1: %\documentclass{aastex}
2: \documentclass[12pt,preprint]{aastex}
3:
4: %\documentclass[12pt,manuscript]{aastex}
5: %\usepackage{graphicx}
6:
7: \begin{document}
8: \def\la{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
9: \def\ga{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
10: \def\lam{$\lambda$}
11: \def\kms{km~s$^{-1}$}
12: \def\vphot{$v_{\rm phot}$}
13: \def\ang{~\AA}
14: \def\syn{SYNOW}
15: \def\dm15{{$\Delta$}$m_{15}$}
16: \def\rsi{$R$(Si~II)}
17: \def\v10{$V_{10}$(Si~II)}
18: \def\wsi{$W_lambda$(Si~II)}
19: \def\vdot{\.v(Si~II)}
20: \def\W575{$W(5750)$}
21: \def\W610{$W(6100)$}
22: \def\6100{the 6100~\AA\ absorption}
23: \def\tex{$T_{\rm exc}$}
24: \def\ve{$v_{\rm e}$}
25:
26: \title {Comparative Direct Analysis of Type~Ia Supernova Spectra.
27: III. Premaximum}
28:
29: \author {David Branch\altaffilmark{1}, M.~A. Troxel\altaffilmark{1},
30: David~J. Jeffery\altaffilmark{1}, Kazuhito Hatano\altaffilmark{2},
31: Miriam Musco\altaffilmark{3}, Jerod Parrent\altaffilmark{1},
32: E.~Baron\altaffilmark{1}, Leeann Chau
33: Dang\altaffilmark{1,}\altaffilmark{4}, D.~Casebeer\altaffilmark{1},
34: Nicholas Hall\altaffilmark{1}, \& Wesley Ketchum\altaffilmark{1}}
35:
36: \altaffiltext{1} {Homer L. Dodge Department of Physics and Astronomy,
37: University of Oklahoma, Norman,~OK 73019; e-mail: branch@nhn.ou.edu}
38:
39: \altaffiltext{2} {Division of Network Business, PionetSoft Corporation,
40: Shibuya-ku, Tokyo 151-0072, Japan}
41:
42: \altaffiltext{3} {Department of Astronomy, Indiana University,
43: Bloomington,~IN 47405}
44:
45: \altaffiltext{4} {Department of Astronomy, Whitman College, Walla
46: Walla,~WA 99362}
47:
48: \begin{abstract}
49:
50: A comparative study of spectra of 21 Type~Ia supernovae (SNe~Ia)
51: obtained about one week before maximum light, and 8 spectra obtained
52: 11 or more days before maximum, is presented. To a large extent the
53: premaximum spectra exhibit the defining characteristics of the four
54: groups defined in Paper~II (core-normal, broad--line, cool, and
55: shallow--silicon). Comparisons with SYNOW synthetic spectra show that
56: all strong features and most weak ones can be accounted for in a
57: plausible way. The issues of detached high--velocity features, the
58: possible ubiquity of carbon clumps, the maximum detectable ejecta
59: velocities, and the possibility of blueshifted emission--line peaks
60: are discussed.
61:
62: \end{abstract}
63:
64: \keywords{supernovae: general}
65:
66: \section{INTRODUCTION}
67:
68: This is the third in a series of papers on a comparative direct
69: analysis of the optical spectra of Type~Ia supernovae (SNe~Ia).
70: Paper~I (Branch et~al. 2005) was concerned with a time series of
71: spectra, from 12 days before to 115 days after the time of maximum
72: light, of the well--observed, spectroscopically normal, Type~Ia
73: SN~1994D. Paper~II (Branch et~al. 2006) concentrated on
74: near--maximum--light spectra of 24 SNe~Ia. In this paper we focus on
75: pre--maximum--light spectra, which form in the outermost layers of the
76: ejected matter and provide clues to the nature of the explosions that
77: cannot be obtained from later spectra that form deeper in the ejecta.
78: Our emphasis is on line identifications, which are needed in the quest
79: to develop good hydrodynamical explosion models for SNe~Ia. When
80: detailed spectrum calculations are carried out for explosion models
81: (e.g., Baron et~al. 2006; Howell et~al. 2001), discrepancies between
82: synthetic and observed spectra invariably appear. Knowing the
83: identities of the observed features provides clues to how the
84: explosion models should be modified in order to achieve better
85: agreement with observation.
86:
87: In Paper~II we divided the maximum--light spectra into four groups:
88: {\sl core--normal; broad--line; cool;} and {\sl shallow--silicon}
89: (denoted CN, BL, CL, and SS, respectively, in the figures and tables
90: of this paper). The group assignments were made on the basis of
91: measurements of the (pseudo) equivalent widths of absorption features
92: near 5750\ang\ and 6100\ang, as well as on the appearance (depth,
93: width, and shape) of the 6100\ang\ absorption, which is produced by
94: the Si~II \lam6355 transition. Although we framed the presentation
95: and discussion in terms of the four groups, in the end we concluded
96: that for the most part the spectra appeared to have a continuous
97: distribution of properties, rather than consisting of discrete
98: subtypes. Therefore, exactly how the group boundaries were drawn was
99: not important.
100:
101: Figure~1, a plot of the equivalent width $W(5750)$ of the absorption
102: feature near 5750\ang\ against the equivalent width \W610\ of the
103: absorption feature near 6100\ang\ in maximum--light spectra, is an
104: update of the corresponding figure in Paper~II, with the following
105: additional events: the core--normal SN~2003cg (Elias--Rosa
106: et~al. 2006), SN~2003du (Gerardy et al. 2004), and SN~2004S
107: (Krisciunas et~al. 2006); the broad--line SN~1983G (McCall
108: et~al. 1984) and SN~2002er (Kotak et al. 2006); and the
109: shallow--silicon SN~1999aa (Garavini et~al. 2004), SN~1999ac (Garavini
110: et~al. 2005), SN~2000E (Valentini et~al. 2003), SN~2005cg (Quimby
111: et~al. 2006), and SN~2003fg [the proposed super--Chandrasekhar SN~Ia,
112: also known as SNLS-03D3bb; Howell et~al. (2006); Jeffery, Branch, \&
113: Baron (2007a); but see also Hillebrandt, Sim, \& R\"opke (2007)].
114: SN~1999ac is a borderline case: in Figure~1 it falls between two
115: core--normals but on the basis of the shape, depth, and width of
116: \6100, we include it with the shallow--silicons.
117:
118: To a large extent the premaximum spectra turn out to exhibit the
119: defining characteristics of the four maximum--light groups of
120: Paper~II, so we use the same terminology in this paper (with all group
121: assignments based on maximum--light, not premaximum, spectra). Again,
122: we are not suggesting that SNe~Ia actually consist of discrete
123: subtypes; instead, we are thinking in terms of three inhomogeneous
124: groups radiating outward (in Figure~1) from the highly homogeneous
125: core normals.
126:
127: In Figure~1, three lines of constant $W(5750)/W(6100)$ ratio are
128: shown. This ratio is similar to although not the same as the well
129: known $R(Si~II)$ ratio (Nugent et~al. 1995), which is based on the
130: depths of the same two absorption features. Considerable diversity at
131: a given $W(5750)/W(6100)$ ratio is evident, especially at the low
132: value of 0.05, e.g., the SN~2005hk, SN~1999ee, and
133: SN~2002bf have similar $W(5750)/W(6100)$ ratios but otherwise thei
134: spectra are very different.
135:
136: In this paper, as in Papers~I and II, we confine our attention to
137: optical spectra, from the Ca~II H\&K feature in the blue
138: ($\sim3700$\ang) to the Ca~II infrared triplet (Ca~II IR3) in the red
139: ($\sim9000$\ang). All spectra have been corrected for the redshifts
140: of the parent galaxies. The original spectra have a range of slopes,
141: owing to intrinsic differences, differences in the amount of reddening
142: by interstellar dust, and observational error. Because we are
143: interested only in the spectral features, and not in the underlying
144: continuum slopes, all spectra in this paper, both observed and
145: synthetic, have been flattened according to the local normalization
146: technique of Jeffery et~al. (2007b). The flattening facilitates
147: comparison of the spectral features. Mild smoothing also has been
148: applied to some of the spectra.
149:
150: We examined all of the SN~Ia premaximum spectra available to us and
151: selected two samples: a ``one--week premax'' sample, consisting of one
152: spectrum each of 21 SNe~Ia observed between day~$-5$ and day~$-9$ with
153: respect to the time of maximum brightness in the $B$ band, and an
154: ``early'' sample, consisting of one spectrum each of eight SNe~Ia
155: observed on or earlier than day~$-11$. (All SNe of the early sample
156: are also in the one--week premax sample.) The SNe~Ia and the epochs
157: of the selected spectra are listed in Table~1.
158:
159: Continuing our attempt to provide an internally consistent
160: quantification of SN~Ia spectra, we have used the \syn\ code to fit
161: all spectra of Table~1. Descriptions of \syn\ and its use can be
162: found in Branch et~al. (2003) and Paper~I\footnote{For consistency
163: with Papers~I and II, in this paper we do not use \syn\ version~2.0,
164: described and used by Parrent et~al. (2007).}. Compared to other
165: supernova synthetic--spectrum codes, SYNOW is simple, but when working
166: in an empirical spirit from the data, as opposed to evaluating an
167: explosion model, SYNOW offers the advantage that the user has direct
168: control over line optical depths as function of radial coordinate,
169: while other codes input composition and density structures and then
170: compute the temperature structure and the line optical depths. Thus
171: SYNOW is the most useful code for making line identifications (pending
172: eventual confirmation by means of detailed calculations for explosion
173: models whose spectra and light curves match the observations). For
174: the present work, we adopt precepts for fitting spectra that are much
175: like those of Paper~II, e.g., the default excitation temperature is
176: 7000~K for the cool SN~1986G and SN~1999by and 10,000~K for all
177: others.
178:
179: \section{CORE--NORMALS}
180:
181: Figure~2 shows the spectra of the seven core--normals of the one--week
182: premax sample. The spectrum of SN~2001el differs from the others; as
183: shown by Mattila et~al. (2005), it resembles the day~$-14$ spectrum of
184: SN~1990N, so it will be discussed below, with the early sample. The
185: other spectra show considerable homogeneity, although the diversity is
186: somewhat greater than in the maximum--light core--normal sample. The
187: 6100\ang\ absorption is deeper in SN~1998bu and SN~2003cg than in the
188: others, and the broad absorption near 4300\ang\ is weaker in SN~1994D
189: and SN~1998aq than in the others. SN~2003du has strong HV\footnote{As
190: in Papers~I and II, PV refers to features that form at or near the
191: photospheric velocity and HV refers to high--velocity detached
192: features.} Ca~II absorptions near 3700\ang\ and 8000\ang\ (although
193: not nearly as strong as in the exceptional SN~2001el). SN~1990N also
194: has a strong absorption near 3700\ang. Diversity in HV Ca~II was
195: present also in maximum--light spectra of core--normals.
196:
197: The SYNOW fitting parameters for core normals of the one--week premax
198: sample are in Table~2. An example fit, to the day~$-7$ spectrum of
199: SN~2003du, is shown in Figure~3. The fit is good and most of the line
200: identifications seem clear. The most obvious discrepancy, the failure
201: of the synthetic spectrum to match the observed absorption near
202: 3980\ang, is primarily caused by Si~II \lam4130, so we can not remove
203: this discrepancy without doing harm to the fit elsewhere.
204:
205: The synthetic absorption produced by Si~III \lam4560 is a bit too blue
206: to match the observed absorption at 4400\ang\ (which is present in
207: most observed spectra of this paper). A better fit could be achieved
208: by using C~III \lam4649 instead of Si~III. But in a synthetic
209: spectrum that contains only Si~III lines the \lam4560 absorption does
210: occur at the right place to account for the 4400\ang\ absorption, and
211: Si~III \lam5743 also accounts for a weak observed absorption near
212: 5560\ang. In most premaximum spectra there is little or no evidence
213: for C~II features, so it would be surprising to see ubiquitous
214: evidence for the high--excitation (29.6 eV) C~III line. Therefore,
215: even though neither Si~III feature fits well in the full synthetic
216: spectrum (the \lam5743 absorption does not even appear as a distinct
217: feature), we assume that these two Si~III lines usually are
218: responsible for the corresponding observed features.
219:
220: The synthetic absorption produced by Si~II \lam5972 does not account
221: well for the observed absorption at 5750\ang. This and the other two
222: discrepancies mentioned above are generic to our fits. The fit could
223: be improved by invoking Fe~III lines with a higher excitation
224: potential, at or approaching 15,000~K, so that lines near 6000\ang\
225: contribute to the 5750\ang\ absorption [see Branch et~al. (2003) on
226: SN~1998aq]. But the problem with the 5750\ang\ absorption occurs in
227: some spectra in which we see no evidence for the stronger Fe~III lines
228: in the blue, so we refrain from using high--excitation Fe~III to
229: resolve this discrepancy. Another way to improve the fit, in some
230: cases, would be to use a higher Si~II optical depth and impose a
231: maximum velocity on Si~II (see \S3 and \S4).
232:
233: In our fits to the core normals of the one--week premax sample, the
234: relative weakness of the absorption near 4300\ang\ in SN~1994D and
235: SN~1998aq is accounted for by using lower Mg~II and Fe~III optical
236: depths (see Table~2). In the same two SNe we have used C~II lines, to
237: allow \lam6580 to improve the fit near the peak of the Si~II \lam6355
238: feature.
239:
240: Figure~4 shows the spectra of the three core--normals (and the two
241: broad--lines, to be discussed in \S3) of the early sample. The
242: diversity among the core--normals of the early sample is much greater
243: than in the one--week premax sample, e.g., the three 6100~\AA\
244: absorptions have distinctly different shapes, and in SN~1994D the
245: absorption trough from 4600 to 5000\ang\ is deeper than in SN~1990N
246: and SN~2003du. The early spectrum of SN~1994D is an example of group
247: crossover: the spectrum is more like that of the broad--line SN~2002bo
248: than like the other core--normals. (On the other hand, the spectrum
249: of the broad--line SN~2002er could pass for core--normal.)
250:
251: The SYNOW fitting parameters for the three core--normals (and the two
252: broad--lines) of the early sample are in Table~3. The deeper
253: 4600\ang\ to 5000\ang\ absorption trough in SN~1994D is accounted for
254: primarily by using stronger HV Fe~II.
255:
256: How to fit the various shapes of the 6100\ang\ absorption is an
257: interesting issue. For SN~1994D we fit it in the conventional way,
258: with a simple PV Si~II component (a fit is shown in Paper~I). For
259: SN~2003du, the feature appears to consist of two marginally resolved
260: components (Stanishev et~al. 2007), so we use both PV and HV Si~II.
261: The fit is shown in Figure~5. (We did not need to use HV Si~II for
262: the one--week premax spectrum of SN~2003du.)
263:
264: The unusual, nearly flat--bottomed 6100\ang\ absorption of the early
265: spectrum of SN~1990N has been discussed by several authors (Jeffery
266: et~al. 1992; Fisher et~al. 1997; Mazzali 2001). Mattila et~al. (2005)
267: showed that the one--week premax spectrum of SN~2001el resembled the
268: early spectrum of SN~1990N (see their Figure~1). As can be seen by
269: comparing our Figures~2 and 4, the HV absorption of the Ca~II infrared
270: triplet is much deeper in SN~2001el (requiring a covering factor
271: approaching unity), but otherwise the two spectra, including the
272: 6100\ang\ absorption, are quite similar. Mattila et~al. attributed
273: the flat--bottomed shape of the 6100\ang\ absorption to line formation
274: in a shell that is geometrically thin compared to the size of the
275: photosphere, but this explanation is not viable. Line formation in a
276: thin shell that is not detached from the photosphere does produce a
277: (shallow) flat--bottomed absorption [see Figure~8 of Jeffery \& Branch
278: (1990)], but the blue edge of the absorption is blueshifted only by an
279: amount that corresponds to the photospheric velocity, which is not the
280: case in the spectra of SN~1990N and SN~2001el. We can nicely fit the
281: flat--bottomed absorptions with a single Si~II component, using a high
282: value of $v_e$ and imposing a maximum velocity. However, since the
283: blue edge of the absorption corresponds to a high velocity comparable
284: to that at which the HV Ca~II and HV Fe~II are forming ($\sim20,000$
285: \kms), it seems more reasonable to assume that the the feature is is
286: an unresolved blend of PV and HV components, and fit the absorption
287: accordingly (see Table~2 for SN~2001el and Table~3 for SN~1990N).
288: Support for this interpretation is provided by a previously
289: unpublished spectrum of SN~1990N obtained at day~$-12$ in bad weather
290: by S.~Benetti and M.~Turatto (Figure~6). The absorption feature
291: resembles the marginally--resolved two--component absorptions of
292: SN~2003du (Figure~5) and SN~2005cg (Figure~13). [see also Garavini et
293: al. 2007 on 05cf]
294:
295: \section{BROAD--LINES}
296:
297: Figure~7 shows the spectrum of the core--normal SN~2003du (for
298: comparison) and the spectra of the four broad--lines of the one--week
299: premax sample. In maximum--light spectra (Paper~II) broad--lines have
300: the same spectral features and line identifications as core--normals,
301: but the equivalent width of \6100\ is larger and the features are
302: generally broader. The four broad--lines of Figure~7 also have these
303: characteristics, although as in Paper~II, the borderline case SN~1992A
304: differs from core--normal mainly in the depth and equivalent width of
305: \6100. SN~2002er is much like SN~1992A except for a weaker absorption
306: near 7450\ang. As at maximum light, SN~1984A is the most extreme
307: broad--line in the sample.
308:
309: The SYNOW fitting parameters for broad--lines of the one--week premax
310: sample are listed in Table~4. As at maximum light (Paper~II), the
311: one--week premaximum spectra of the broad--lines can be fitted with
312: the same ions as used for the core--normals, but with higher values of
313: $v_e$ and/or $\tau$.\footnote{When inspecting fitting parameters one
314: should keep in mind the partial degeneracy of $v_e$ and $\tau$. For
315: example, for SN~2002bo the synthetic absorption produced by Si~II
316: \lam6355 with $\tau = 7$ and $v_e = 2000$ \kms\ is roughly similar to
317: the corresponding absorption in SN~1992A with $\tau = 45$ and $v_e =
318: 1000$ \kms.} Figure~8 shows our fit to the day~$-7$ spectrum of the
319: extreme broad--line SN~1984A. For Si~II a high value of $v_e=4000$
320: \kms\ is used and a maximum velocity of 27,000~\kms\ is imposed.
321:
322: The spectra of the two broad--lines of the early sample are shown in
323: Figure~4, and fitting parameters are in Table~3. Fe~III is not used
324: for the early spectra of the broad--lines, but its features could be
325: hidden by the strong HV Fe~II lines. Figure~9 shows our fit to
326: the day~$-14$ spectrum of SN~2002bo. In an attempt to account for the
327: 5750\ang\ absorption with Si~II \lam5972 we have used a large optical
328: depth of 30 for Si~II and imposed a maximum velocity of 25,000 \kms.
329: A maximum velocity of 22,000 \kms\ has been imposed on S~II.
330:
331: \section{COOLS}
332:
333: Figure~10 shows the spectrum of the core--normal SN~2003du and the
334: spectra of the three cools of the one--week premax sample. For
335: reasons given in Paper~II, the borderline SN~1989B was included in the
336: cool group even though it did not show the blue trough from about
337: 4000\ang\ to 4300\ang\ that is characteristic of the other cools.
338: Here again SN~1989B has a higher \W610\ value than core--normals and
339: resembles SN~1986G at wavelengths longer than about 4400\ang, while
340: lacking the blue trough. SN~1989B remains a borderline case.
341:
342: The SYNOW fitting parameters for cools of the one--week premax sample
343: are listed in Table~5. At maximum light (Paper~II) our strategy when
344: attempting to simultaneously account for the 5750~\AA\ and 6100~\AA\
345: absorptions was to use a high optical depth and impose a maximum
346: velocity on Si~II (as we have done in this paper for some of the
347: broad--lines; see \S3). Figure~11 shows our fit to the day~$-5$
348: spectrum of SN~1999by, where we have used the same strategy with only
349: partial success; in the synthetic spectrum, the 5750~\AA\ absorption
350: is too weak even though the 6100~\AA\ is too strong. We regard the
351: line identifications in Figure~11 to be definite, except for Sc~II,
352: which helps in one place but cannot be considered definite.
353:
354: There are no cools in the early sample. Only a small number of cools
355: have been observed at all, and because they have short rise times and
356: are dim relative to other SNe~Ia, no spectra have been obtained at
357: times as early as day~$-11$.
358:
359: \section{SHALLOW--SILICONS}
360:
361: Figure~12 shows the spectra of the core--normal SN~2003du and the
362: seven shallow--silicons of the one--week premax sample. These were
363: classified as shallow--silicon on the basis of their maximum--light
364: spectra but they also have shallow--silicon at one--week premax.
365: (SN~1999ac is borderline in this respect but it is more like the
366: shallow--silicons than the core normals in other respects, such as in
367: having weak S~II absorptions.) There is plenty of diversity, however,
368: extending to the extreme cases of SN~1991T and its near twin SN~1997br
369: which at a glance do not necessarily have Si~II at all (although we do
370: include it in our fits), and to SN~2005hk. SN~2005hk is in our
371: one--week premax sample of shallow--silicons (with fitting parameters
372: in Table~6 the same as used in Chornock et~al. 2006), and SN~2002cx
373: was in our maximum--light sample of shallow--silicons. But as
374: discussed in Paper~II and elsewhere (Li et~al. 2003; Branch
375: et~al. 2004b; Jha et~al. 2006; Chornock et~al. 2006; Stanishev
376: et~al. 2007; Phillips et~al. 2007), in spite of certain spectroscopic
377: similarities to SN~1991T--likes, the SN~2002cx--likes appear to be
378: physically distinct from other SNe~Ia.
379:
380: The fitting parameters for the shallow--silicons of the one--week
381: premax sample are in Table~6. The shape of the 6100\ang\ absorption
382: in SN~2005cg is especially interesting, because it appears to consist
383: of marginally resolved PV and HV components of Si~II. Our fit is
384: shown in Figure~13. Note that in Figure~12 the 6100\ang\ absorption
385: of SN~1999ee bears some resemblance to that of SN~2005cg. As with the
386: one--week premax spectrum of SN~2001el and the early spectrum of
387: SN~1990N, the 6100\ang\ absorption of SN~1999ee can be fit with a
388: single broad PV component or an unresolved combination of a PV and an
389: HV feature. For the same reasoning as in the discussion of SN~1990N
390: and SN~2001el and the results of Garavini et~al. (2007) and Stanishev
391: et~al. (2007), we use the two components.
392:
393: Figure~14 shows the spectra of the core--normal SN~2003du and the
394: three shallow--silicons of the early sample. Again SN~1999aa and
395: SN~1991T do have shallow silicon, and again SN~1999ac is borderline.
396: Figure~15 shows our fit to the day~$-12$ spectrum of SN~1999aa. Like
397: Garavini et~al. (2004), we have resorted to C III, but the
398: identification is not definite. When fitting the HV Ca~II infrared
399: triplet absorption, the HV Ca~II H\&K absorption comes out too deep;
400: this happens in some of our other fits, although less than here. We
401: believe the absorption feature near 5500\ang\ to be Si~III \lam5743,
402: but as usual it does not appear as a distinct feature in the synthetic
403: spectrum.
404:
405: \section{DISCUSSION}
406:
407: The previous sections have been organized in terms of the four groups
408: of Paper~II, with all group assignments based on the appearance of
409: near--maximum--light spectra. To a large extent the one--week premax
410: spectra of this paper retain the defining characteristics of the four
411: groups, i.e., group assignments based on one--week premax spectra
412: would be much the same. To a lesser extent, the same is true of the
413: early spectra of this paper, e.g., almost all shallow--silicons are
414: distinct from core--normals (Figure~14). But there is some crossover
415: between groups in the early sample, e.g., Figure~4 shows that the
416: early spectra of the core--normal SN~1994D and the broad--line
417: SN~2002er are not very different. In any case, as in Paper~II we see
418: little or no evidence that SNe~Ia actually break up into distinct
419: groups, except for the SN~2002cx--likes (and possibly the cools, but
420: see Pastorello et~al. 2007 on SN~2004eo for evidence of a connection
421: between core normals and cools).
422:
423: The SYNOW fits to the premaximum spectra of this paper generally are
424: of the same quality as the fits to the maximum--light spectra of
425: Paper~II. All strong features and most weak ones can be accounted for
426: in a plausible way. Most line identifications seem clear, although as
427: discussed in previous sections there are some ambiguities, and some of
428: the weak features remain unidentified. For the most part our line
429: identifications in premaximum spectra are consistent with those
430: suggested previously on the basis of SYNOW (or SYNOW--like) synthetic
431: spectra, as well as synthetic spectra calculated with the
432: Mazzali--Lucy Monte Carlo code. Such studies have been carried out
433: for SN~1990N (Jeffery et~al. 1992; Mazzali et~al. 1993; Mazzali 2001),
434: SN~1991T (Ruiz--Lapuente et~al. 1992; Jeffery et~al. 1992; Mazzali,
435: Danziger, \& Turatto 1995; Fisher et~al. 1999), SN~1997br (Hatano
436: et~al. 2002), SN~1998aq (Branch et~al. 2003), SN~1999aa (Garavini
437: et~al. 2004), SN~1999ac (Garavini et~al. 2005), SN~1999ee (Mazzali
438: et~al. 2005a), SN~2002bo (Benetti et~al. 2004; Stehle et~al. 2005),
439: SN~2003cg (Elias-Rosa et~al. 2006), and SN~2005hk (Chornock
440: et~al. 2006).
441:
442: One of the notable differences from line identifications in some of
443: the previous papers is that neither in this paper nor in Paper~II do
444: we use PV Fe~II or Co~II (except in SN~2005hk). In Paper~I on
445: SN~1994D we first used weak PV Fe~II at day ~+4, and at later epochs
446: it became much stronger. This appears to be consistent with the
447: argument of Kasen \& Woosley (2007) that the development of extensive
448: line blocking by PV Fe~II (and Co~II) lines soon after maximum light
449: is the key to understanding the SN~Ia width--luminosity relation
450: (Phillips et~al. 1999). We will return to this issue in Paper~IV, on
451: post--maximum SN~Ia spectra.
452:
453: We do use PV Fe~III for the one--week premax sample, except in the
454: cools. It is commonly said, correctly, that SN~1991T--likes are
455: characterized by conspicuous Fe~III features. It is not true, though,
456: that the Fe~III features are particularly strong in SN~1991T--likes.
457: At one--week premax we use Fe~III optical depths for the
458: shallow--silicons that are comparable to or lower than for the
459: core--normals and the broad--lines. The same was true at maximum
460: light (although we failed to mention it in Paper~II). At maximum and
461: one--week premax, Fe~III features are conspicuous in shallow--silicons
462: only because there is less competition from other features.
463:
464: It is interesting to consider some of the implications of the fitting
465: parameters, especially the line optical depths, in the context of the
466: local--thermodynamic--equilbrium (LTE) calculations of Hatano
467: et~al. (1999a). For example, Figure~5 of Hatano et~al. shows that in
468: LTE near 7500~K the Si~II optical depth can exceed unity in a
469: composition in which hydrogen and helium have burned to carbon and
470: oxygen while the mass fraction of other elements, including silicon,
471: is only solar. This suggests that the presence of Si~II in the
472: spectrum does not necessarily require synthesized silicon. However,
473: for this composition the optical depth of Fe~II is predicted to be
474: greater than that of Si~II, which according to us is not the case in
475: premaximum or near--maximum spectra. Figure~6 of Hatano et~al. shows
476: that for a carbon--burned composition, which does include synthesized
477: silicon, the Si~II optical depth exceeds that of Fe~II by a large
478: factor. Therefore, since we do not use PV Fe~II in premaximum and
479: maximum spectra, we expect that wherever PV Si~II is detected, it is
480: produced by synthesized, rather than primordial, silicon. Another
481: interesting example is that for the core--normals of the one--week
482: premax sample, the Si~II optical depths range from four to eight (and
483: for the core--normals of the maximum--light sample it ranged from six
484: to 13). Figure~6 of Hatano et~al., for the carbon--burned
485: composition, shows that a temperature difference of only 800~K is
486: sufficient to change the Si~II optical depth by a factor of two. This
487: temperature sensitivity makes the high degree of homogeneity among the
488: core--normals all the more remarkable.
489:
490: The HV features are intriguing. Following the discovery of HV Ca~II
491: and HV Fe~II in SN~1994D (Hatano et~al. 1999b), HV Ca~II has been
492: recognized by many authors: see Li et~al. (2001), Thomas
493: et~al. (2004), and Branch et~al. (2004a) on SN~2000cx; Wang
494: et~al. (2003), Kasen et~al. (2003), and Mattila et~al. (2005) on
495: SN~2001el; Mazzali et~al. (2005a) on SN~1999ee; Gerardy et~al. (2004)
496: and Stanishev et~al. (2007) on SN~2003du; Quimby et~al. (2006) on
497: SN~2005cg; and Garavini et~al. (2007) on SN~2005cf. In premaximum
498: spectra of SNe~Ia, HV Ca~II is ubiquitous (Mazzali 2005b; and this
499: paper), and in Paper~II and we regularly invoked HV Ca~II even at
500: maximum light. The identification of HV Fe~II, which usually appears
501: at modest strength in a crowded spectral region, is perhaps not
502: definite, but we have invoked it regularly both in this paper and in
503: Paper~II at maximum light. One thing that is clear, from flux spectra
504: (Thomas et~al. 2004; Tanaka et~al. 2006) and especially from
505: polarization measurements (Wang et~al. 2003; Kasen et~al. 2003), is
506: that in at least some cases the HV Ca~II features form in asymmetrical
507: structures. The {\sl origin} of the HV features is unclear. One
508: possibility is that for one reason or another they form naturally in
509: the ejecta (Hatano et~al. 1999b). An alternative, favored by Gerardy
510: et~al. (2004) and Quimby et~al. (2006), is that they form in a shell
511: of restricted velocity interval and elevated density that is produced
512: by interaction between the ejecta and circumstellar matter (an
513: accretion disk, a filled Roche lobe, or a common envelope). Our
514: identification of HV Fe~II, if correct, may hint that the features can
515: form naturally in the ejecta, because more often than not we detached
516: HV Fe~II at a lower velocity than HV Ca~II. In the spectra of the
517: early sample, it is unlikely that the HV Fe~II features are produced
518: by iron resulting from decay the decay of $^{56}$Ni and $^{56}$Co,
519: because lines of HV Ni~II and HV Co~II also would be expected (see
520: Figure~9 of Hatano et~al. 1999a), but it is possible that the HV Fe~II
521: features are produced by iron that has been freshly synthetisized
522: along with $^{56}$Ni. Another hint that the HV features may arise
523: naturally in the ejecta is that in some cases we see HV Si~II at about
524: the same velocity as HV Ca~II and HV Fe~II. In the circumstellar
525: interaction scenario, the blue wing of the Si~II absorption is
526: predicted to be truncated rather than extended (Gerardy et~al. 2004;
527: Quimby et~al. 2006). In any case, it is interesting that that HV
528: features, of Ca~II, Fe~II, and Si~II typically are detached not far
529: from 20,000 \kms. The density and/or composition structure of most
530: SNe~Ia seems to have something special around 20,000 \kms, probably
531: including a density increase (Mazzali et~al. 2005a; Mazzali 2005b).
532:
533: The day~$-14$ spectrum of the broad--line SN~2002bo (Figure~8 and
534: Table~3) raises an issue about the definition of HV and PV features.
535: We did not use HV Ca~II or HV Si~II, and what we called HV Fe~II was
536: only mildly detached, at 22,000 \kms, from the photosphere at 20,000
537: \kms. Thus the absence of HV Ca~II or HV Si~II is a matter of
538: definition: Ca~II and Si~II are forming at 20,000 \kms, as in other
539: SNe~Ia, but they are not referred to as HV simply because of the high
540: velocity at the photosphere.
541:
542: The issue of carbon also is interesting for constraining explosion
543: models. As discussed in \S2, a C~III line could be used to improve
544: our fits for core--normals but we suspect that it would be spurious.
545: We have used C~III to improve the fits for several shallow--silicons
546: but in no case is the identification convincing. Marion et~al (2006)
547: have established that C~I lines are not detected [except in cools:
548: Howell et~al. (2001)], but their conclusion that carbon is much less
549: abundant than oxygen depends strongly on their assumption of LTE at
550: 5000~K. A temperature of 5000~K seems too low to account for the
551: strengths of other lines (e.g., Fe~III) and at higher temperatures the
552: upper limit on the carbon--to--oxygen ratio would be much lower (see
553: Figure~5 of Hatano et~al. 1999a).
554:
555: Lines of C~II offer the best prospects for detecting the presence of
556: carbon. Evidence has been presented for the strongest optical C~II
557: line, \lam6580, in SN~1998aq (Branch et~al. 2003); SN~1990N (Jeffery
558: et~al. 1992; Mazzali 2001); SN~1999ac (Garavini et~al. 2005); and
559: SN~1994D and SN~1996X (Branch et~al. 2003 and Paper~I). Thomas
560: et~al. (2007) presented the most convincing evidence, in SN~2006D. A
561: \lam6580 absorption was clearly present at day~$-7$ and day~$-5$, and
562: it weakened as maximum light was approached. In this paper we have
563: invoked C~II in these events and only one other: the early spectrum of
564: SN~2003du, but in this case C~II \lam6580 is being used mainly just to
565: beat down the Si~II \lam6355 emission peak and the identification is
566: not convincing. Thus, even at premaximum, C~II is elusive, or, as
567: Thomas et~al. put it, sporadic, perhaps due to its tendency to appear
568: in clumps. If the clumps are few in number and none happen to be in
569: front of the photosphere, no absorption features will be seen. Even
570: if a clump is in front of the photosphere, if it has sufficiently high
571: line--of--sight velocity (about 15,000 \kms\ or more) the \lam6580
572: absorption may be blueshifted into the Si~II absorption and not be
573: detectable. Therefore, although C~II is seen in only a fraction of
574: premaximum SN~Ia spectra, the ubiquitous presence of carbon clumps in
575: the outer layers of SNe~Ia is not excluded. Unburned carbon at the
576: observed velocities is not characteristic of published
577: delayed--detonation models.
578:
579: Another issue to be addressed on the basis of premaximum spectra is
580: that of the maximum detectable ejection velocities. From Figures~2,
581: 4, 7, 10, 12, and 14 we can see that Ca~II absorption is generally
582: detectable up to about 25,000 \kms\ and sometimes up to about 30,000
583: \kms. The highest maximum velocity that has been imposed on Ca~II is
584: 34,000 \kms\ for SN~2001el. Detectable optical depth in Ca~II can
585: easily be produced with a solar mass fraction of Ca~II, so the maximum
586: velocities of synthesized Ca~II may be lower. From the same figures
587: we can see that Si~II absorption is generally detectable up to 15,000
588: \kms\ and sometimes above 20,000 \kms. The highest maximum velocity
589: that has been imposed on Si~II is 27,000 \kms\ for the one--week
590: premax spectrum of SN~1984A. We have argued above that this
591: absorption is produced by synthesized silicon. This is consistent
592: with the 11,900 \kms\ that Mazzali, Benetti, \& Hillebrandt (2007)
593: adopted for a lower limit to the top of the silicon--rich layer, but
594: the silicon--rich layer seems to generally extend to much higher
595: velocities. Such high silicon velocities are not characteristic of
596: pure deflagration models. The presence of carbon at intermediate
597: velocities and silicon at high velocities could be consistent with the
598: pulsating reverse detonation models of Bravo \& Garcia--Senz (2006).
599:
600: Another velocity--related issue is the possibility that some of the
601: P--Cygni emission components are blueshifted, as they sometimes are in
602: SNe~II (Chugai 1988; Dessart \& Hiller 2005). From measurements of the
603: wavelengths of flux maxima Blondin et~al. (2006) inferred
604: emission--line blueshifts of up to 8000 \kms\ in premaximum SN~Ia
605: spectra. The assumptions on which SYNOW is based (sharp photosphere
606: and resonant scattering) do not allow blueshifts of unblended
607: emission--line peaks, yet our fits do not in general have a problem
608: reproducing the wavelengths of the observed flux maxima. In SYNOW
609: synthetic spectra the flux maxima are affected by line blending;
610: sometimes they are a bit too red, consistent with an emission--line
611: blueshift, but equally often they are a bit too blue. In SNe~Ia, line
612: blending is a serious problem for establishing that emission--line
613: peaks are blueshifted.
614:
615: The premaximum spectra of SNe~Ia exhibit a rich diversity, which even
616: within each of our four groups probably is multidimensional. The
617: outer layers of SN~Ia ejecta differ in composition, density, and
618: temperature structure, and they have asymmetries - probably both
619: global shape asymmetries as well as composition clumping.
620: Understanding the observational diversity in terms of the physical
621: differences is a challenging task that will require more observations
622: and modeling. Flux spectra have much to tell us, but in view of the
623: asymmetries, polarization spectra are a high priority.
624:
625: We are grateful to Stefano Benetti and Massimo Turatto for providing
626: the previously unpublished spectrum of SN~1990N, and to all observers
627: who have provided their published spectra. This work has been
628: supported by NSF grants AST-0204771 and AST-0506028, and NASA LTSA
629: grant NNG04GD36G.
630:
631: \clearpage
632:
633: \begin{references}
634:
635: \reference{} Anupama, G. C, Sahu, D. K., \& Jose, J. 2005, A\&A, 429, 667
636:
637: \reference{} Baron, E., Bongard, S., Branch, D., \&
638: Hauschildt,~P.~H. 2006, ApJ, 645, 480
639:
640: \reference{} Benetti, S., et al. 2004, MNRAS, 348, 261
641:
642: \reference{} Blondin, S., et al. 2006, ApJ, 131, 1648
643:
644: \reference{} Branch, D., Baron, E., Hall,~N., Melakayil,~M., \&
645: Parrent,~J. 2005, PASP, 117, 545 (Paper~I)
646:
647: \reference{} Branch, D., et al. 2003, AJ, 126, 1489
648:
649: \reference{} Branch, D., et al. 2004a, ApJ, 606, 413
650:
651: \reference{} Branch, D., et al. 2004b, PASP, 116, 903
652:
653: \reference{} Branch, D., et al. 2006, PASP, 118, 560 (Paper~II)
654:
655: \reference{} Bravo, E., \& Garcia--Senz, D. 2006, ApJ, 642, L157
656:
657: \reference{} Chornock, D., et al. 2006, PASP, 118, 722
658:
659: \reference{} Chugai, N. N. 1988, Sov. Astron. Lett., 14(5), 334
660:
661: \reference{} Cristiani, S., et al. 1992, A\&A, 259, 63
662:
663: \reference{} Dessart, L., \& Hillier, D. J. 2005, A\&A, 437, 667
664:
665: \reference{} Elias--Rosa, N., et al. 2006, MNRAS, 369, 1880
666:
667: \reference{} Fisher, A., Branch, D., Hatano, K., \& Baron,~E. 1999,
668: MNRAS, 304, 67
669:
670: \reference{} Fisher, A., Branch, D., Nugent, P., \& Baron,~E. 1997,
671: ApJ, 481, L89
672:
673: \reference{} Garavini, G., et al. 2004, AJ, 128, 387
674:
675: \reference{} Garavini, G., et al. 2005, AJ, 130, 2278
676:
677: \reference{} Garavini, G., et al. 2007, A\&A, in press
678:
679: \reference{} Garnavich, P. M., et al. 2004, ApJ, 613, 1120
680:
681: \reference{} Gerardy, C., et al. 2004, ApJ, 607, 391
682:
683: \reference{} Hamuy, M., et al. 2002, AJ, 124, 417
684:
685: \reference{} Hatano, K., Branch, D., Fisher, A., Baron,~E., \&
686: Filippenko,~A.~V. 1999b, ApJ, 525, 881
687:
688: \reference{} Hatano, K., Branch, D., Fisher, A., Millard,~J., \&
689: Baron,~E. 1999a, ApJS, 121, 233
690:
691: \reference{} Hatano, K., Branch, D., Qiu, Y. L., Baron,~E.,
692: Thielemann,~F.-K., \& Fisher, A. 2002, New Astr., 7, 441
693:
694: \reference{} Hernandez, M., et al. 2000, MNRAS, 319, 223
695:
696: \reference{} Hillebrandt, W., Sim, S. A., \& R\"opke,~F.~K. 2007,
697: A\&A, 465, L17
698:
699: \reference{} Howell, D. A., H\"oflich, P., Wang, L., \&
700: Wheeler,~J.~C. 2001, ApJ, 556, 302
701:
702: \reference{} Howell, D. A., et~al. 2006, Nature, 443, 308
703:
704: \reference{} Jeffery, D. J. \& Branch, D. 1990, in Supernovae,
705: eds. J.~C.~Wheeler, T.~Piran, \& S.~Weinberg (Singapore: World
706: Scientific), 149
707:
708: \reference{} Jeffery, D. J., Branch, D., \& Baron, E. 2007a, ApJ,
709: submitted
710:
711: \reference{} Jeffery, D. J., Ketchum, W., Branch, D., Baron,~E.,
712: Elmhamdi,~A., \& Danziger,~I.~J. 2007b, ApJS, in press
713:
714: \reference{} Jeffery, D. J., Leibundgut, B., Kirshner, R. P.,
715: Benetti,~S., Branch,~D., \& Sonneborn,~G. 1992, ApJ, 397, 304
716:
717: \reference{} Jha, S., et~al. 2006, AJ, 132, 189
718:
719: \reference{} Kasen, D., \& Woosley, S. E. 2007, ApJ, 656, 661
720:
721: \reference{} Kasen, D., et al. 2003, ApJ, 593, 788
722:
723: \reference{} Kotak, R., et al. 2006, A\&A, 436, 1021
724:
725: \reference{} Krisciunas, K., et al. 2006, AJ, 133, 58
726:
727: \reference{} Leibundgut, B., Kirshner, R. P., Filippenko, A. V.,
728: Shields,~J.~S, Foltz,~C.~B., Phillips,~M.~M., \& Sonneborn,~S. 1991,
729: ApJ, 371, L23
730:
731: \reference{} Li, W., et al. 1999, AJ, 117, 2709
732:
733: \reference{} Li, W., et al. 2001, PASP, 113, 1178
734:
735: \reference{} Li, W., et al. 2003, PASP, 115, 453
736:
737: \reference{} Marion, G. H., H\"oflich, P., Wheeler,~J.~C.,
738: Robinson,~E.~L., Gerardy,~C.~L., \& Vacca,~W.~D. 2006, ApJ, 645, 1392
739:
740: \reference{} Mattila, S., et~al. 2005, A\&A, 443, 649
741:
742: \reference{} Mazzali, P. A. 2001, MNRAS, 321, 341
743:
744: \reference{} Mazzali, P. A., Benetti, S., \& Hillebrandt, W. 2007,
745: Science, 315, 825
746:
747: \reference{} Mazzali, P. A., Danziger, I. J., \& Turatto, M. 1995,
748: A\&A 297, 509
749:
750: \reference{} Mazzali, P. A., Lucy, L. B., Danziger, I. J.,
751: Gouiffes,~C., Cappellaro,~E., \& Turatto,~M. 1993,
752: A\&A 269, 423
753:
754: \reference{} Mazzali, P. A., et~al. 2005a, MNRAS, 357, 200
755:
756: \reference{} Mazzali, P. A., et~al. 2005b, ApJ, 623, L37
757:
758: \reference{} McCall, M. L., Reid, N., Bessell, M. S., \&
759: Wickramasinghe,~D. 1984, MNRAS, 210, 839
760:
761: \reference{} Meikle, W. P. S., et al. 1996, MNRAS, 281, 263
762:
763: \reference{} Nugent, P., Phillips, M. M., Baron, E., Branch,~D., \&
764: Hauschildt,~P.~H. 1995, ApJ, 455, L147
765:
766: \reference{} Parrent, J., et al. 2007, PASP, 119, 135
767:
768: \reference{} Pastorello et~al. 2007, MNRAS, 377, 1531
769:
770: \reference{} Phillips, M. M., Lira, P., Suntzeff, N. B.,
771: Schommer,~R.~A., Hamuy,~M., \& Maza,~J. 1999, AJ,
772: 118, 1766
773:
774: \reference{} Phillips, M. M., Wells, L. A., Suntzeff, N. B.,
775: Hamuy,~M., Leibundgut,~B., Kirshner,~R.~P., \& Foltz,~C.~B. 1992, AJ,
776: 103, 1632
777:
778: \reference{} Phillips, M. M., et al. 2007, PASP, 119, 360
779:
780: \reference{} Quimby, R., et al. 2006, ApJ, 636, 400
781:
782: \reference{} Ruiz--Lapuente, P., Cappellaro, E., Turatto, M.,
783: Gouiffes,~C., Danziger,~I.~J., Della Valle, M., \& Lucy,~L.~B. 1992,
784: ApJ, 387, L33
785:
786: \reference{} Stanishev, V., et al. 2007, A\&A, in press
787:
788: \reference{} Stehle, M., Mazzali, P. A., Benetti, S., \&
789: Hillebrandt,~W. 2005, MNRAS, 360, 1231
790:
791: \reference{} Tanaka, M., Mazzali, P. A., Maeda, K., \&
792: Nomoto,~K. 2006, ApJ, 645, 470
793:
794: \reference{} Thomas, R. C., Branch, D., Baron, E., Nomoto,~K., Li,~W., \&
795: Filippenko,~A.~V. 2004, ApJ, 601, 1019
796:
797: \reference{} Thomas, R. C., et~al. 2007, ApJ, 654, L53
798:
799: \reference{} Valentini, G., et al. 2003, 595, 779
800:
801: \reference{} Wang, L., et al. 2003, ApJ, 591, 1110
802:
803: \reference{} Wegner, G., \& McMahan, R. K. 1987, AJ, 92, 287
804:
805: \reference{} Wells, L. A., et al. 1994, AJ, 108, 2233
806:
807: \end{references}
808:
809: \clearpage
810:
811: \begin{figure}
812: \includegraphics[width=.8\textwidth,angle=270]{f1.eps}
813: \caption{An update of Figure~1 of Paper~II: $W(5750)$ is plotted
814: against $W(6100)$ for spectra obtained within three days of maximum
815: light. Core--normal SNe~Ia are shown as {\sl circles}, broad--line
816: SNe~Ia as {\sl squares}, cool SNe~Ia as {\sl diamonds}, and
817: shallow--silicon SNe~Ia as {\sl triangles}. The labels of the {\sl
818: dashed lines} refer to the $W(5750)/W(6100)$ ratio.}
819: \end{figure}
820:
821: \clearpage
822:
823: \begin{figure}
824: \includegraphics[width=.8\textwidth,angle=0]{f2.eps}
825: \caption{Spectra of seven core--normals of the one--week premax
826: sample. The spectra have been flattened as described in the text.
827: Vertical displacements are arbitrary and narrow absorptions near
828: 7600\ang\ and 6900\ang\ are telluric. Vertical {\sl dashed lines}
829: refer to Ca~II \lam3945 and \lam8579 blueshifted by 25,000 \kms\ and
830: Si~II \lam6355 blueshifted by 15,000 \kms.}
831: \end{figure}
832:
833: \clearpage
834:
835: \begin{figure}
836: \includegraphics[width=.8\textwidth,angle=270]{f3.eps}
837: \caption{The day~$-7$ spectrum of the core--normal SN~2003du ({\sl
838: solid line}), from Anupama et~al. (2005), is compared with a
839: synthetic spectrum ({\sl dashed line}).}
840: \end{figure}
841:
842: \clearpage
843:
844: \begin{figure}
845: \includegraphics[width=.8\textwidth,angle=0]{f4.eps}
846: \caption{Spectra of three core--normals and two broad--lines of
847: the early sample. Vertical {\sl dashed lines}
848: refer to Ca~II \lam3945 blueshifted by 30,000 \kms,
849: Si~II \lam6355 blueshifted by 20,000 \kms, and Ca~II \lam8579 by
850: 25,000 \kms. }
851: \end{figure}
852:
853: \clearpage
854:
855: \begin{figure}
856: \includegraphics[width=.8\textwidth,angle=270]{f5.eps}
857: \caption{The day~$-11$ of the core--normal SN~2003du ({\sl solid
858: line}), from Stanishev et~al. (2007), is compared with a synthetic
859: spectrum ({\sl dashed line}).}
860: \end{figure}
861:
862: \clearpage
863:
864: \begin{figure}
865: \includegraphics[width=.8\textwidth,angle=0]{f6.eps}
866: \caption{Three spectra of SN~1990N are compared: days~$-14$ and $-8$,
867: from Leibundgut et~al. (1991), and a previously unpublished spectrum
868: obtained on day~$-12$ by S.~Benetti and M.~Turatto. Vertical {\sl
869: dashed lines} refer to Si~II \lam6355 blueshifted by 20,000 and
870: 12,000 \kms\ and Ca~II \lam8579 blueshifted by 25,000 \kms. The
871: 6100\ang\ absorption in the day~$-12$ spectrum appears to be a
872: marginally resolved two--component (PV + HV) feature.}
873: \end{figure}
874:
875: \clearpage
876:
877: \begin{figure}
878: \includegraphics[width=.8\textwidth,angle=0]{f7.eps}
879: \caption{Spectra of one core--normal and four broad--line SN~Ia of
880: the one--week premax sample. Vertical {\sl dashed lines}
881: refer to Ca~II \lam3945 and \lam8579 blueshifted by 25,000 \kms\ and
882: Si~II \lam6355 blueshifted by 20,000 \kms. }
883: \end{figure}
884:
885: \clearpage
886:
887: \begin{figure}
888: \includegraphics[width=.8\textwidth,angle=270]{f8.eps}
889: \caption{The day~$-7$ spectrum of the extreme broad--line SN~1984A
890: ({\sl solid line}), from Wegner \& McMahan (1987), is compared with
891: a synthetic spectrum ({\sl dashed line}).}
892: \end{figure}
893:
894: \clearpage
895:
896: \begin{figure}
897: \includegraphics[width=.8\textwidth,angle=270]{f9.eps}
898: \caption{The day~$-14$ spectrum of the broad--line SN~2002bo ({\sl
899: solid line}), from Benetti et~al. (2004), is compared with a
900: synthetic spectrum ({\sl dashed line}).}
901: \end{figure}
902:
903: \clearpage
904:
905: \begin{figure}
906: \includegraphics[width=.8\textwidth,angle=0]{f10.eps}
907: \caption{Spectra of one core--normal and three cools of the one--week
908: premax sample. Vertical {\sl dashed lines} refer to Ca~II \lam3945
909: blueshifted by 20,000 \kms\ and Si~II \lam6355 blueshifted by 15,000
910: \kms. }
911: \end{figure}
912:
913: \clearpage
914:
915: \begin{figure}
916: \includegraphics[width=.8\textwidth,angle=270]{f11.eps}
917: \caption{The day~$-5$ spectrum of the cool SN~1999by ({\sl solid
918: line}), from Garnavich et~al. (2004), is compared with a synthetic
919: spectrum ({\sl dashed line}).}
920: \end{figure}
921:
922: \clearpage
923:
924: \begin{figure}
925: \includegraphics[width=.8\textwidth,angle=0]{f12.eps}
926: \caption{Spectra of one core--normal and seven shallow--silicons of
927: the one--week premax sample. Vertical {\sl dashed lines} refer to
928: Ca~II \lam3945 and \lam8579 blueshifted by 25,000 \kms\ and Si~II
929: \lam6355 blueshifted by 15,000 \kms. }
930: \end{figure}
931:
932: \clearpage
933:
934: \begin{figure}
935: \includegraphics[width=.8\textwidth,angle=270]{f13.eps}
936: \caption{The day~$-9$ spectrum of the shallow--silicon SN~2005cg ({\sl
937: solid line}), from Quimby et~al. (2006), is compared with a
938: synthetic spectrum ({\sl dashed line}).}
939: \end{figure}
940:
941: \clearpage
942:
943: \begin{figure}
944: \includegraphics[width=.8\textwidth,angle=0]{f14.eps}
945: \caption{Spectra of one core--normal and three shallow--silicons of
946: the early sample. Vertical {\sl dashed lines} refer to Ca~II \lam3945
947: and \lam8579 blueshifted by 25,000 \kms\ and Si~II \lam6355
948: blueshifted by 20,000 \kms. }
949: \end{figure}
950:
951: \clearpage
952:
953: \begin{figure}
954: \includegraphics[width=.8\textwidth,angle=270]{f15.eps}
955: \caption{The day~$-12$ spectrum of the shallow--silicon SN~1999aa
956: ({\sl solid line}), from Garavini et~al. (2004), is compared with a
957: synthetic spectrum ({\sl dashed line}).}
958: \end{figure}
959:
960: \clearpage
961:
962: \begin{deluxetable}{llll}
963: %\tabletypesize{\scriptsize}
964: %\rotate
965: \tablenum{1}
966: %\tablewidth{1000pt}
967: \setlength{\tabcolsep}{4pt}
968:
969: \tablecaption{The SN~Ia Sample}
970:
971: \tablehead{\colhead{SN} & \colhead{Epochs} & \colhead{Galaxy} &
972: \colhead{References} \\
973:
974: \colhead{} & \colhead{(days)} & \colhead{} & \colhead{} }
975:
976: \startdata
977:
978: 1984A BL & $-$7 & NGC 4419 & Wegner \& McMahan (1987)\\
979:
980: 1986G CL & $-$6 & NGC 5128 & Cristiani et al. (1992)\\
981:
982: 1989B CL & $-$7 & NGC 3627 & Wells et al. (1994)\\
983:
984: 1990N CN & $-$8, $-$14 & NGC 4639 & Leibundgut et al. (1991)\\
985:
986: 1991T SS & $-$7, $-$11 & NGC 4527 & Phillips et~al. (1992)\\
987:
988: 1992A BL & $-$6 & NGC 1380 & P.~Challis, unpublished\\
989:
990: 1994D CN & $-$8, $-12$ & NGC 4526 & Meikle et~al. (1996)\\
991:
992: 1997br SS & $-7$ & ESO 576-G40 & Li et~al. (1999)\\
993:
994: 1998aq CN & $-$8 & NGC 3982 & Branch et al. (2003)\\
995:
996: 1998bu CN & $-$6 & NGC 3368 & Hernandez et~al. (2000)\\
997:
998: 1999aa SS & $-7$, $-$12 & NGC 4469 & Garavini et~al. (2004)\\
999:
1000: 1999ac SS & $-9$, $-15$ & NGC 2848 & Garavini et~al. (2005) \\
1001:
1002: 1999by CL & $-$5 & NGC 2841 & Garnavich et al. (2004)\\
1003:
1004: 1999ee SS & $-$7 & IC 5179 & Hamuy et al. (2002)\\
1005:
1006: 2001el CN & $-8$ & NGC 1448 & Mattila et al. (2005)\\
1007:
1008: 2002bo BL & $-$6, $-$14 & NGC 3190 & Benetti et
1009: al. (2004)\\
1010:
1011: 2002er BL & $-7$, $-11$ & UGC 10743 & Kotak et al. (2006)\\
1012:
1013: 2003cg CN & $-7$ & NGC 3169 & Elias--Rosa et al. (2006)\\
1014:
1015: 2003du CN & $-$7, $-11$ & NGC 1921 & Anupama et al. (2005), Stanishev
1016: et al. (2007)\\
1017:
1018: 2005cg SS & $-9$ & NGC 9290 & Quimby et al. (2006)\\
1019:
1020: 2005hk SS & $-5$ & UGC 272 & Chornock et al. (2006)\\
1021:
1022: \enddata
1023: \end{deluxetable}
1024:
1025: \clearpage
1026:
1027: \begin{deluxetable}{lccccccc}
1028: %\tabletypesize{\scriptsize}
1029: \rotate
1030: \tablenum{2}
1031: %\tablewidth{1000pt}
1032: \setlength{\tabcolsep}{4pt}
1033:
1034: \tablecaption{Fitting Parameters for Core--Normal SNe~Ia of the
1035: One--Week Premax Sample}
1036:
1037: \tablehead{\colhead{Parameter} & \colhead{SN~1990N} &
1038: \colhead{SN~1994D} & \colhead{SN~1998aq} & \colhead{SN~1998bu}
1039: & \colhead{SN~2001el} & \colhead{SN~2003cg} & \colhead{SN~2003du} }
1040:
1041: \startdata
1042:
1043: \vphot\ (\kms) & 12,000 & 13,000 & 13,000 & 11,000 & 15,000 & 13,000 &
1044: 13,000 \\
1045:
1046: \ve\ (\kms)& 1000 & 1000 & 1000 & 1000 & 2000 & 1000 & 1000\\
1047:
1048: $\tau$(C~II) & \nodata & 0.8 & 0.4/14 & \nodata & \nodata & \nodata &
1049: \nodata \\
1050:
1051: $\tau$(O~I) & \nodata & 0.5/14 & \nodata & 0.6/12 & \nodata & 0.8 & 0.3 \\
1052:
1053: $\tau$(Mg~II) & 1.2 & 0.5 & 0.4 & 1.5 & 0.5 & 0.7 & 0.5 \\
1054:
1055: $\tau$(Si~II) & 4 & 5 & 5 & 8 & 2& 8 & 5 \\
1056:
1057: $\tau$(HV Si~II) & \nodata & \nodata & \nodata & \nodata &
1058: 0.5/20 & 8 & 5 \\
1059:
1060: $\tau$(Si~III) & 1.1 & 0.6 & 1.8 & 1.8 & 0.2 & 1.7 & 2.2 \\
1061:
1062: $\tau$(S~II) & 1.7 & 1.5 & 1.5 & 2 & 0.4 & 3 & 1.8\\
1063:
1064: $\tau$(Ca~II) & 5(2) & 10 & 4 & 25 & \nodata & 25 & 6(2) \\
1065:
1066: $\tau$(HV Ca~II) & 1.8(4)/20 & 2.5(3)/23 & \nodata &1.6(2)/20 &
1067: 20(8)/23[34] & 5/22 & 5(3)/21 \\
1068:
1069: $\tau$(HV Fe~II) & 0.2(3)/17 & 0.25(2)/19 & 0.2(2)/19 & 0.2(2)/17 &
1070: 1/20 & 0.4/19 & 0.4/19 \\
1071:
1072: $\tau$(Fe~III) & 1.2 & 0.5 & 0.7 & 1.5 & \nodata & 1.5 & 0.8 \\
1073:
1074: \enddata
1075:
1076: \tablenotetext{a}{For each ion the optical depth, $\tau$, is the
1077: optical depth at the photosphere or detachment velocity of the ion's
1078: reference line (ordinarily the ion's strongest line in the optical
1079: spectrum). When a value of \ve\ other than the default value in row~2
1080: is used, the value (in units of 1000~\kms) is given in parentheses.
1081: Minimum and maximum velocities (in units of 1000~\kms) are preceded by
1082: a forward slash. Maximum velocities are in square brackets.}
1083:
1084: \end{deluxetable}
1085:
1086: \begin{deluxetable}{lccccc}
1087: %\tabletypesize{\scriptsize}
1088: \rotate
1089: \tablenum{3}
1090: %\tablewidth{1000pt}
1091: \setlength{\tabcolsep}{4pt}
1092:
1093: \tablecaption{Fitting Parameters for Three Core--Normal and Two
1094: Broad--line SNe~Ia of the Early Sample}
1095:
1096: \tablehead{\colhead{Parameter} & \colhead{SN~1990N} &
1097: \colhead{SN~1994D} & \colhead{SN~2003du} & \colhead{SN~2002bo}&
1098: \colhead{SN~2002er} }
1099:
1100: \startdata
1101:
1102: \vphot\ (\kms) & 14,000 & 14,000 & 14,000 & 20,000 & 15,000 \\
1103:
1104: \ve\ (\kms)& 2000 & 2000 & 1000 & 2000 & 2000 \\
1105:
1106: $\tau$(C~II) & 0.5 & 0.6 & 0.6 & \nodata & \nodata \\
1107:
1108: $\tau$(O~I) & \nodata & 0.5 & 0.3/15 & 0.2(3)/21 & \nodata \\
1109:
1110: $\tau$(Mg~II) & \nodata & \nodata & 0.6 & 1 & 1 \\
1111:
1112: $\tau$(Si~II) & 1.6 & 8 & 8 & 30[25] & 6[20]\\
1113:
1114: $\tau$(HV Si~II) & 0.7(3)/20 & \nodata & 0.5(2)/20 & \nodata & \nodata \\
1115:
1116: $\tau$(Si~III) & 0.3 & 0.7 & 2.5 & 0.6 & 0.2\\
1117:
1118: $\tau$(S~II) & 0.5 & 0.6 & 1.8 & 1[22] & 1.5[16]\\
1119:
1120: $\tau$(Ca~II) & \nodata & 200 & 12(2) & 20(10) & 10\\
1121:
1122: $\tau$(HV Ca~II) & 8(6)/24[30] & 200(6)/23 & 12(5)/22 & \nodata & 100(6)/23\\
1123:
1124: $\tau$(HV Fe~II) & 0.3(3)/20 & 1.2(3)/20 & 0.3(3)/19 & 1(3)/22 & 0.7(3)/20\\
1125:
1126: $\tau$(Fe~III) & 0.5 & \nodata & 0.8 & \nodata & \nodata\\
1127:
1128: \enddata
1129:
1130: \tablenotetext{a}{For column descriptions see the note to Table~2.}
1131:
1132: \end{deluxetable}
1133:
1134: \clearpage
1135:
1136: \begin{deluxetable}{lcccc}
1137: %\tabletypesize{\scriptsize}
1138: \rotate
1139: \tablenum{4}
1140: %\tablewidth{1000pt}
1141: \setlength{\tabcolsep}{4pt}
1142:
1143: \tablecaption{Fitting Parameters for Broad--Line SNe~Ia of the
1144: One--Week Premax Sample}
1145:
1146: \tablehead{\colhead{Parameter} & \colhead{SN~1984A} &
1147: \colhead{SN~1992A} & \colhead{SN~2002bo} &\colhead{SN~2002er} }
1148:
1149: \startdata
1150:
1151: \vphot\ (\kms) & 16,000 & 14,000 & 14,000 &16,000\\
1152:
1153: \ve\ (\kms)& 2000 & 1000 & 2000 &1000\\
1154:
1155: $\tau$(O~I) & \nodata & 1.5/16 & 0.15(3)/15 &0.1 \\
1156:
1157: $\tau$(Mg~II) & 3 & 1.2(2) & 0.8 &1.7(2)\\
1158:
1159: $\tau$(Si~II) & 10(4)/[27] & 45 & 7 & 35\\
1160:
1161: $\tau$(Si~III) & 0.6 & 0.5 & 0.5 &0.7 \\
1162:
1163: $\tau$(S~II) & 2.3 & 2.5 & 1.7 &3.5\\
1164:
1165: $\tau$(Ca~II) & 20(6) & 50 & 120 &150\\
1166:
1167: $\tau$(HV Ca~II) & 20(6)/21 &2(20)/24[28] & 5(3)/21 & 8(20)/24[29]\\
1168:
1169: $\tau$(HV Fe~II) & 1.7(3)/24 & 0.8/19 & 1(3)/19 & 1.5/20\\
1170:
1171: $\tau$(Fe~III) & 1.4 & 0.5 & 1 &0.4 \\
1172:
1173: \enddata
1174: \tablenotetext{a}{For column descriptions see the note to Table~2.}
1175: \end{deluxetable}
1176: %\end{document}
1177:
1178: \clearpage
1179:
1180: \begin{deluxetable}{lccc}
1181: %\tabletypesize{\scriptsize}
1182: \rotate
1183: \tablenum{5}
1184: %\tablewidth{1000pt}
1185: \setlength{\tabcolsep}{4pt}
1186:
1187: \tablecaption{Fitting Parameters for Cool SNe~Ia of the One--Week
1188: Premax Sample}
1189:
1190: \tablehead{\colhead{ } & \colhead{SN~1989B} & \colhead{SN~1986G} &
1191: \colhead{SN~1999by} }
1192:
1193: \startdata
1194:
1195: \vphot\ (\kms) & 12,000 & 12,000 & 11,000 \\
1196:
1197: \ve\ (\kms)& 1000 & 1000 & 1000 \\
1198:
1199: $\tau$(O~I) & 1/13 & \nodata & 4 \\
1200:
1201: $\tau$(Mg~I) & \nodata & 2.5 & 1.8 \\
1202:
1203: $\tau$(Mg~II) & 1/13 & \nodata & 2 \\
1204:
1205: $\tau$(Si~II) & 20 & 40(2)/[19] & 60/[14]\\
1206:
1207: $\tau$(Si~III) & 0.7 & \nodata & \nodata\\
1208:
1209: $\tau$(S~II) & 2 & 2 & 1\\
1210:
1211: $\tau$(Ca~II) & 10(2) & \nodata & 40 \\
1212:
1213: $\tau$(HV Ca~II) & 1(2)/20 & \nodata & \nodata \\
1214:
1215: $\tau$(Sc~II) & \nodata & \nodata & 1.5 \\
1216:
1217: $\tau$(Ti~II) & \nodata & \nodata & 1.5 \\
1218:
1219: $\tau$(HV Fe~II) & 1/17 & \nodata & \nodata \\
1220:
1221: $\tau$(Fe~III) & 1.5 & \nodata & \nodata \\
1222:
1223: \enddata
1224: \tablenotetext{a}{For column descriptions see the note to Table~2.}
1225: \end{deluxetable}
1226:
1227: \clearpage
1228:
1229: \begin{deluxetable}{lccccccc}
1230: %\tabletypesize{\scriptsize}
1231: \rotate
1232: \tablenum{6}
1233: %\tablewidth{1000pt}
1234: \setlength{\tabcolsep}{4pt}
1235:
1236: \tablecaption{Fitting Parameters for Shallow--Silicon SNe~Ia of the
1237: One--Week Premax Sample}
1238:
1239: \tablehead{\colhead{ } & \colhead{SN~1999ee} & \colhead{SN~1991T} &
1240: \colhead{SN~1997br} & \colhead{SN2005hk} & \colhead{SN~1999aa} &
1241: \colhead{SN~1999ac} & \colhead{SN~2005cg}}
1242:
1243: \startdata
1244:
1245: \vphot\ (\kms)& 13,000 & 10,000 & 12,000 & 5000 & 11,000 & 13,000 & 15,000\\
1246:
1247: \ve\ (\kms)& 2000 & 1000 & 1000 & 2000 & 2000 &2000 & 1000\\
1248:
1249: $\tau$(C~III) & \nodata & 0.3 & \nodata & 0.3 & \nodata &
1250: \nodata &\nodata\\
1251:
1252: $\tau$(O~I) & 0.1 & 0.1 & \nodata & 0.08/6 & 0.1/12 & 0.1/14 &0.5/17 \\
1253:
1254: $\tau$(Mg~II) & 0.4(3) & 0.3 & \nodata & 0.35/6 & 0.3(3) & 0.4(3)/14 &0.3(4)\\
1255:
1256: $\tau$(Si~II) & 0.9 & 0.2 & 0.2 & 0.2 & 0.6 & 1.8 &4 \\
1257:
1258: $\tau$(HV Si~II) & 0.4/19 & \nodata & \nodata & \nodata & \nodata &
1259: \nodata & 0.45(2)/21\\
1260:
1261: $\tau$(Si~III) & 0.5 & 1.3 & 0.6 & 0.3 & 0.5 & 0.9 &0.8\\
1262:
1263: $\tau$(S~II) & 0.4 & \nodata & \nodata & 0.25 & 0.5 & 0.4 &1\\
1264:
1265: $\tau$(Ca~II) & 5 & 1 & 0.3 & 0.8 & 1.5(3) & 4(3) &15\\
1266:
1267: $\tau$(HV Ca~II) & 4(9)/20[28] & 0.6(3)/18 & 0.2(2)/18 & \nodata &
1268: 1.3(9)/21[28] & 1.5(6)/19[23] &5(20)/23[32]\\
1269:
1270: $\tau$(Ti~II) & \nodata & \nodata & \nodata & 0.04 & \nodata &
1271: \nodata &\nodata\\
1272:
1273: $\tau$(HV Fe~II) & 0.5/22 & \nodata & \nodata & \nodata & 0.2/21 &
1274: 0.3/19 &\nodata\\
1275:
1276: $\tau$(Fe~III) & 0.6 & 0.7(2) & 0.8(2) & 0.7 & 0.6 & 0.5 &0.8\\
1277:
1278: $\tau$(Co~II) & \nodata & \nodata & \nodata & 0.2 & \nodata &
1279: \nodata &\nodata\\
1280:
1281: $\tau$(Ni~II) & \nodata & \nodata & \nodata & 0.1 & \nodata &
1282: \nodata &\nodata\\
1283:
1284: $\tau$(Ni~III) & \nodata & 0.1(2) & 0.15 & 0.1 & \nodata &\nodata &\nodata\\
1285: \enddata
1286: \tablenotetext{a}{For column descriptions see the note to Table~2.}
1287: \end{deluxetable}
1288:
1289: %\end{document}
1290:
1291: \clearpage
1292:
1293: \begin{deluxetable}{lcccc}
1294: %\tabletypesize{\scriptsize}
1295: \rotate
1296: \tablenum{7}
1297: %\tablewidth{1000pt}
1298: \setlength{\tabcolsep}{4pt}
1299:
1300: \tablecaption{Fitting Parameters for Shallow--Silicon SNe~Ia of the
1301: Early Sample}
1302:
1303: \tablehead{\colhead{Parameter} &
1304: \colhead{SN~1999aa} & \colhead{SN~1999ac} & \colhead{SN~1991T} }
1305:
1306: \startdata
1307:
1308: \vphot\ (\kms) & 16,000 & 17,000 & 14,000 \\
1309:
1310: \ve\ (\kms) & 1000 & 1000 & 1000 \\
1311:
1312: $\tau$(C~II) & \nodata & 0.2/19 & \nodata \\
1313:
1314: $\tau$(C~III) & 0.7 & \nodata & 0.5 \\
1315:
1316: $\tau$(O~I) & 0.3/17 & 1 & \nodata \\
1317:
1318: $\tau$(Si~II) & 0.4/17 & 5 & 0.2 \\
1319:
1320: $\tau$(Si~III) & 0.6 & 2.5 & 2 \\
1321:
1322: $\tau$(S~II) & 0.5 & 1.2 & 0.5 \\
1323:
1324: $\tau$(Ca~II) & 1 & 4(2) & 1 \\
1325:
1326: $\tau$(HV Ca~II) & 7/25 & 1.5(5)/20 & \nodata \\
1327:
1328: $\tau$(HV Fe~II) & 0.5/22 & \nodata & \nodata \\
1329:
1330: $\tau$(Fe~III) & 2 & 1.5 & 2 \\
1331:
1332: \enddata
1333: \tablenotetext{a}{For column descriptions see the note to Table~2.}
1334: \end{deluxetable}
1335:
1336: \end{document}
1337: