533: 
534: 
535: \subsubsection{Iron-group Elements}
536: \label{sec:02boFe}
537: 
538: An absorption feature around 4300 \AA\ results mainly from
539: \ion{Mg}{ii} $\lambda$4481 and \ion{Fe}{iii} $\lambda$4420, with a smaller
540: contribution by \ion{Si}{iii} $\lambda$4553.
541: Around 4800 \AA, many \ion{Fe}{ii} and \ion{Si}{ii} lines, 
542: \eg \ion{Fe}{ii} $\lambda$5169 and \ion{Si}{ii} $\lambda$5056, 
543: produce the strong absorption feature.
544: To account for these strong features, the mass fraction of Fe, 
545: including both stable Fe and decay products of \Nifs, 
546: should be $X$(Fe)=0.022 at the photosphere.
547: 
548: Our method of estimating the Fe abundance is illustrated in Figure 
549: \ref{fig:02boFe}, right panel. For solar Fe abundance, the absorption at 
550: 4950 \AA\ (mainly \ion{Fe}{ii} $\lambda$5169) is weaker than the observed 
551: feature (red solid line).
552: On the other hand, Fe-rich ejecta [$X$(Fe)$\gsim 0.10$] produce too  
553: strong a feature (dotted line).
554: 
555: The abundance of radioactive \Nifs\ synthesized in the explosion can be 
556: estimated using the Ni and Co lines that are present around 2000-4000\AA.
557: Since at $t \sim$ 10 days around 30\% of \Nifs\ has not yet decayed, 
558: the strong \ion{Ni}{II} $\lambda$4067 line is useful to constrain 
559: the mass fraction.
560: The procedure is illustrated in Figure \ref{fig:02boFe}, left panel.
561: For a larger \Nifs\ mass fraction [$X$(\Nifs)$_0=0.45$], 
562: the absorption at 3900 \AA\ 
563: appears (red dotted line), which is inconsistent with the observed feature.
564: On the other hand, when only a solar abundance of \Nifs\ is assumed, the 
565: flux level at $\lambda < 3500$ \AA\ becomes too high (red solid line).
566: This is because of insufficient blocking of (near) UV photons by iron group 
567: lines.
568: [Our code treats the blocking by iron group elements correctly,
569: considering photon branching (Lucy 1999; Mazzali 2000)].
570: 
571: A \Nifs\ mass fraction $X$(\Nifs)$_0$ = 0.045 is derived.
572: Although the observed flux level at $\lambda < 3500$ \AA\ is often uncertain, 
573: the upper limit of $X$(\Nifs)$_0$ is strict because it is derived from 
574: the absence of \ion{Ni}{II} $\lambda$4067.
575: Therefore, most Fe detected above the photosphere must be stable 
576: [$X$(Fe)$_0$= 0.020].
577: 
578: 
579: \subsubsection{Silicon and Carbon}
580: 
581: The ratio of the two \ion{Si}{ii} absorptions at 5800\AA\ and 6000\AA\ 
582: is reproduced well.
583: A strong \ion{Si}{ii} $\lambda$6355 suggests that the photosphere resides 
584: in the Si-rich layer.
585: The mass fraction of Si is estimated as $X$(Si)$\sim$0.50.
586: The emission peak of \ion{Si}{ii} $\lambda$6355 is well developed,
587: and there is no evidence of \ion{C}{ii} $\lambda$6578.
588: The blue wing of the synthetic \ion{Si}{ii} absorption is weak, 
589: suggesting that the line may be contaminated by a HVF, 
590: as are the \ion{Ca}{ii} lines (see \S 3.1.4).
591: At 7500 \AA, a weak absorption is seen, which is often attributed to
592: \ion{O}{i} $\lambda$7774.
593: However, in the synthetic spectrum the feature is mainly due to 
594: \ion{Mg}{ii} $\lambda$7869 and \ion{Si}{ii} $\lambda$7849.
595: These lines are actually stronger than the \ion{O}{i} line if  
596: \ion{Mg}{ii} $\lambda$4481 and \ion{Si}{ii} $\lambda$6355 are reproduced.
597: 
598: 
599: \subsubsection{Calcium}
600: 
601: The spectrum around \ion{Ca}{ii} H\&K and the \ion{Ca}{ii} IR triplet is 
602: enlarged in the insets of Figure \ref{fig:02bo}.
603: The thin cyan line shows a synthetic spectrum computed with the original W7 
604: density structure and homogeneous abundances, setting $X$(Ca)=0.02.
605: While the observed features have a blue wing extending to $v \sim 30,000$ \kms\ 
606: (\ion{Ca}{ii} H\&K) and $27,000$ \kms\ (\ion{Ca}{ii} IR triplet), the wing of 
607: the synthetic absorption disappears at $v \sim 27,000$ \kms\ and 
608: $v \sim 23,000$ \kms, respectively.
609: This difficulty has been reported in many papers, \eg  Hatano et al. (1999) for 
610: SN 1994D, Gerardy et al. (2004) for SN 2003du, Mazzali et al. (2005a) for 
611: SN 1999ee and Mazzali et al. (2005b) for the objects in this paper.
612: To reproduce these features
613: \footnote{Here the width of the feature and the strength of the absorption 
614: relative to the continuum are compared because of the poor fit of the 
615: continuum level.}, 
616: we test several cases.
617: 
618: First, we modified the abundance distribution, keeping the W7 density structure. 
619: Although the features can be reproduced with $X$(Ca) $\gsim$ 0.20 at
620: $v > 22,000$ \kms, such a large fraction is very unlikely because the Ca mass
621: fraction in the most Ca-rich layer is $\lsim$ 0.1 in all explosion models
622: (see \eg Iwamoto et al. 1999). 
623: Secondly, when we use the DD density structure, the
624: observed profile can be reproduced with $X$(Ca)=0.02 at $v > 22,000$ \kms\
625: (insets of Fig. \ref{fig:02bo}, dotted red lines) because of the higher
626: density at $v > 20,000$ \kms.
627: 
628: Finally, we test the interesting possibility that interaction with CSM leads 
629: to some abundance mixing (Gerardy et al. 2004). 
630: As discussed by Mazzali et al. (2005a), when a hydrogen-rich CSM is mixed with 
631: the ejecta, the electron density increases, favouring recombination. 
632: Therefore some fraction of \ion{Ca}{iii}, which is the dominant state (see 
633: \S \ref{sec:HVG}), becomes \ion{Ca}{ii}, making the \ion{Ca}{ii} line stronger.
634: Here we assume that the interaction causes a density jump by a factor of 4 at 
635: $v > 22,000$ \kms, corresponding to $0.01 \Msun$ of material.
636: Hydrogen is assumed to be mixed into the shocked region and we find that 
637: H$\alpha$ does not appear as long as $X$(H) $< 0.30$
638: ($M$(H)$\sim 5 \times 10^{-3} \Msun$)
639: With this model, the mass fraction of Ca in the high-velocity layers is reduced 
640: to $X$(Ca)=0.02 (with the W7 density structure; dashed blue line in insets of 
641: Fig. \ref{fig:02bo}) and $X$(Ca)=0.006 (with the DD density structure), 
642: respectively. 
643: The deceleration following the interaction is not considered in our 
644: computation and a more realistic hydrodynamic calculation is needed to 
645: confirm this scenario.
646: 
647: \begin{figure}
648: \begin{center}
649: %\plotone{f02.eps}
650: \includegraphics[scale=0.85]{f02.eps} 
651: \figcaption{
652: SN 2002bo: Observed spectrum around 3500 \AA\ ({\it left}) and 4900 \AA\ 
653: ({\it right}) compared with synthetic profiles for different metal 
654: abundances.
655: \label{fig:02boFe}}
656: \end{center}
657: \end{figure}
658: 
659: 
660: 
661: \subsection{SN 2002dj (Day $-11$, HVG)}
662: 
663: 
664: \begin{figure}
665: \begin{center}
666: %\plotone{f03.eps}
667: \includegraphics[scale=0.85]{f03.eps} 
668: \figcaption{
669: Observed spectrum of the HVG SN 2002dj at day $-11$ (black solid) 
670: and a model with W7-CSM density/abundance structure (blue dashed).
671: As in Fig. \ref{fig:02bo}, the insets show \ion{Ca}{ii} H\&K (left) and 
672: \ion{Ca}{ii} IR (right). The W7-CSM spectrum is compared to W7 (thin cyan) 
673: and DD (dotted red) models, both with homogeneous abundance and without 
674: CSM interaction.
675: \label{fig:02dj}}
676: \end{center}
677: \end{figure}
678: 
679: 
680: \subsubsection{General Properties}
681: 
682: SN 2002dj has a spectral sequence and LC similar to SN 2002bo 
683: (Pignata et al. 2005), and is also an HVG SN.
684: The spectrum $11$ days before B maximum is well reproduced by a synthetic 
685: spectrum with 
686: $\log L$ (\ergs) = 42.44 and $\vph = 14,000$ \kms\ (Fig. \ref{fig:02dj}).
687: Almost all lines are the same as in the spectrum of SN 2002bo, although the 
688: velocities are higher because the epoch is earlier.
689: The photospheric temperature, 10,200 K, is also similar.
690: 
691: 
692: \subsubsection{Iron-group Elements}
693: 
694: \ion{Fe}{ii} and \ion{Fe}{iii} lines contribute to the features 
695: near 4200 and 4800 \AA, as in SN 2002bo, but they are weaker.
696: In fact, the absorption at 4200 \AA\ is caused mostly 
697: by \ion{Mg}{ii} $\lambda$4481.
698: Utilizing the feature around 4800 \AA\ (a combination of \ion{Si}{ii} and 
699: \ion{Fe}{ii} lines) as in \S \ref{sec:02boFe}, the mass fraction 
700: of Fe is evaluated as $X$(Fe) $\sim 0.005$ at the photosphere.
701: The features at 3500 and 3900\AA\ require $X$(\Nifs)$_0$ = 0.005.
702: Therefore, the Fe causing the observed spectral features is again 
703: mostly stable Fe.
704: Although the values are smaller than those of SN 2002bo, they are still 
705: larger than the solar abundances by a factor of 3.5 and 70 for Fe and Ni, 
706: respectively.
707: 
708: 
709: \subsubsection{Silicon and Carbon}
710: 
711: The \ion{Si}{ii} $\lambda$5979 line is quite strong, which is consistent with 
712: the low photospheric temperature in the synthetic spectrum.
713: The \ion{Si}{ii} $\lambda$6355 line in SN 2002dj is the broadest of the 6 SNe.
714: In the model, the mass fraction of Si is $X$(Si)$= 0.50$ at the photosphere. 
715: The synthetic profile does not show enough absorption in the blue wing, as was 
716: the case for SN 2002bo, suggesting the presence of a \ion{Si}{ii} HVF.
717: The emission profile of \ion{Si}{ii} $\lambda$6355 is well developed and there 
718: is no evidence of an absorption by \ion{C}{ii} $\lambda$6578.
719: 
720: 
721: \subsubsection{Calcium}
722: 
723: The \ion{Ca}{ii} HVF in SN 2002dj is the strongest of the 6 SNe. Even assuming
724: $X$(Ca)=0.80 at the corresponding velocity range, the profile is not reproduced
725: with the W7 density structure (the thin cyan line in insets of Fig. \ref{fig:02dj} 
726: is a homogeneous abundance model with $X$(Ca)=0.02). 
727: The blue dashed line is a synthetic spectrum obtained with a density jump at 
728: $v > 23,000$ \kms\ (adding 0.005 $\Msun$), where $X$(H) = 0.30 and 
729: $X$(Ca) = 0.10. This hydrogen mass fraction is an upper limit for the absence of
730: H$\alpha$ again ($M$(H) $\sim 2 \times 10^{-3} \Msun$). 
731: Using the DD density structure, the Ca mass fraction at high-velocity layers 
732: is reduced to $X$(Ca)=0.02 (assuming no CSM interaction, red dotted lines in 
733: the insets of Fig. \ref{fig:02dj}) and $X$(Ca)=0.005 (assuming CSM 
734: interaction), respectively.
735: 
736: 
737: \subsection{SN 2002er (Day $-7.4$, HVG)}
738: 
739: \begin{figure}
740: \begin{center}
741: %\plotone{f04.eps}
742: \includegraphics[scale=0.85]{f04.eps}
743: \figcaption{
744: Observed spectrum of the HVG SN 2002er at day $-7.4$ (black solid) 
745: and a model with W7-CSM density/abundance structure (blue dashed).
746: As in fig. \ref{fig:02bo}, the left inset shows \ion{Ca}{ii} H\&K, the right 
747: inset \ion{Ca}{ii} IR. The W7-CSM spectrum is compared to W7 (thin cyan) and 
748: DD (dotted red) models, both with homogeneous abundance and without CSM 
749: interaction.
750: \label{fig:02er}}
751: \end{center}
752: \end{figure}
753: 
754: \subsubsection{General Properties}
755: 
756: The spectra of the HVG SN 2002er were studied by Kotak et al. (2005). We use 
757: the earliest spectrum that covers the \ion{Ca}{ii} IR triplet ($-7.4$ days).
758: A synthetic spectrum with $\log L$ (\ergs) $=  42.98$ and $\vph = 9,500$ \kms\
759: provides the best fit (Fig. \ref{fig:02er}).
760: The lower $\vph$ agrees with the fact that this SN has lower line velocities 
761: than other HVG SNe (Kotak et al. 2005; B05). At the same time, 
762: $T_{\rm ph}= 17,300$ K, higher than in both SNe 2002bo and 2002dj.
763: This results from the combination of the higher $L$ and lower $\vph$.
764: 
765: 
766: \subsubsection{Iron-group Elements}
767: 
768: \ion{Fe}{iii} lines contribute to the blue absorption near 
769: 4300\AA\ (\ion{Fe}{iii} $\lambda$4396, 4420, 4431).
770: At 4800\AA, \ion{Fe}{iii} lines (\eg $\lambda$5127) dominate over
771: \ion{Fe}{ii} lines (\eg $\lambda$5018), 
772: because the temperature is higher than in SNe 2002bo and 2002dj.
773: To get sufficiently strong Fe lines, $X$(Fe)=0.013 near the photosphere.
774: Although the lack of spectral coverage at $\lambda < 3800$\AA\ makes it 
775: difficult to estimate the abundances of Ni and Co (see \S \ref{sec:02boFe}), 
776: an upper limit for these elements can be estimated as $X$(\Nifs)$_0<$0.025
777: considering the absence of \ion{Ni}{ii} $\lambda$4067.
778: Therefore, the mass fraction of stable Fe is $X$(Fe)$_0=$0.012-0.013.
779: 
780: 
781: \subsubsection{Silicon and Carbon}
782: 
783: Around 4300\AA, \ion{Si}{iii} ($\lambda$4553, 4658) lines are stronger than 
784: in other HVG SNe, indicating a higher temperature.
785: The blue wing of \ion{Si}{ii} $\lambda$6355 is also reproduced fairly well, 
786: and there is no clear evidence of a \ion{Si}{ii} HVF, 
787: unlike SNe 2002bo and 2002dj.
788: \ion{C}{ii} $\lambda$6578 is not seen, but the \ion{Si}{ii} emission peak 
789: is weaker than in SNe 2002bo and 2002dj.
790: A weak absorption near 7500\AA\ is mostly
791: due to \ion{Si}{ii} $\lambda$7849, and the contribution of 
792: \ion{Mg}{ii} $\lambda$7869 is small.
793: 
794: \subsubsection{Calcium}
795: 
796: The \ion{Ca}{ii} IR triplet is again not accounted for by the density 
797: structure of W7 (thin cyan lines in the insets of Fig. \ref{fig:02er} show 
798: the profile of one-zone abundance model with $X$(Ca)=0.05).
799: With the original W7 density structure, a Ca-dominated layer 
800: [$X$(Ca) $\sim$ 0.40] is required for the feature, which is highly unrealistic.
801: The line depth and width can be reproduced with a density jump at 
802: $v > 21,000$ \kms, increasing the mass by 0.015 $\Msun$ and 
803: containing $X$(H) = 0.30 ($M$(H)$\sim 5 \times 10^{-3} \Msun$) 
804: and $X$(Ca) = 0.05, which is similar to SNe 2002bo and 2002dj.
805: If the DD model density profile is used instead, the feature is reproduced with 
806: $X$(Ca)=0.05 at $v > 21,000$ \kms\ even without CSM interaction 
807: (insets of Fig. \ref{fig:02er}, dotted red lines).
808: If CSM interaction is considered, this is further reduced to $X$(Ca)=0.007.
809: 
810: 
811: \subsection{SN 2001el (Day $-9.0$, LVG)}
812: 
813: \begin{figure}
814: \begin{center}
815: %\plotone{f05.eps}
816: \includegraphics[scale=0.85]{f05.eps}
817: \figcaption{
818: Observed spectrum of the LVG SN 2001el at day $-9$ (black solid) and a model 
819: with W7-CSM density/abundance structure (green dashed). The red solid line 
820: shows the profile without C in the ejecta.
821: As in fig. \ref{fig:02bo}, the left inset shows \ion{Ca}{ii} H\&K, the right 
822: inset \ion{Ca}{ii} IR. The W7-CSM spectrum is compared to W7 (thin cyan) 
823: and DD (dotted red) models, both with homogeneous abundance and without 
824: CSM interaction.
825: \label{fig:01el}}
826: \end{center}
827: \end{figure}
828: 
829: \subsubsection{General Properties}
830: 
831: The LVG SN 2001el was studied in spectroscopy (Mattila et al. 2005) and 
832: spectropolarimetry (Wang et al. 2003).
833: The best-fitting synthetic spectrum has $\log L$ (\ergs) $= 42.87$ and 
834: $\vph = 10,500$ \kms\ (Fig. \ref{fig:01el}). 
835: This leads to $T_{\rm ph} = 17,400$ K, which is similar to SN 2002er, as can 
836: be expected since the spectra are similar.
837: 
838: 
839: \subsubsection{Iron-group Elements}
840: 
841: As in SN 2002er, the features around 4300 \AA\ and 4800 \AA\ are
842: attributed mostly to \ion{Fe}{iii} lines. 
843: An iron mass fraction $X$(Fe) $=$ 0.055 at the photosphere is needed 
844: to reproduce these features.
845: Relatively reliable flux data at $\lambda \sim 3300-3500$ \AA\ 
846: makes it possible to estimate the mass fractions of Ni and Co reliably 
847: (see \ref{sec:02boFe}):
848: $X$(\Nifs)$_0$=0.19 provides the best fit.
849: Thus the mass fraction of stable Fe is estimated as $X$(Fe)$_0$=0.048.
850: 
851: 
852: \subsubsection{Silicon and Carbon}
853: 
854: The narrow \ion{Si}{iii} feature near 4400 \AA\ has similar strength as in 
855: SN 2002er, supporting the similar photospheric velocity and temperature.
856: The \ion{Si}{ii} $\lambda$6355 absorption has a boxy profile and is not 
857: reproduced well by the synthetic spectrum, suggesting a contribution from HVFs.
858: Similar profiles were also seen in SNe 1990N 
859: (Leibundgut et al. 1991; Mazzali 2001) and 2005cf (Garavini et al. 2007).
860: 
861: The \ion{Si}{ii} $\lambda$6355 emission peak is not as round and 
862: well developed in SN 2001el as in SNe 2002bo and 2002dj.
863: The synthetic spectrum computed including C 
864: (Fig. \ref{fig:01el}, dashed green line) reproduces the feature well. 
865: If C is not included, the emission peak becomes too strong
866: (Fig. \ref{fig:01el}, red line).
867: This may be a sign of the presence of C. 
868: The C abundance required for the feature is, however, only $X$(C) $\sim 0.005$.
869: The mass of C is estimated as M(C) $\sim 0.02 \Msun$ assuming the W7 density 
870: structure and a homogeneous C mass fraction above the photosphere.
871: Since \ion{C}{ii} $\lambda$6578 is the strongest optical C line, the presence 
872: of this amount of C does not affect other parts of the spectrum.
873: 
874: 
875: \subsubsection{Calcium}
876: 
877: As in other SNe, the \ion{Ca}{ii} HVFs cannot be reproduced with the W7 density 
878: structure (the thin cyan line in insets of Fig.\ref{fig:01el} shows the profile 
879: of a homogeneous abundance model with $X$(Ca)=0.02).
880: The HVFs in SN 2001el are very strong and are not reproduced, even using 
881: $X$(Ca)=0.80 at high velocities.
882: A better match is obtained using a density jump of a factor of 4 at 
883: $v \gsim 21,500$ \kms, adding a mass of 0.015 $\Msun$, mixed-in hydrogen 
884: ($X$(H)=0.30, $M$(H)$\sim 5 \times 10^{-3} \Msun$), and $X$(Ca)=0.10 (insets 
885: of Fig \ref{fig:01el}, dashed green line), but the synthetic \ion{Ca}{ii} IR 
886: triplet is too narrow.
887: It may be made broader by increasing Ca mass fraction 
888: or using a larger density jump, but both 
889: changes seem unlikely and lead to a worse fit of \ion{Ca}{ii} H\&K.
890: A large Ca mass fraction, $X$(Ca)=0.10 in the high-velocity layers, 
891: is needed even with the DD density structure (insets of Fig. \ref{fig:01el}, dotted red lines).
892: If CSM interaction is considered for the DD model, this is, however, 
893: reduced to $X$(Ca)=0.02.
894: 
895: 
896: \subsection{SN 2003cg (Day $-7.6$, LVG)}
897: 
898: \begin{figure}
899: \begin{center}
900: %\plotone{f06.eps}
901: \includegraphics[scale=0.85]{f06.eps}
902: \figcaption{
903: Observed spectrum of the LVG SN 2003cg at day $-7.6$ (black solid) and a model 
904: with W7-CSM density/abundance structure (green dashed). This model uses  
905: stratification with three abundance zones (\S \ref{sec:03cgSi}).
906: As in Fig. \ref{fig:02bo}, the left inset shows \ion{Ca}{ii} H\&K, the right 
907: inset \ion{Ca}{ii} IR. The W7-CSM spectrum is compared to W7 (thin cyan) 
908: and DD (dotted red) models, both with homogeneous abundance and without 
909: CSM interaction.
910: \label{fig:03cg}}
911: \end{center}
912: \end{figure}
913: 
914: 
915: \subsubsection{General Properties}
916: The LVG SN 2003cg was studied by Elias-Rosa et al. (2006).
917: A synthetic spectrum with $\log L$ (\ergs) $=42.84$ and $\vph=9,000$\,\kms\ 
918: provides the best fit (Fig. \ref{fig:03cg}).
919: The model has $\Tph=17,000$\,K, similar to SNe 2002er and 2001el  
920: but higher than SNe 2002bo and 2002dj.
921: 
922: 
923: \subsubsection{Iron-group Elements}
924: 
925: To account for the strength of the \ion{Fe}{ii} and \ion{Fe}{iii} 
926: features at 4000-4900 \AA, $X$(Fe)$\sim$ 0.036.
927: We estimate an upper limit for \Nifs\ of $X$(\Nifs)$_0< 0.066$
928: from the absence of the \ion{Ni}{ii} $\lambda$4067 line.
929: The mass fraction of stable Fe is therefore $X$(Fe)$_0= 0.033-0.036$. 
930: A similar result is obtained in Elias-Rosa et al. (2006), but
931: the mass fraction of stable Fe they find is only $X$(Fe)$_0$=0.015.
932: The main difference is that $\vph$ is lower in our model,
933: leading to a higher temperature and suppressing \ion{Fe}{ii}. 
934: In our model, the fraction of \ion{Fe}{ii} is smaller than 
935: in Elias-Rosa et al. (2006) by a factor of 3.
936: 
937: 
938: \subsubsection{Silicon and Carbon}
939: \label{sec:03cgSi}
940: 
941: SN 2003cg has the narrowest \ion{Si}{ii} $\lambda$6355 line of our 6 SNe, 
942: although the line velocities are typical for LVG (Elias-Rosa et al. 2006).
943: To reproduce the narrow feature, we use a layer at 
944: $v \sim 17,000 - 20,500$ \kms\ with a smaller Si mass fraction 
945: [$X$(Si) $= 0.08$] than at the photosphere [$X$(Si) $= 0.40$].
946: There is no evidence of a \ion{Si}{ii}  HVF, which seems consistent with the 
947: weakness of the \ion{Ca}{ii} HVF (\S 3.5.4).
948: The emission peak is somewhat suppressed, as in SNe 2002er and 2001el.
949: Elias-Rosa et al. (2006) identified this as \ion{C}{ii} $\lambda$6578, who   
950: obtained a strong synthetic \ion{C}{ii} line useing $X$(C)=0.05.
951: We find that $0.002$ is an upper limit at the photosphere.
952: However, C is not necessarily needed and the identification is not conclusive.
953: 
954: 
955: \subsubsection{Calcium}
956: 
957: While the \ion{Ca}{ii} IR HVF might be present in SN 2003cg, it is the 
958: weakest of the 6 SNe.
959: In fact, a synthetic spectrum computed with the W7 density structure and 
960: homogeneous abundances [$X$(Ca)=0.02] (insets of Figure \ref{fig:03cg}, 
961: thin cyan line) provides a reasonably good match to the observed profile,
962: although a better match is obtained for a slightly higher Ca abundance 
963: [$X$(Ca)=0.03] at high velocity than at the photosphere [$X$(Ca)=0.02].
964: In the context of CSM interaction, \ie a density jump by a factor of 4 
965: and $X$(H)=0.30 ($M$(H)$\sim 5 \times 10^{-3} \Msun$) at $v \gsim 21,000$ \kms, 
966: $X$(Ca)=0.001 is enough for the HVF.
967: With the DD density structure a smaller fraction of Ca is required, 
968: as expected [$X$(Ca)=0.001 (0.0001) without (with) CSM interaction].
969: 
970: 
971: \subsection{SN 2003du (Day $-11.0$, LVG)}
972: 
973: \begin{figure}
974: \begin{center}
975: %\plotone{f07.eps}
976: \includegraphics[scale=0.85]{f07.eps}
977: \figcaption{
978: Observed spectrum of the LVG SN 2003du at day $-11$ (black solid) and a model 
979: with W7-CSM density/abundance structure (green dashed).
980: As in Fig. \ref{fig:02bo}, the left inset shows \ion{Ca}{ii} H\&K, the right 
981: inset \ion{Ca}{ii} IR. The W7-CSM spectrum is compared to W7 (thin cyan) and 
982: DD (dotted red) models, both with homogeneous abundance and without 
983: CSM interaction.
984: \label{fig:03du}}
985: \end{center}
986: \end{figure}
987: 
988: 
989: \subsubsection{General Properties}
990: 
991: The LVG SN 2003du was studied by Stanishev et al. (2007).
992: A synthetic spectrum with $\log L$ (\ergs) $=42.71$ and $\vph = 10,000$ \kms\ 
993: gives the best agreement with the observed spectrum (Fig. \ref{fig:03du}).
994: These parameters lead to $T_{\rm ph}=18,600$ K, which is close to that of 
995: other LVG SNe.
996: 
997: 
998: \subsubsection{Iron-group Elements}
999: 
1000: The \ion{Fe}{iii} feature near 4300 \AA\ is weaker than in SNe 2002el and 
1001: 2003cg, but the temperature is not significantly different, suggesting a 
1002: smaller Fe abundance at the photosphere: $X$(Fe) = 0.003 (determined as in 
1003: \S \ref{sec:02boFe}) is enough to obtain this feature.
1004: We estimate the Ni mass fraction as $X$(\Nifs)$_0$=0.005 from the feature at 
1005: 3900 \AA.
1006: Thus, almost all the Fe contributing to the spectrum is likely to be stable.
1007: 
1008: 
1009: \subsubsection{Silicon and Carbon}
1010: 
1011: The \ion{Si}{ii} $\lambda$6355 line is relatively weak and can be 
1012: reproduced with a Si mass fraction of $X$(Si)=0.3.
1013: In spite of the low velocity of the absorption minimum, the blue wing of the 
1014: absorption extends to $v \sim 20,000$ \kms.
1015: Although this could be a sign of high-velocity material as in SNe 2002bo, 
1016: 2002dj and 2001el, the line profile is different from other SNe, and the 
1017: high-velocity absorption ($v \sim 15,000 - 20,000$ \kms) is relatively weak
1018: (Mazzali et al. 2005b).
1019: 
1020: The emission peak of the line is suppressed, as seems common among LVG SNe.
1021: Including carbon with $X$(C) $\sim$ 0.003 improves the fit. If the 
1022: identification is correct, the mass of C is estimated to be M(C)$\sim0.001\Msun$
1023: assuming the W7 density structure and a homogeneous C mass fraction
1024: above the photosphere.
1025: 
1026: 
1027: \subsubsection{Calcium}
1028: 
1029: The \ion{Ca}{ii} IR triplet at day $-5$ was studied by Gerardy et al. (2004).
1030: The spectrum at day $-11$ also shows a conspicuous HVF (Fig. \ref{fig:03du}), 
1031: with profile and strength similar to those of SN 2001el. A very large 
1032: high-velocity Ca mass fraction [$X$(Ca)$>$0.7] is required if W7 is used 
1033: (thin cyan lines in the insets of Fig. \ref{fig:03du} show the profile of a 
1034: homogeneous abundance model with $X$(Ca)=0.016).
1035: For the CSM interaction scenario, we need $X$(Ca)=0.11 at high velocies 
1036: ($v \gsim 22,000$ \kms), where the density is increased by a factor of 4, 
1037: adding 0.012 $\Msun$, and $X$(H) is assumed to be 0.30 
1038: ($M$(H)$\sim 5 \times 10^{-3} \Msun$, dashed green line in 
1039: insets of Fig. \ref{fig:03du}).
1040: Even if the DD density structure is adopted, a high mass fraction of Ca at high 
1041: velocity is required [$X$(Ca)=0.10]. This is reduced to $X$(Ca)=0.015 when CSM 
1042: interaction is assumed(see Table \ref{tab:abun}).
1043: 
1044: \section{DISCUSSION}
1045: \label{sec:discussion}
1046: 
1047: \begin{figure}
1048: \begin{center}
1049: %\plotone{f08.eps}
1050: \includegraphics[scale=0.85]{f08.eps}
1051: \figcaption{
1052: Observed \ion{Si}{ii} $\lambda$6355 features of 3 HVG SNe Ia (top, blue) 
1053: and 3 LVG SNe Ia (bottom, green).
1054: Velocities and shapes of the blue wing differ among the objects. 
1055: They range from $15,000$ \kms\ to $25,000$ \kms\ irrespective of the SN group 
1056: (HVG, \S \ref{sec:HVG} or LVG, \S \ref{sec:LVG}).
1057: The black dashed lines correspond to Doppler velocities of $v = 10,000$ and 
1058: $13,000$ \kms, which are the typical photospheric velocities of LVG and HVG, 
1059: respectively.
1060: \label{fig:si}}
1061: \end{center}
1062: \end{figure}
1063: 
1064: \begin{figure*}
1065: \begin{center}
1066: %\plotone{f09.eps}
1067: \includegraphics[scale=0.95]{f09.eps}
1068: \figcaption{
1069: Radial distribution of (a) radiation temperature (K),
1070: (b) number density of electrons (${\rm cm^{-3}}$) 
1071: and (c)-(j) ionization fraction of each element.
1072: Blue lines show the values for a HVG SN (2002bo) 
1073: and green lines for a LVG SN (2001el).
1074: The two dominant ionization states are shown inside each panel; solid lines 
1075: represent the lower state (\eg \ion{C}{i}, \ion{O}{i}, \ion{Mg}{ii}).
1076: In panel (c), no blue lines appear, because C is not included in the models of 
1077: HVG SNe.
1078: \label{fig:ion}}
1079: \end{center}
1080: \end{figure*}
1081: 
1082: We have modelled the pre-maximum spectra of 6 different SNe Ia 
1083: (see Table \ref{tab:prop}).
1084: Three of these (2002bo, 2002dj and 2002er) are classified as High Velocity 
1085: Gradient (HVG), and are characterized by a large Si line velocity that evolves 
1086: rapidly after maximum brightness. 
1087: The other three SNe (2001el, 2003cg, and 2003du), classified as Low Velocity 
1088: Gradient (LVG), have a smaller Si line velocity that evolves slowly after 
1089: maximum.
1090: The parameters of our models are shown in Tables \ref{tab:param} (luminosity 
1091: and photospheric velocity) and \ref{tab:abun} (abundances).
1092: 
1093: In the following, we summarize the properties of HVG (\S \ref{sec:HVG}) 
1094: and LVG (\S \ref{sec:LVG}) SNe.
1095: Their different appearance stems basically 
1096: from a different photospheric velocity.
1097: The presence of iron-peak elements in the outer layers is revisited
1098: and its implications are discussed in \S \ref{sec:Fe}.
1099: Then, the origin of the difference between HVG and LVG is discussed in 
1100: \S \ref{sec:HVGLVG}.
1101: The variety of High Velocity Features (HVFs) is very large independently of
1102: the classification of HVG or LVG.
1103: The causes for HVFs are also discussed in \S \ref{sec:HVF}.
1104: Finally, possible influences to the LC width-luminosity relation are
1105: discussed (\S \ref{sec:WLR}).
1106: 
1107: 
1108: \subsection{The Properties of HVG SNe}
1109: \label{sec:HVG}
1110: 
1111: 
1112: 
1113: HVG SNe 2002bo and 2002dj  have the highest line velocities and the lowest
1114: photospheric temperatures of the 6 SNe.  The higher temperature for SN 2002er
1115: results from the high $L$ and the low $\vph$. SN 2002er seems to be
1116: intrinsically more luminous than SN 2002bo at a similar epoch (Table 2),
1117: although it has a larger decline rate.  This is also supported by the analysis
1118: of nebular spectra (Mazzali et al. 2007).  The velocity of SN 2002er is
1119: intrinsically lower than that of other HVG SNe (B05; Kotak et al. 2005). 
1120: Although a SN may be classified as HVG, its $T_{eff}$ can be as high as that of
1121: LVG SNe if the SN has both a high enough $L$ and low enough $\vph$. This
1122: suggests continuity between two groups (Branch et al. 2006, 2007).
1123: 
1124: The \ion{Si}{ii} $\lambda$6355 lines of SNe 2002bo and 2002dj have 
1125: well-defined P-Cygni profiles with blue wings reaching $v \sim 23,000$ \kms\
1126: (blue lines in Fig. \ref{fig:si}). In the spectra of these SNe, \ion{C}{ii} is
1127: not seen below at least $v = 20,000$ \kms.  This suggests that the abundance
1128: distribution of HVG SNe resembles that of 1D delayed detonation models
1129: (Khokhlov et al. 1991a; H\"oflich \& Khokhlov 1996; Iwamoto et al. 1999) rather
1130: than that of the 1D deflagration model W7 (Nomoto et al. 1984).
1131: 
1132: 
1133: The computed physical conditions in the ejecta of a representative HVG SN,
1134: 2002bo, and those of a typical LVG SN of similar $\Delta m_{15}$, and thus 
1135: presumably similar luminosity, SN 2001el, are shown in Figure \ref{fig:ion}. 
1136: In the ionization plot, solid lines represent the lower of two states marked in
1137: each panel. The temperature structure, the number density of electrons, and the
1138: ionization state of SN 2002bo and SN 2001el computed under the assumptions
1139: described in \S \ref{sec:method} are rather similar at all layers with $v >
1140: 12,900$\,\kms.  Consequently, the ionization fractions are also similar.
1141: 
1142: B05 show the evolution of the parameter $\cal R$(Si), the ratio of the 
1143: \ion{Si}{ii} lines at 5800 and 6000 \AA.
1144: $\cal R$(Si) is thought to be a temperature indicator (Nugent et al. 1995).
1145: At pre-maximum epochs ($\lsim -10$ days), $\cal R$(Si) is larger in HVG
1146: than in LVG SNe (see Fig. 2 in B05). 
1147: As we have shown, a typical HVG SN has a lower photospheric temperature before 
1148: maximum.
1149: The fact that SNe 2002bo and 2001el have a similar $\Delta m_{15}$, and 
1150: therefore should have a similar luminosity, suggests that the lower temperature 
1151: of SN 2002bo at pre-maximum epochs reflects its higher photospheric velocity.
1152: Therefore, $\cal R$(Si) is not a good indicator of SN luminosity at  
1153: pre-maximum phases, when the dispersion of the photospheric velocity among 
1154: different SNe is still very large (B05).
1155: On the other hand, since the diversity in photospheric velocity is small at 
1156: epochs near maximum, both $\cal R$(Si) 
1157: and the temperature more closely reflects the intrinsic luminosity.
1158: 
1159: 
1160: \subsection{The Properties of LVG SNe}
1161: \label{sec:LVG}
1162: 
1163: The three LVG SNe have similar photospheric velocities and temperatures.
1164: The temperature structure, electron density, and ionization state of 
1165: SN 2001el are shown in Figure \ref{fig:ion} (green lines).
1166: The run of all physical quantities is similar to HVG SN 2002bo at comparable 
1167: velocities, but the values at the photosphere are significantly different 
1168: because $\vph$ is lower for LVG.
1169: In particular, important ions such as \ion{Si}{ii}, \ion{S}{ii}, \ion{Ca}{ii}, 
1170: and \ion{Fe}{ii} are a factor of $> 10$ less abundant near the photosphere of 
1171: LVG SNe (at $v \sim 10,000$ \kms) then near the photosphere of HVG SNe 
1172: (at $v \sim 13,000$ \kms).
1173: 
1174: The \ion{C}{ii} $\lambda$6578 line is likely to be present in the LVG SNe
1175: 2001el and 2003du, and possibly also in SN 2003cg near the photosphere 
1176: ($v \sim 10,000$ \kms). The mass fraction of C required to suppress the 
1177: \ion{Si}{ii} $\lambda$6355 emission peak is however only 0.002-0.003.
1178: The mass of C in SNe 2001el and 2003du is only $\sim 0.001 \Msun$ if we 
1179: assume a homogeneous, spherically symmetric distribution above the photosphere.
1180: Even if we consider that we cannot estimate the C abundance
1181: at $v \gsim 18,000$ \kms\ because of the blend with the \ion{Si}{ii} line, 
1182: and assume $X$(C)=0.5 at $v > 18,000$ \kms, 
1183: the mass of C is at most $\sim 0.01 \Msun$. 
1184: This is still less than predicted by the W7 model ($\sim 0.05 \Msun$).
1185: 
1186: Although \ion{C}{i} is used to estimate C abundance in Marion et al. (2006),
1187: \ion{C}{i} is not the dominant ionization state as shown in Figure 
1188: \ref{fig:ion} (solid line), and the exact amount of \ion{C}{i} is very
1189: sensitive to the temperature structure. 
1190: In our calculations, \ion{C}{i} lines never become visible even if \ion{C}{ii}
1191: makes a deep absorption, suggesting that \ion{C}{ii} is a better indicator of
1192: carbon abundance in this temperature range.
1193: 
1194: Alternatively, the emission peak of \ion{Si}{ii} $\lambda$6355 may be
1195: suppressed if Si is detached from the photosphere (see Branch et al. 2002 for 
1196: the case of H$\alpha$). This is however not the case for the objects presented 
1197: here because the velocity of \ion{Si}{ii} decreases to $v < 9,000$ \kms\ at 
1198: subsequent epochs and Si is abundant at the corresponding velocities. 
1199: In addition to this, the fact that the emission profile of the \ion{Si}{ii} 
1200: line is well-peaked at maximum light even in LVG suggests that the suppression 
1201: is related to the conditions of the outer layers.
1202: 
1203: 
1204: \subsection{Fe-group elements in the outer ejecta}
1205: \label{sec:Fe}
1206: 
1207: Both \ion{Fe}{ii} and \ion{Fe}{iii} lines are clearly present in pre-maximum
1208: spectra.  We showed that about a few percent of Fe in mass fraction, including
1209: both stable Fe and decay products of \Nifs, are required to match the observed
1210: lines (\S \ref{sec:02boFe}). Lines of both ions (\ion{Fe}{ii} and
1211: \ion{Fe}{iii}) are correctly reproduced, which suggests that the computation of
1212: the ionization state of Fe is reliable. The fact that Fe is not highly abundant
1213: indicates that the photosphere resides outside the Fe-rich layers.
1214: 
1215: In order to distinguish stable Fe (\eg $^{54}$Fe) from the decay product of
1216: \Nifs\ (\ie $^{56}$Fe), the abundances of Ni and Co were quantified from the
1217: spectra (\S \ref{sec:02boFe}). In 4 of our SNe, we are able to estimate the
1218: \Nifs\ abundance near the photosphere, while only upper limits could be
1219: obtained in the remaining 2 cases. If \Nifs\ accounted for all the present Fe
1220: abundance at such an early epoch (when only 3\% of \Nifs\ has decayed to
1221: \Fefs), the ejecta would be dominated by \Nifs. This is, however, clearly
1222: inconsistent with the observed spectra and the parameters obtained from the
1223: modeling. Therefore, we conclude that not only \Nifs, but also stable Fe must
1224: exist at the photosphere. The respective abundances obtained are always larger
1225: than solar suggesting that the origin is not the progenitor but nucleosynthesis
1226: during the explosion.
1227: 
1228: The metal abundances in the outer layers suggest that the abundances are  
1229: not sharply stratified with velocity. 
1230: Even in 1D, the region where \Nifs\ is synthesized extends to Si-rich layers 
1231: because the temperature changes smoothly during the explosion.
1232: $^{54}$Fe can also be synthesized in Si-rich layers (\eg Iwamoto et al. 1999).
1233: Another explanation for Fe-group elements in outer layers may be large 
1234: scale mixing, as seen in recent three-dimensional deflagration models 
1235: (\eg Gamezo et al. 2003; R\"opke \& Hillebrandt 2005; R\"opke et al. 2006). 
1236: 
1237: The presence of Fe-group elements in the outer layers and the diversity of
1238: their abundances and a low abundance of unburned elements may hinder the use of 
1239: pre-maximum spectra as indicators of the progenitor metallicity (Lentz et al. 
1240: 2000).
1241: In fact, larger diversity than predicted by model computations is seen in 
1242: UV spectra (Ellis et al. 2007).
1243: This may reflect diversity in the abundances of Fe-group elements in the outer 
1244: layers, or in the photospheric temperature (\S \ref{sec:HVG} and \ref{sec:LVG}).
1245: 
1246: 
1247: \subsection{The Origin of the Difference between HVG and LVG SNe}
1248: \label{sec:HVGLVG}
1249: 
1250: The difference between HVG and LVG SNe can be studied comparing the properties
1251: of SNe with similar decline rates ($\Delta m_{15}$). For an optimal comparison
1252: at a similar epoch, we again take SNe 2002bo (HVG; $-8.0$ days, $\Delta m_{15}
1253: = 1.17$) and 2001el (LVG; $-9$ days, $\Delta m_{15} = 1.15$) as examples. 
1254: 
1255: The marginal detection of carbon in the emission peak of 
1256: \ion{Si}{ii} $\lambda$6355 in LVG SNe suggests that
1257: the burning is less powerful in LVG SNe than in HVG SNe. 
1258: It is also interesting that other SNe where C was detected (SN 1998aq, 
1259: Branch et al. 2003; and possibly SN 1994D, Branch et al. 2005) are also LVG
1260: \footnote{Since the number of LVG seems to be larger than that of HVG 
1261: (B05; Mazzali et al. 2007), this could be a statistical effect.}.
1262: In addition, we find that HVG SNe do not have carbon at least 
1263: up to $v\sim 20,000$ \kms.
1264: 
1265: If more burning occurs in HVG than in LVG SNe, the former may have more 
1266: kinetic energy.
1267: This would tend to make the photospheric velocity of HVG higher. The kinetic 
1268: energy difference between HVG and LVG can estimated by a simple analysis.
1269: We assume that the outermost layer are fully burned to Si in HVG while half 
1270: of the material at $v > 18,000$ \kms\ is oxygen and carbon in LVG.
1271: Since the mass above $v=18,000$ \kms\ is $\sim 0.02 \Msun$ assuming the W7 
1272: density structure, the difference in nuclear energy release is
1273: $\sim 0.02 \times 10^{51}$ erg, which is less than $\sim 2\%$ of the typical 
1274: kinetic energy of SNe Ia models.
1275: The expected difference in photospheric velocity is however only $\lsim 1$\%.
1276: Therefore, it seems unlikely that the difference between HVG and LVG can 
1277: be accounted for by a different kinetic energy alone. 
1278: Yet, such a difference in kinetic energy cannot be ruled out.
1279: 
1280: Alternatively, HVG SNe may have more massive ejecta than LVG SNe.  
1281: The limiting mass of a spinning WD can exceed the Chandrasekhar mass of the 
1282: static WD, $\sim 1.38 \Msun$ (Uenishi, Nomoto \& Hachisu 2003; 
1283: Yoon \& Langer 2004, 2005; Dom\'inguez et al. 2006).  
1284: Since the density structure of a rotating WD is flatter than that of a static 
1285: one, the photospheric velocity may be larger before maximum but then evolve 
1286: more rapidly.  Unless the WD mass is extremely large -- such as the 2$\Msun$ 
1287: suggested for SN 2003fg by Howell et al. (2006), the expansion kinetic 
1288: energy would not be significantly reduced by a higher binding energy.
1289: However, there is no evidence that HVG SNe tend to synthesize more \Nifs, 
1290: which does not support this interpretation.
1291: 
1292: Even though the mass and the kinetic energy of the ejecta are similar in HVG 
1293: and LVG, a different Fe-group abundance in the outer region could make their 
1294: appearance different.
1295: Some amount of Fe-group elements is present in the outer layers of all 6
1296: SNe (see \S \ref{sec:Fe}). 
1297: Line velocities are expected to be higher if more Fe-group elements are present
1298: in the outer layers because of the larger opacity of these elements.
1299: 
1300: Our results show that the near-photospheric abundance of Fe-group elements is
1301: not smaller in HVG than in LVG, even though the photospheric velocity is higher
1302: in HVG by $\sim 2-3,000$ \kms. This might imply that HVG SNe have more Fe-group
1303: elements if the metal abundance is a decreasing function of radius (velocity).
1304: Therefore, the abundances of Fe-group elements in the outer layers can be a
1305: cause of spectral diversity, while differences in the kinetic energy or the
1306: progenitor mass are less likely. To confirm this hypothesis, the metal
1307: abundance in the spectra of LVG with $\vph \gsim 13,000$ \kms\ should be
1308: investigated with earlier spectra.
1309: 
1310: 
1311: \subsection{High-Velocity Features}
1312: \label{sec:HVF}
1313: 
1314: \begin{figure}
1315: \begin{center}
1316: %\plotone{f10.eps}
1317: \includegraphics[scale=0.85]{f10.eps}
1318: \figcaption{
1319: \ion{Ca}{ii} IR triplet profiles (a mean wavelength of 8567 \AA\ is assumed) 
1320: of 3 HVG SNe Ia (top, blue) and 3 LVG SNe Ia (bottom, green).
1321: In LVG SNe (SNe 2001el and 2003du in our sample and SN 2003kf in Mazzali et al. 
1322: 2005b), HVFs tend to be detached, which is likely the consequence of the high 
1323: photospheric temperature and the low photospheric velocity 
1324: (see \S \ref{sec:HVF}).
1325: On the other hand, the strength of the high-velocity absorption 
1326: seems to be random (Tanaka et al. 2006). 
1327: The dashed black, vertical lines correspond to Doppler velocities of
1328: $v = 10,000$ and $13,000$ \kms, which are the typical photospheric velocities
1329: of LVG and HVG, respectively. 
1330: \label{fig:ca}}
1331: \end{center}
1332: \end{figure}
1333: 
1334: 
1335: High-velocity features ($v \gsim 20,000$ \kms) are seen in both \ion{Ca}{ii}
1336: H\&K and in the IR triplet. Except for SN 2003cg, HVFs are not present in 
1337: synthetic spectra based on W7. They can be reproduced using the DD density
1338: structure and a relatively large Ca abundance. Alternatively, they can be also
1339: explained by an enhancement in density of the outer layer ($v > 20-22,000$
1340: \kms) of a factor of 4 from the W7 value assuming some H ($\sim 30$\% in mass
1341: fraction) in this region. Interaction with CSM may produce the latter
1342: situation; mixing of H into the ejecta increases the \ion{Ca}{ii} fraction by a
1343: factor of 5-10 as it favors recombination (Mazzali et al. 2005a).
1344: 
1345: The profiles of the \ion{Ca}{ii} IR triplet show variations (Fig.
1346: \ref{fig:ca}).  In HVG SNe (blue lines), the profile is round and the 
1347: high-velocity component tends to blend with the photospheric component. In LVG
1348: SNe, on the other hand, the profiles are less blended because of the weakness 
1349: of the photospheric component (green lines, see also SN 2003kf in Mazzali et
1350: al. 2005b). This is a natural consequence of the high $\Tph$ and the low 
1351: $\vph$ of LVG SNe.  The high temperature suppresses \ion{Ca}{ii} and
1352: \ion{Si}{ii} lines near the photosphere, and the low photospheric velocity
1353: makes the position of the photospheric component redder. The combination of
1354: these two effects results in a weaker and redder photospheric absorption, so
1355: that the high-velocity component becomes more detached in LVG SNe.
1356: 
1357: The strength of the high-velocity component also varies from SN to SN. As our
1358: models show, this is not a temperature effect, because the temperature in the
1359: outer region is similar in all cases (Fig. \ref{fig:ion}).  Although we could
1360: reproduce the strength of the HVF increasing the Ca abundance, the abundances
1361: that are required show a large spread (of a factor of 100; Table
1362: \ref{tab:abun}). This diversity may be a line-of-sight effect if HVFs are
1363: caused by aspherical structures, like a torus or clumps (Tanaka et al. 2006),
1364: as also suggested by the high polarization level of the feature (Wang et al.
1365: 2003; Kasen et al. 2003).
1366: 
1367: The very strong, broad blue wing of the \ion{Si}{ii} $\lambda$6355 absorption 
1368: in SNe 2002bo, 2002dj and 2001el is not perfectly explained by our synthetic 
1369: spectra (Figures \ref{fig:02bo}, \ref{fig:02dj} and \ref{fig:01el}). This  
1370: suggests the presence of HVFs in \ion{Si}{ii} as well as in \ion{Ca}{ii} lines,
1371: although the velocity of the blue edge of the \ion{Si}{ii} absorption is lower
1372: than that of the \ion{Ca}{ii} lines (Figures \ref{fig:si} and \ref{fig:ca}).
1373: On the other hand, SNe 2002er, 2003cg and 2003du lack this feature (Figures
1374: \ref{fig:02er}, \ref{fig:03cg}, and \ref{fig:03du}). 
1375: 
1376: 
1377: The origin of HVFs is debated (\eg Hatano et al. 1999; Wang et al. 2003; Thomas
1378: et al. 2004; Gerardy et al. 2004; Mazzali 2005a; Kasen \& Plewa 2005; Tanaka et
1379: al. 2006). The difficulty in using the DD model is that it still requires a high
1380: Ca abundance in the high-velocity layers ($v \gsim 23,000$ \kms; Table
1381: \ref{tab:abun}) despite the absence of a very strong Si absorption at similar 
1382: velocities. One problem with the CSM interaction scenario is that it  requires
1383: the accumulation  of a relatively large mass. The estimated mass in the
1384: high-velocity shell is $\sim 0.01 \Msun$. In order to accumulate this mass, a
1385: mass loss rate $\dot{M} \sim 10^{-4}\ \Msun\ {\rm yr}^{-1}$ is required for a
1386: wind velocity of 10 \kms. This mass loss rate is higher than that constrained
1387: from X-ray, optical and radio observations (\eg Cumming et al. 1996; Mattila et
1388: al. 2005; Immler et al. 2006;  Hughes et al. 2007; Panagia et al. 2006). A
1389: combination of these two scenarios  (\ie DD model and CSM interaction) may relax
1390: these requirements (Gerardy et al. 2004).
1391: 
1392: Alternatively, HVFs may reflect the pre-SN abundances of the progenitor WD.
1393: In the single-degenerate scenario, the WD accretes matter from a companion star
1394: (see Nomoto et al. 1994 for a review).
1395: At an accrretion rate suitable for stable H shell burning (Nomoto 1982),
1396: He shell flashes are rather weak (Taam 1980; Fujimoto \& Sugimoto 1982),
1397: and their products can accrete onto the WD.
1398: 
1399: The products of He shell flashes could be Ca-rich depending on conditions 
1400: such as the pressure at the burning shell (Hashimoto, Hanawa \& Sugimoto 1983).
1401: After many cycles of He flashes, when the mass of the WD reaches the 
1402: Chandrasekhar mass, the outermost layers of the WD should consist not only
1403: of C and O but also of heavier elements.
1404: If such matter resides in the outermost layers after the explosion,
1405: it could cause HVFs.
1406: To investigate this scenario further, the growth of the WD should be studied 
1407: in detail.
1408: 
1409: If \ion{Si}{ii} and \ion{Ca}{ii} HVFs are caused by the explosion itself or by 
1410: CSM interaction, they should correlate (see Fig. 12 in Tanaka et al. 2006).
1411: In fact, this seems to be the case for SNe 2002bo, 2002dj, 2001el, and 2003cg, 
1412: but the correlation is not clear for SNe 2002er and 2003du
1413: \footnote{Si and Ca HVF may correlate in an earlier spectrum of SN 2003du. 
1414: See Stanishev et al. (2007).}
1415: (Figs. \ref{fig:si} and \ref{fig:ca}).
1416: This may be useful to discriminate the origin of the HVFs.
1417: 
1418: 
1419: \subsection{Implications for the Width-Luminosity Relation}
1420: \label{sec:WLR}
1421: 
1422: The relation between LC width and luminosity in SNe Ia indicates that the 
1423: properties of SNe Ia are determined by a single dominant parameter.
1424: However, dispersion is present in the relation, and it may due to 
1425: uncertainties in distance and reddening or to intrinsic properties of 
1426: SNe Ia (Mazzali \& Podsiadlowski 2006).
1427: Understanding the origins of the dispersion is important for 
1428: a more precise cosmological use of SNe Ia.
1429: 
1430: The earliest spectra of SNe Ia are clearly not determined by a single
1431: parameter, \eg SNe with a similar maximum luminosity have different line
1432: velocities. Therefore, the variation in the early phase spectra could, to some
1433: extent, reveal an intrinsic dispersion in the width-luminosity relation. Since
1434: the variation of properties is larger in HVG than in LVG SNe (B05; Mazzali et
1435: al. 2007), the HVG group at least could be related to the dispersion in the
1436: width-luminosity relation.
1437: 
1438: A first parameter to look at in this context is the kinetic energy of the
1439: ejecta. HVG SNe may have higher kinetic energy than LVG by at most $\sim 2$\% 
1440: (\S \ref{sec:HVGLVG}). Since the light curve width ($\tau_{\rm LC}$) scales
1441: roughly as $\tau_{\rm LC} \propto \KE^{-1/4}$  (Arnett 1982), where $\KE$ is
1442: the kinetic energy of the ejecta, the LC width of HVG SNe may become narrower
1443: than that of LVG SNe. This effect should not be significant 
1444: given the dependence of $\tau_{\rm LC}$. 
1445: However, an earlier LC rise, accompanied by a brighter
1446: luminosity at maximum even when the ejecta contain the same amount of \Nifs, is
1447: a change orthogonal to the observed width-luminsity relation. 
1448: Thus, differences in $\KE$ could cause an intrinsic dispersion 
1449: in the width-luminosity relation.
1450: 
1451: The \Nifs\ distribution could also affect the LC rise time.  If HVG SNe have an
1452: extended \Nifs\ distribution as suggested in \S \ref{sec:HVGLVG}, their LC may
1453: peak earlier than in the case where \Nifs\ is confined in the innermost layers
1454: because the diffusion time is shorter for the outer \Nifs. An early LC rise
1455: would make the peak luminosity brighter even for the same \Nifs\ mass.
1456: Consequently, SNe with similar $\Delta m_{15}$ may have different peak
1457: luminosities in certain conditions of mixing (\eg Mazzali \& Podsiadlowski
1458: 2006, Woosley et al. 2007).
1459: 
1460: 
1461: \section{CONCLUSIONS}
1462: \label{sec:conclusions}
1463: 
1464: The outermost ejecta of Type Ia SNe are studied 
1465: by modeling very early spectra.
1466: Of the 6 SNe we studied, 2 (SNe 2002bo and 2002dj) are classified as HVG, 
1467: being characterized by a high photospheric velocity and a low photospheric 
1468: temperature. All three LVG SNe (SNe 2001el, 2003cg and 2003du) have uniform
1469: properties and are characterized by a low photospheric velocity and a high
1470: photospheric temperature. The properties of SN 2002er, which is classified as
1471: HVG, are close to those of LVG, suggesting that there is continuity between the
1472: groups. This is consistent with suggestions by Branch et al. (2006, 2007)
1473: 
1474: HVG SNe have a \ion{Si}{ii} $\lambda$6355 line with a pronounced emission peak
1475: and a broad blue absorption, suggesting that Si is present at $v > 20,000$
1476: \kms. The abundance distribution in HVG SNe is similar to that of a delayed 
1477: detonation model, \ie the burning front reaches the outermost layers. On the
1478: other hand, the \ion{Si}{ii} $\lambda$6355 emission profile of LVG SNe tends to
1479: be suppressed, possibly because of the presence of \ion{C}{ii} $\lambda$6578.
1480: This suggests that the burning front in LVG SNe is weaker than in HVG SNe.
1481: However, the mass fraction of carbon at the photosphere of LVG SNe 
1482: is only $X$(C) $\lsim 0.01$, and
1483: the estimated C mass is less than $\sim 0.01 \Msun$, which is less 
1484: than predicted by W7 ($\sim 0.05 \Msun$).
1485: 
1486: The difference in the photospheric velocity explains the different appearance 
1487: of HVG and LVG SNe at pre-maximum phases through the temperature difference.
1488: At the highest velocities ($v \gsim 20,000$ \kms), however, the variety of the 
1489: \ion{Ca}{ii} and \ion{Si}{ii} features is not explained by this scenario.
1490: This may require additional factors such as asphericity or diversity 
1491: in the element abundance of the progenitor WD.
1492: 
1493: Both stable Fe and \Nifs\ are detected at the photosphere ($v = 9-14,000$ \kms)
1494: in all 6 SNe, suggesting that Fe-group elements are always present in the 
1495: outer layers.
1496: The differece in their abundances may be an important reason for the 
1497: diversity among SN Ia spectra.
1498: On the other hand, the difference in kinetic energy between HVG and LVG SNe 
1499: is too small to be the origin of the spectral diversity.
1500: 
1501: The diversity of SNe Ia seen in early phase spectra could cause the intrinsic
1502: dispersion in the LC width-luminosity relation, through the difference in
1503: kinetic energy and the amount of iron-group elements in the outer layers.
1504: 
1505: 
1506: \acknowledgments
1507: M.T. is supported by the JSPS (Japan Society for the Promotion of Science) 
1508: Research Fellowship for Young Scientists.
1509: G.P. acknowledges support by the Proyecto FONDECYT 3070034.
1510: This research was supported in part by the Grant-in-Aid for 
1511: Scientific Research (18104003, 18540231)
1512: and the 21st Century COE Program (QUEST) from the JSPS and MEXT of Japan.
1513: 
1514: 
1515: \begin{thebibliography}{}
1516: 
1517: \bibitem{abb85} Abbott, D.C., \& Lucy, L.B. 1985, ApJ, 288, 679
1518: 
1519: \bibitem{arn82} Arnett, W.D. 1982, ApJ, 253, 785
1520: 
1521: \bibitem{ast06} Astier, P., et al. 2006, A\&A, 447, 31
1522: 
1523: \bibitem{Bar06} Baron, E., Bongard, S., Branch, D., Hauschildt, P. H. 2006, ApJ, 645, 480
1524: 
1525: \bibitem{ben04} Benetti, S., et al. 2004, MNRAS, 348, 261 
1526: 
1527: \bibitem{ben05} Benetti, S., et al. 2005, ApJ, 623, 1011 (B05)
1528: 
1529: \bibitem{bon06} Bongard, S., Baron, E., Smadja, G., Branch, D., \& Hauschildt, P.H. 2006, ApJ, 647, 513
1530: 
1531: \bibitem{bra02} Branch, D., et al. 2002, ApJ, 566, 1005
1532: 
1533: \bibitem{bra03} Branch, D., et al. 2003, AJ, 126, 1489  
1534: 
1535: \bibitem{bra05} Branch, D., Baron, E., Hall, N., Melakayil, M., \& Parrent, J. 2005, PASP, 117, 545
1536: 
1537: \bibitem{bra06} Branch, D., et al. 2006, PASP, 118, 560 
1538: 
1539: \bibitem{bra07} Branch, D., et al. 2007, PASP, 119, 709
1540: 
1541: \bibitem{cum96} Cumming, R. J., Lundqvist, P., Smith, L. J., Pettini, M., King, D. L. 1996, MNRAS, 283, 1335
1542: 
1543: \bibitem{dom06} Dom\'inguez, I.,  Piersanti, L.,  Bravo, E.,  Tornambe, A., Straniero, O., \& Gagliardi, S. 2006, ApJ, 644, 21
1544: 
1545: \bibitem{eli06} Elias-Rosa, N., et al. 2006, MNRAS, 369, 1880
1546: 
1547: \bibitem{ell07} Ellis, R.S., et al. 2007, ApJ, in press (arXiv:0710.3896)
1548: 
1549: \bibitem{fis97} Fisher A., Branch D., Nugent P., Baron E., 1997, ApJ, 481, L89
1550: 
1551: \bibitem{fuj82} Fujimoto, M. Y., \& Sugimoto, D. 1982, 257, 291
1552: 
1553: \bibitem{gam03} Gamezo, V.N., Khokhlov, A., Oran, E.S., Chtchelkanova, A.Y. \& Rosenberg, R.O. 2003, Science, 299, 77
1554: 
1555: \bibitem{gam06} Gamezo, V.N., Khokhlov, A., \& Oran, E.S., 2005, ApJ, 623, 337
1556: 
1557: \bibitem{gar05} Garavini, G., et al. 2005, AJ, 130, 2278
1558: 
1559: \bibitem{gar07} Garavini, G., et al. 2007, A\&A, 471, 527
1560: 
1561: \bibitem{ger04} Gerardy, C. L., et al.  2004, ApJ, 607, 391
1562: 
1563: \bibitem{hat99} Hatano, K., Branch, D., Fisher, A., Baron, E. \& Filippenko, A.V. 1999, ApJ, 525, 881
1564: 
1565: \bibitem{hat00} Hatano, K., Branch, D., Lentz, E.J., Baron, E., Filippenko, A.V. \& Garnavich, P.M. 2000, ApJ, 543, L49 
1566: 
1567: \bibitem{hac06} Hachinger, S., Mazzali, P.A., \& Benetti, S. 2006, MNRAS, 370, 299
1568: 
1569: \bibitem{has83} Hashimoto, M.-A., Hanawa, T., \& Sugimoto, D. 1983, PASJ, 35, 1
1570: 
1571: \bibitem{hic07} Hicken, M., Garnavich, P.M., Prieto, J.L., Blondin, S., DePoy, D.L., Kirshner, R.P., \& Parrent, J. 2007, ApJ, 669, L17 
1572: 
1573: \bibitem{hof96} H\"oflich P., \& Khokhlov, A. 1996, ApJ, 457, 500
1574: 
1575: \bibitem{how06} Howell, D.A., et al. 2006, Nature, 443, 308
1576: 
1577: \bibitem{hug07} Hughes, J. P., Chugai, N., Chevalier, R., Lundqvist, P., \& Schlegel, E. 2007, ApJ, in press (arXiv:0710.3190)
1578: 
1579: \bibitem{imm06} Immler, S., et al. 2006, ApJ, 648, L119
1580: 
1581: \bibitem{iwa99} Iwamoto, K., Brachwitz, F., Nomoto, K., Kishimoto, N., Umeda, H., Hix, W.P., \& Thielemann, F-K. 1999, ApJS, 125, 439
1582: 
1583: \bibitem{kas03} Kasen, D., et al. 2003, ApJ, 593, 788
1584: 
1585: \bibitem{kas05} Kasen, D., \& Plewa, T. 2005, ApJ, 622, L41
1586: 
1587: \bibitem{kho91a} Khokhlov, A.M. 1991a, A\&A, 245, L25	
1588: 
1589: \bibitem{kho91b} Khokhlov, A.M. 1991b, A\&A, 245, 114	
1590: 
1591: \bibitem{kle78} Klein, R. I., \& Castor, J. I. 1978, ApJ, 220, 902
1592: 
1593: \bibitem{kno03} Knop, R.A., et al. 2003, ApJ, 598, 102
1594: 
1595: \bibitem{kot05} Kotak, R., et al. 2005, A\& A, 436, 1021
1596: 
1597: \bibitem{lei91} Leibundgut, B., Kirshner, R. P., Filippenko, A. V., Shields, J. C., Foltz, C. B., Phillips, M. M., \&  Sonneborn, G. 1991, ApJ, 371, L23
1598: 
1599: \bibitem{len00} Lentz, J.E., Baron, E., Branch, D., Hauschildt, P.H., \& Nugent, P. 2000, ApJ, 530, 966
1600: 
1601: \bibitem{luc99} Lucy, L.B. 1999, A\&A, 345, 211
1602: 
1603: \bibitem{mar06} Marion, G.H., H\"oflich, P., Wheeler, J.C., Robinson, E.L., Gerardy, C.L., \& Vacca, W.D. 2006, ApJ, 645, 1392
1604: 
1605: \bibitem{mat05} Mattila, S., Lundqvist, P., Sollerman, J., Kozma, C., Baron, E., Fransson, C., Leibundgut, B., \& Nomoto, K. 2005, A\&A, 443, 649
1606: 
1607: \bibitem{ml93} Mazzali, P.A. \& Lucy, L.B. 1993, A\&A, 279, 447
1608: 
1609: \bibitem{maz93} Mazzali, P.A., Lucy, L.B., Danziger, I.J., Gouiffes, C., Cappellaro, E., \& Turatto, M. 1993, A\&A, 269, 423
1610: 
1611: \bibitem{maz98}	Mazzali, P. A., Cappellaro, E., Danziger, I. J., Turatto, M., \& Benetti, S. 1998, ApJ, 499, L49
1612: 
1613: \bibitem{maz00} Mazzali, P.A. 2000, A\&A, 363, 705
1614: 
1615: \bibitem{maz01} Mazzali, P.A. 2001, MNRAS, 321, 341
1616: 
1617: \bibitem{maz05a} Mazzali, P.A., Benetti, S.,  Stehle, M., Branch, D., Deng, J., Maeda, K., Nomoto, K. \& Hamuy, M.  2005a, MNRAS, 357, 200
1618: 
1619: \bibitem{maz05b} Mazzali, P.A., et al. 2005b, ApJ, 623, L37
1620: 
1621: \bibitem{maz06} Mazzali, P.A., \& Podsiadlowski, P. 2006, MNRAS, 369, L19
1622: 
1623: \bibitem{maz07} Mazzali, P.A., R\"opke, F.K., Benetti, S., \& Hillebrandt, W. 2007, Science, 315, 825
1624: 
1625: \bibitem{nom76} Nomoto, K., Sugimoto, D., \& Neo, S. 1976, Ap\&SS, 39, L37
1626: 
1627: \bibitem{nom82} Nomoto, K. 1982, ApJ, 253, 798
1628: 
1629: \bibitem{nom84} Nomoto, K., Thielemann, F.-K., \& Yokoi, K. 1984, ApJ, 286, 644
1630: 
1631: \bibitem{nom94} Nomoto, K., Yamaoka, H., Shigeyama, T., Kumagai, S., \& Tsujimoto, T. 1994, Bludman S. A., Mochkovitch R., Zinn-Justin, eds., Les Houches Session LIV, Supernovae, Elsevier, Amsterdam, 199
1632: 
1633: \bibitem{nug95} Nugent, P., Phillips, M., Baron, E., Branch, D., \& Hauschildt, P. 1995, ApJ, 455, L147
1634: 
1635: \bibitem{pan06} Panagia, N., Van Dyk, S. D., Weiler, K. W., Sramek, R. A., Stockdale, C. J., \& Murata, K. P. 2006, ApJ, 646, 369 
1636: 
1637: \bibitem{per99} Perlmutter, S., et al. 1999, ApJ, 517, 565
1638: 
1639: \bibitem{phi93} Phillips, M.M. 1993, ApJ, 413, L105 
1640: 
1641: \bibitem{phi99} Phillips, M.M., Lira, P., Suntzeff, N. B., Schommer, R. A., Hamuy, M., \&  Maza, J. 1999, AJ, 118, 2688 
1642: 
1643: \bibitem{pig04} Pignata, G., et al. 2004, MNRAS, 355, 178 
1644: 
1645: \bibitem{pig05} Pignata, G., et al. 2005, ASP Conf. Ser. 342, 1604-2004: Supernovae as Cosmological Lighthouses, ed. M. Turatto, S. Benetti, L. Zampieri, \& W. Shea (San Francisco: ASP), 266
1646: 
1647: \bibitem{pin00} Pinto, P.A., \& Eastman, R.G. 2000, ApJ, 530, 757
1648: 
1649: \bibitem{ple04} Plewa, T., Calder, A.C., \& Lamb, D. Q.  2004, ApJ, 612, L37
1650: 
1651: \bibitem{qui06}Quimby, R., H\"oflich, P., Kannappan S.J., Rykoff E., Rujopakarn W., Akerlof C.W., Gerardy C.L., \& Wheeler J.C. 2006, ApJ, 636, 400
1652: 
1653: \bibitem{rei02} Reinecke, M., Hillebrandt, W. \& Niemeyer, J. C. 2002, A\& A, 391, 1167
1654: 
1655: \bibitem{rie98} Riess, A. G., et al. 1998, AJ, 116, 1009
1656: 
1657: \bibitem{rie04} Riess, A. G., et al. 2004, ApJ, 607, 665
1658: 
1659: \bibitem{roe05} R\"opke, F., K. \&  Hillebrandt, W. 2005, A\& A, 431, 635
1660: 
1661: \bibitem{roe06}	R\"opke, F. K., Gieseler, M., Reinecke, M., Travaglio, C., \& Hillebrandt, W. 2006, A\&A, 453, 203
1662: 
1663: \bibitem{sta07} Stanishev, V. et al. 2007, A\& A, 469, 645
1664: 
1665: \bibitem{ste05} Stehle, M., Mazzali, P.A., Benetti, S., \& Hillebrandt, W. 2005, MNRAS, 360, 1231
1666: 
1667: \bibitem{taa80} Taam, R. E. 1980, ApJ, 237, 142
1668: 
1669: \bibitem{tan06} Tanaka, M., Mazzali, P.A., Maeda, K., \& Nomoto, K. 2006, ApJ, 645, 470
1670: 
1671: \bibitem{tho04} Thomas, R. C., Branch, D., Baron, E., Nomoto, K., Li, W., Filippenko, A. V. 2004, ApJ, 601, 1019
1672: 
1673: \bibitem{tho06} Thomas, R. C, et al. 2007, ApJ, 654, L53
1674: 
1675: \bibitem{tra04} Travaglio, C., Hillebrandt, W., Reinecke, M., \& Thielemann, F.-K. 2004, A\& A, 425, 1029
1676: 
1677: \bibitem{uen04} Uenishi, T., Nomoto, K., \& Hachisu, I. 2003, ApJ, 595, 1094
1678: 
1679: \bibitem{wan03} Wang, L., et al., 2003, ApJ, 591, 1110
1680: 
1681: \bibitem{woo07} Wood-Vasey, W.M., et al. 2007, ApJ, 666, 694
1682: 
1683: \bibitem{woo06} Woosley, S. E., Kasen, D., Blinnikov, S., \& Sorokina, E. 2007, ApJ, 662, 487
1684: 
1685: \bibitem{yoo04} Yoon, S.C., \& Langer, N. 2004, A\&A, 419, 623
1686: 
1687: \bibitem{yoo05} Yoon, S.C., \& Langer, N. 2005, A\&A, 435, 967
1688: 
1689: \end{thebibliography}
1690: 
1691: 
1692: \end{document}
1693: