1: %% Beginning of file 'snid.tex'
2: %%
3: %% This is a sample manuscript marked up using the
4: %% AASTeX v5.x LaTeX 2e macros.
5:
6: %\documentclass[12pt,preprint]{aastex}
7: \documentclass[numberedappendix]{emulateapj}
8:
9: %% manuscript produces a one-column, double-spaced document:
10: %% \documentclass[manuscript]{aastex}
11:
12: %% preprint2 produces a double-column, single-spaced document:
13: %% \documentclass[preprint2]{aastex}
14:
15: %% packages
16: \usepackage{graphicx}
17:
18: %% new commands
19: \def\kms{\hbox{$~$km$~$s$^{-1}$}}
20: \def\l{\ifmmode\lambda\else$\lambda$\fi}
21: \def\snia{SN~Ia}
22: \def\sneia{SNe~Ia}
23: \newcommand{\beq}{\begin{equation}}
24: \newcommand{\eeq}{\end{equation}}
25: \newcommand{\bea}{\begin{eqnarray}}
26: \newcommand{\eea}{\end{eqnarray}}
27:
28: %% You can insert a short comment on the title page using the command below.
29: \slugcomment{Accepted for publication in Ap.~J.}
30: \shorttitle{Time Dilation in SN~Ia Spectra}
31: \shortauthors{Blondin et al.}
32:
33:
34: \begin{document}
35:
36: %Title of paper
37: \title{Time Dilation in Type I\lowercase{a} Supernova Spectra
38: at High Redshift\footnotemark[1]}
39:
40: \footnotetext[1]{\vspace{0.00cm}Based on observations made with ESO
41: Telescopes at the Paranal Observatory under programs 67.A-0361,
42: 267.A-5688, 078.D-0383, and 080.D-0477; at the Gemini Observatory and
43: NOAO, which are operated by the Association of Universities for
44: Research in Astronomy, Inc., under cooperative agreements with the
45: NSF; with the Magellan Telescopes at Las Campanas Observatory; and at the
46: W.~M.~Keck Observatory, which was made possible by the generous
47: financial support of the W.~M.~Keck Foundation.}
48:
49: \author{
50: S.~Blondin,\altaffilmark{2}
51: T.~M.~Davis,\altaffilmark{3,4}
52: K.~Krisciunas,\altaffilmark{5}
53: B.~P.~Schmidt,\altaffilmark{6}
54: J.~Sollerman,\altaffilmark{3}
55: W.~M.~Wood-Vasey,\altaffilmark{2}
56: A.~C.~Becker,\altaffilmark{7} \\
57: P.~Challis,\altaffilmark{2}
58: A.~Clocchiatti,\altaffilmark{8}
59: G.~Damke,\altaffilmark{9}
60: A.~V.~Filippenko,\altaffilmark{10}
61: R.~J.~Foley,\altaffilmark{10}
62: P.~M.~Garnavich,\altaffilmark{11}
63: S.~W.~Jha,\altaffilmark{12} \\
64: R.~P.~Kirshner,\altaffilmark{2}
65: B.~Leibundgut,\altaffilmark{13}
66: W.~Li,\altaffilmark{10}
67: T.~Matheson,\altaffilmark{14}
68: G.~Miknaitis,\altaffilmark{15}
69: G.~Narayan,\altaffilmark{16}
70: G.~Pignata,\altaffilmark{17} \\
71: A.~Rest,\altaffilmark{9,16}
72: A.~G.~Riess,\altaffilmark{18,19}
73: J.~M.~Silverman,\altaffilmark{10}
74: R.~C.~Smith,\altaffilmark{9}
75: J.~Spyromilio,\altaffilmark{13}
76: M.~Stritzinger,\altaffilmark{3,20} \\
77: C.~W.~Stubbs,\altaffilmark{2,16}
78: N.~B.~Suntzeff,\altaffilmark{5}
79: J.~L.~Tonry,\altaffilmark{21}
80: B.~E.~Tucker,\altaffilmark{6}
81: and A.~Zenteno\altaffilmark{22}
82: }
83:
84: \altaffiltext{2}{Harvard-Smithsonian Center for Astrophysics, 60
85: Garden Street, Cambridge, MA 02138; {sblondin@cfa.harvard.edu} }
86: \altaffiltext{3}{Dark Cosmology Centre, Niels Bohr Institute, University of Copenhagen, Juliane Maries Vej 30, DK--2100 Copenhagen~\O, Denmark.}
87: \altaffiltext{4}{Department of Physics, University of Queensland, QLD, 4072, Australia.}
88: \altaffiltext{5}{Department of Physics, Texas A\&M University, College Station, TX 77843-4242.}
89: \altaffiltext{6}{The Research School of Astronomy and Astrophysics, The Australian National University, Mount Stromlo and Siding Spring Observatories, via Cotter Road, Weston Creek, PO 2611, Australia.}
90: \altaffiltext{7}{Department of Astronomy, University of Washington, Box 351580, Seattle, WA 98195-1580.}
91: \altaffiltext{8}{Pontificia Universidad Cat\'olica de Chile, Departamento de Astronom\'ia y Astrof\'isica, Casilla 306, Santiago 22, Chile.}
92: \altaffiltext{9}{Cerro Tololo Inter-American Observatory, National Optical Astronomy Observatory, Casilla 603, La Serena, Chile.}
93: \altaffiltext{10}{Department of Astronomy, University of California, Berkeley, CA 94720-3411.}
94: \altaffiltext{11}{Department of Physics, University of Notre Dame, 225 Nieuwland Science Hall, Notre Dame, IN 46556-5670.}
95: \altaffiltext{12}{Department of Physics and Astronomy, Rutgers University, 136 Frelinghuysen Road, Piscataway, New Jersey 08854.}
96: \altaffiltext{13}{European Southern Observatory, Karl-Schwarzschild-Strasse 2, D-85748 Garching, Germany.}
97: \altaffiltext{14}{National Optical Astronomy Observatory, 950 North Cherry Avenue, Tucson, AZ 85719-4933.}
98: \altaffiltext{15}{Fermilab, P.O. Box 500, Batavia, IL 60510-0500.}
99: \altaffiltext{16}{Department of Physics, Harvard University, 17 Oxford Street, Cambridge, MA 02138.}
100: \altaffiltext{17}{Departamento de Astronom{\'{\i}}a, Universidad de Chile, Casilla 36-D, Santiago, Chile.}
101: \altaffiltext{18}{Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218.}
102: \altaffiltext{19}{Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218.}
103: \altaffiltext{20}{Las Campanas Observatory, Carnegie Observatories, Casilla 601, La Serena, Chile.}
104: \altaffiltext{21}{Institute for Astronomy, University of Hawaii, 2680 Woodlawn Drive, Honolulu, HI 96822.}
105: \altaffiltext{22}{Department of Astronomy, University of Illinois at Urbana-Champaign, 1002 West Green St, Urbana, IL 61801-3080.}
106:
107:
108:
109:
110:
111: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
112: %%
113: %% Abstract
114: %%
115: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
116:
117: \begin{abstract}
118: We present multiepoch spectra of 13 high-redshift Type Ia supernovae
119: (\sneia) drawn from the literature, the ESSENCE and SNLS projects, and
120: our own separate dedicated program on the ESO Very Large Telescope. We
121: use the Supernova Identification (SNID) code of Blondin \& Tonry to
122: determine the spectral ages in the supernova rest frame. Comparison
123: with the observed elapsed time yields an apparent aging rate
124: consistent with the $1/(1+z)$ factor (where $z$ is the redshift)
125: expected in a homogeneous, isotropic, expanding universe. These
126: measurements thus confirm the expansion hypothesis, while
127: unambiguously excluding models that predict no time dilation, such as
128: Zwicky's ``tired light'' hypothesis.
129: We also test for power-law dependencies of the aging rate on
130: redshift. The best-fit exponent for these models is consistent with the
131: expected $1/(1+z)$ factor.
132: \end{abstract}
133:
134:
135: \keywords{cosmology: miscellaneous --- supernovae: general}
136:
137:
138:
139:
140:
141: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
142: %%
143: %% Introduction
144: %%
145: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
146:
147: \section{Introduction}\label{sect:intro}
148:
149: The redshift, $z$, is a fundamental observational quantity in
150: Friedman-Lema{\^{\i}}tre-Robertson-Walker (FLRW) models of the
151: universe. It relates the frequency of light emitted from a distant
152: source to that detected by a local observer by a factor $1/(1+z)$. One
153: important consequence is that the observed
154: rate of any time variation in the intensity of emitted radiation will
155: also be proportional to $1/(1+z)$ (see \citealt{Weinberg:1972} and
156: Appendix~\ref{sect:math}).
157:
158: Due to their large luminosities (several billion times that of the sun)
159: and variability on short timescales ($\sim20$~d from explosion
160: to peak luminosity; \citealt{Riess/etal:1999c,Conley/etal:2006b}), Type
161: Ia supernovae
162: (\sneia) are ideally suited to probe these time-dilation effects
163: across a large fraction of the observable universe. The suggestion to
164: use supernovae as ``cosmic clocks'' was proposed by Wilson more than six
165: decades ago \citep{Wilson:1939} and tested on light curves of
166: low-redshift \sneia\ in the mid-1970s \citep{Rust:1974}, but only
167: since the mid-1990s has this effect been unambiguously detected in the light
168: curves of high-redshift objects \citep{SN1995K,Goldhaber/etal:2001}.
169:
170: These latter studies show that the light curves of distant \sneia\ are
171: consistent with those of nearby \sneia\ whose time axis is dilated by
172: a factor $(1+z)$. However, there exists an intrinsic variation in the
173: width of \snia\ light curves which is related to their peak
174: luminosities \citep{Phillips:1993}, such that more luminous \sneia\
175: have broader light curves (Fig.~\ref{fig:lcbol}). This
176: ``width-luminosity'' relation is derived using low-redshift \sneia\
177: where the time-dilation effect, if any, is negligible
178: \citep{Phillips:1993,Hamuy/etal:1995,Riess/Press/Kirshner:1995,Phillips/etal:1999,MLCS2k2}.
179:
180: %%% Fig. 1
181: \begin{figure}
182: \epsscale{1.1}
183: \plotone{f1.eps}
184: \caption{\label{fig:lcbol}
185: Bolometric light
186: curves of 5 low-redshift \sneia\ taken from \cite{Stritzinger:2005}
187: (from top to bottom: SNe~1991T, 1999ee, 1994D, 1992A, 1993H, and
188: 1991bg). More luminous \sneia\ have broader light curves.
189: SN~1991bg is an example of intrinsically subluminous \sneia\
190: (maximum $L_{\rm bol} < 10^9 L_{\rm sun}$), which are less likely to
191: be found at high redshifts.}
192: \end{figure}
193:
194: It is problematic to disentangle this intrinsic variation of
195: light-curve width with luminosity and the effect of time dilation.
196: To directly test the time-dilation hypothesis one needs to accurately
197: know the distribution of light-curve widths at $z\approx0$ and its
198: potential evolution with redshift, whether due to a selection effect
199: (not taken into account by \citealt{Goldhaber/etal:2001}) or an evolution
200: of the mean properties of the \snia\ sample with redshift--- as
201: possibly observed by \cite{Howell/etal:2007}. Moreover, one needs to
202: probe sufficiently high redshifts ($z\gtrsim0.4$, as done by
203: \citealt{SN1995K,Goldhaber/etal:2001}) such that the observed widths of
204: the \snia\ light curves are broader than the intrinsic width of any
205: nearby counterpart.
206:
207: Furthermore, one might argue that
208: at high redshift we are preferentially finding the
209: brighter events (akin to a Malmquist bias).
210: Such a selection effect would produce a spurious relation in
211: which there would be broader light curves at higher redshifts,
212: without any time dilation.
213:
214: The spectra of \sneia\ provide an alternative and a more
215: reliable way to measure the apparent aging rate of distant
216: objects. Indeed, the spectra of \sneia\ are remarkably homogeneous at a
217: given age, such that the age of a \snia\ can be determined from a
218: single spectrum with an accuracy of 1--3 d--- with no reference to
219: its corresponding light curve \citep{Riess/etal:1997a,Howell/etal:2005,SNID}.
220: More importantly, the spectra of \sneia\ spanning a range of luminosities
221: (and hence different intrinsic light-curve widths) evolve uniformly
222: over time \citep{Matheson/etal:2008}.
223: The use of spectra thus avoids
224: the degeneracy between intrinsic light-curve width and time-dilation
225: effects. While there are some notable examples of deviations from
226: homogeneity in several \sneia\ (e.g., SN~2000cx, \citealt{Li/etal:2001b};
227: SN~2002cx, \citealt{Li/etal:2003}; SN~2002ic, \citealt{Hamuy/etal:2003};
228: SN~2003fg, \citealt{03D3bb}; SN~2006gz, \citealt{SN2006gz}), these
229: outliers are readily identifiable spectroscopically through
230: comparison with a large database of supernova spectra (see
231: Section~\ref{sect:age} and \citealt{SNID}).
232:
233: As of today there are two published examples of aging rate
234: measurements using spectra of a single \snia\ (SN~1996bj
235: at $z=0.574$, \citealt{Riess/etal:1997a}; SN~1997ex at $z=0.362$,
236: \citealt{Foley/etal:2005}). In both cases, the null hypothesis of no time
237: dilation is excluded with high significance ($>95\%$).
238:
239: In this paper we present data on 13 high-redshift ($0.28 \le z \le
240: 0.62$) \sneia\ for which we have multiepoch spectra. We
241: use the Supernova Identification (SNID) code of \citet{SNID} to infer
242: the age of each spectrum in the supernova rest frame, from which we
243: determine the apparent aging rate of each \snia. These aging rate
244: measurements are then used to test the $1/(1+z)$ time-dilation
245: hypothesis expected in an expanding universe. The data enable us for
246: the first time to directly test the time-dilation hypothesis {\it over
247: a large redshift range}.
248:
249: This paper is organized as follows.
250: In Section~\ref{sect:age} we explain how one can determine the age
251: of a \snia\ based on a single spectrum and present the SNID algorithm
252: used for this purpose. The aging rate measurements are presented in
253: Section~\ref{sect:aging}, and the time-dilation hypothesis (amongst
254: others) is tested against the data in Section~\ref{sect:tdil}, with
255: the help of model selection statistics (information
256: criteria). Conclusions follow in Section~\ref{sect:ccl}.
257:
258:
259:
260: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
261: %%
262: %% Age
263: %%
264: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
265:
266: \section{Determining the Age of a SN~I\lowercase{a} Spectrum}\label{sect:age}
267:
268: %-----------------------------------------------------------------
269: \subsection{SN~I\lowercase{a} Spectral Evolution}
270:
271: The spectra of \sneia\ consist of blended spectral
272: lines, with a profile shape characteristic of stellar
273: outflows. This line profile (also known as a ``P Cygni'' profile)
274: consists of an emission component symmetric about the line center, and
275: an absorption component that is blueshifted by the $\sim10,000$\,\kms\
276: expansion velocity of the SN ejecta \citep{Pinto/Eastman:2000b}. The
277: expansion also causes a Doppler broadening of both components, such
278: that a typical spectroscopic feature in \snia\ spectra has a width of
279: $\sim100$\,\AA.
280: As the ejecta expand, the photosphere recedes in the comoving frame
281: of the supernova, such that the spectra probe deeper layers
282: of the ejecta with time. Given the homologous nature of the
283: expansion (velocity proportional to radius), and the chemical
284: stratification in the SN ejecta
285: \citep{W7,Stehle/etal:2005,Mazzali/etal:2008}, deeper layers
286: correspond to lower expansion velocities and an increased
287: abundance of iron-peak elements. The impact on the spectra is
288: twofold. First, the blueshift of \snia\ spectral lines decreases with
289: time (by as much as $\sim1000$\,\kms\ per day;
290: \citealt{Benetti/etal:2005,Blondin/etal:2006}). Second, due to the
291: varying chemical composition at the photosphere, the relative
292: shapes and strengths of spectral features change on a timescale of
293: days.
294:
295: This complex spectral evolution is nonetheless predictable to a large
296: extent. At a given age, the spectra are remarkably homogeneous among
297: different ``normal'' \sneia. According to \cite{Li/etal:2001a}, these
298: constitute $\sim65$\% of the local \snia\ sample, the rest consisting of
299: intrinsically subluminous ($\sim15$\%) or overluminous ($\sim20\%$)
300: events, whose spectra show deviations from those of normal \sneia. Subluminous
301: \sneia\ are less likely to be found at high redshifts; in fact, no
302: such object has been spectroscopically confirmed in any high-redshift
303: supernova search to this day
304: \citep[e.g.,][]{Matheson/etal:2005,Howell/etal:2005}. In what follows
305: we only consider normal and overluminous \sneia. In a separate paper
306: we will show that none of the \sneia\ in the high-redshift sample
307: presented here (see Section~\ref{sect:tdil}) has a spectrum or light
308: curve consistent with the subluminous variety of \sneia.
309:
310: The spectroscopic homogeneity of \sneia\ holds even when we consider
311: both normal and overluminous objects in a representative sample of
312: nearby events. In Fig.~\ref{fig:meanspec} we show the mean spectrum
313: for the 22 low-redshift \sneia\ for which we report an aging rate
314: measurement (see Section~\ref{sect:tdil}), at four different
315: ages. While there is an intrinsic spectral variance among these
316: different \sneia\ --- some spectroscopic features correlate with
317: luminosity (e.g., \citealt{Nugent/etal:1995,Matheson/etal:2008}), the
318: average deviations from the mean spectrum are small, and all spectra
319: evolve in a similar manner over the course of several days, {\it
320: independent of light-curve width}.
321:
322: %%% Fig. 2
323: \begin{figure}
324: \epsscale{1.1}
325: \plotone{f2.eps}
326: \caption{\label{fig:meanspec}
327: Standard (light gray) and maximum (dark gray) deviation from the mean
328: spectrum (black line) for the 22 low-redshift \sneia\ for which we
329: report an aging rate measurement (see Section~\ref{sect:tdil}), at
330: four different ages --- given in days from $B$-band maximum light. A
331: low-order curve has been divided out from each spectrum to reveal the
332: relative shapes and strengths of the various spectroscopic features.
333: }
334: \end{figure}
335:
336: Both the homogeneity and rapid evolution of \snia\ spectra enables an
337: accurate determination of the age of a single spectrum. We explain how
338: this is achieved in practice in the following section.
339:
340:
341: %-----------------------------------------------------------------
342: \subsection{\label{sect:snid}The SNID Algorithm}
343:
344: Given a large database of finely time-sampled \snia\ spectral
345: templates, we can determine the age of a given input spectrum by
346: finding the best-match template(s) in the database. There are several
347: standard techniques to do this (see \citealt{SNID} for a
348: review). In this paper, we use an implementation of the correlation
349: techniques of \cite{Tonry/Davis:1979}: SNID \citep{SNID}.
350: SNID automatically determines
351: the type, redshift, and age of a supernova spectrum. We refer the
352: reader to that paper for a more detailed discussion.
353:
354: The redshift of the input spectrum is a free parameter in SNID,
355: although it can be fixed to a specific value.
356: Comparison of the SNID redshifts with those determined from
357: narrow emission and absorption lines in the host-galaxy
358: spectra (typically accurate to $\lesssim100$\,\kms; see \citealt{UZC})
359: yields a dispersion about the one-to-one correspondence of only
360: $\sigma_z \approx 0.005$ out to a redshift $z \approx 0.8$ \citep{SNID}.
361:
362: Similarly, comparison of the SNID ages with those determined using the
363: corresponding light curves yields a typical accuracy $<3$ d,
364: comparable to other algorithms \citep{Riess/etal:1997a,Howell/etal:2005}.
365: However, the age error is systematically overestimated. In this
366: paper, we estimate the error as follows: each spectrum in the SNID
367: database is trimmed to match the rest-frame wavelength range of the
368: input spectrum, and is correlated with all other spectra in the
369: database (except those corresponding to the same supernova). The age
370: error is then given by the mean variance of all template spectra whose
371: SNID age is within one day of the initial estimate.
372:
373: The success of SNID and similar algorithms lies primarily in the
374: completeness of the spectral database. In Fig.~\ref{fig:thist} we
375: show the age distribution of the \snia\ templates used in SNID for
376: this paper (these do not include subluminous \sneia). This database
377: comprises 959 spectra of 79 low-redshift ($z \lesssim 0.05$) \sneia\
378: with ages between $-15$ and $+50$~d from maximum light. The spectra
379: are taken from the literature, from public databases (such as
380: SUSPECT\footnote{http://bruford.nhn.ou.edu/$\sim$suspect/index1.html}
381: or the CfA Supernova
382: Archive\footnote{http://www.cfa.harvard.edu/supernova/SNarchive.html}),
383: or from a set of unpublished spectra from the CfA Supernova Program. A
384: full reference to all spectra in the SNID database is given by
385: \cite{SNID}. It is important to note that each template
386: spectrum is shifted to zero redshift and that each template age is
387: corrected for the expected $(1+z)$ time-dilation factor. Because all
388: the template \sneia\ are at low redshift ($z \lesssim 0.05$), this is
389: a very small correction, and we will see in Section~\ref{sect:tdil}
390: that this has no impact on the aging rate measurements.
391: Thus, SNID determines the age a supernova would have at $z=0$ --- that is,
392: {\it in the supernova rest frame}.
393:
394: %%% Fig. 3
395: \begin{figure}
396: \epsscale{1.1}
397: \plotone{f3.eps}
398: \caption{\label{fig:thist}
399: Age distribution of the 79 \snia\ templates used in SNID (hatched
400: histogram). There are a total of 959 spectra with ages between $-15$
401: and $+50$~d from $B$-band maximum light. The open histogram shows
402: the subsample of 145 spectra from 22 \sneia\ (restricted to ages
403: between $-10$ and $30$~d from maximum) for which we report an aging
404: rate measurement (see Section~\ref{sect:tdil}).}
405: \end{figure}
406:
407: The number of \sneia\ shown in Fig.~\ref{fig:thist} is
408: large enough that we can select a subsample (shown as an
409: open histogram) on which to conduct age determinations and aging rate
410: measurements on low-redshift \sneia. The size of this subsample
411: is set by the requirement that removing it from the SNID database
412: would leave a sufficient number of templates in a given age bin for a
413: reliable age determination (see \citealt{SNID}). It was also chosen to
414: include a sufficient number of intrinsically overluminous \sneia:
415: indeed, there are five such \sneia\ (SNe~1997br, 1998ab, 1999dq,
416: 1999gp, and 2001eh) in this subsample, accounting for $\sim20$\% by
417: number of objects and spectra. For these specific tests, the templates
418: corresponding to the 22 \sneia\ in this subsample are temporarily
419: removed from the SNID database to avoid biasing the age determination.
420:
421: We deliberately restrict this subsample to ages between
422: $-10$ and +30~d from maximum light. Before $-10$ d, the number
423: of spectral templates in the SNID database drops rapidly, and the age
424: determination is inaccurate. Past +30 d, the spectra of \sneia\ do
425: not evolve as rapidly as around maximum light, and the age
426: determination is less precise \citep{SNID}.
427:
428:
429:
430: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
431: %%
432: %% Aging rate
433: %%
434: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
435:
436: \section{Aging Rate at High Redshifts}\label{sect:aging}
437:
438:
439: %-----------------------------------------------------------------
440: \subsection{Spectroscopic Data}
441:
442: Our aging rate measurements at high redshifts are based on a sample of
443: 35 spectra of 13 \sneia\ in the redshift range $0.28 \le z \le
444: 0.62$. These include previously published data by the High-$Z$
445: Supernova Search Team (SN~1996bj, \citealt{Riess/etal:1997a}), the Supernova
446: Cosmology Project (SN~1997ex, \citealt{Foley/etal:2005}; SN~2001go,
447: \citealt{Lidman/etal:2005}), and the ESSENCE project (SNe~2002iz, b027, and
448: 2003js, \citealt{Matheson/etal:2005}). For SN~2001go we present our own
449: reductions of the three epochs of spectroscopic data obtained from
450: the ESO Science Archive
451: Facility\footnote{http://www.eso.org/sci/archive/ .}, as only the first
452: spectrum was published by \cite{Lidman/etal:2005}. The spectra of
453: SN~04D2an (the highest-redshift \snia\ in this sample) were obtained
454: by members of the Supernova Legacy Survey (SNLS), and will be
455: published as part of a larger sample of SNLS data by St\'ephane Basa and
456: coworkers. SN~2006tk will be published alongside the complete ESSENCE
457: supernova dataset in the near future.
458:
459: The other five \sneia\ (SNe~2006mk, 2006sc, 2007tg, 2007tt, 2007un) are
460: ESSENCE targets which were observed spectroscopically through two
461: dedicated ``Target-of-Opportunity'' programs at the ESO Very Large
462: Telescope\footnote{Programs 078.D-0383 and 080.D-0477; PI: Jesper
463: Sollerman.}.
464:
465: Details on the instrumental setup and data reduction can
466: be found in the aforementioned references. The rest of the data will
467: be presented more thoroughly in a separate publication. All of the data
468: are shown in Fig.~\ref{fig:spec}.
469:
470: %%% Fig. 4
471: \begin{figure*}
472: \epsscale{1.1}
473: \plotone{f4.eps}
474: \caption{\label{fig:spec}Multiepoch spectra of the 13 high-redshift \sneia\ used in
475: this study, binned to 10\,\AA\ (gray). The vertical offset
476: between each spectrum is for clarity only, and does not reflect
477: differences in flux density ($F_\lambda$;
478: erg$~$s$^{-1}$cm$^{-2}$\AA$^{-1}$) between them. In each plot, the
479: age of the
480: supernova increases downwards, and the observed time (in days) from
481: the first spectrum is indicated. Overplotted in black is a smoothed
482: version using the inverse-variance-weighted Gaussian algorithm of
483: \citet{Blondin/etal:2006}.}
484: \end{figure*}
485:
486: We use SNID to determine the redshift of each spectrum. The redshift
487: of a given supernova ($z_{\rm SN}$) is then
488: computed as the error-weighted mean of the SNID redshifts ($z_{\rm
489: SNID}$) for each of its spectra (Table~\ref{table:zsnid}). For ten
490: \sneia, we also have a redshift determination from the host
491: galaxy ($z_{\rm GAL}$). Comparison with the supernova redshift shows
492: an excellent agreement between the two measurements (better than
493: 1\%). For the three remaining \sneia\ (SNe~b027, 2006tk, and 2007tg), only
494: the SNID redshift is available, but the
495: different redshift measurements for individual spectra all agree to
496: within 1$\sigma$, and we are confident about their accuracy. In what
497: follows, we will use the galaxy redshift when available for the age
498: and aging rate measurements. Given the excellent agreement between
499: $z_{\rm GAL}$ and $z_{\rm SN}$, this choice has negligible impact on
500: our results.
501:
502: %%% Table 1
503: \begin{deluxetable}{lccc}
504: %\tabletypesize{\scriptsize}
505: \tablewidth{0pt}
506: \tablecaption{\label{table:zsnid}Comparison of galaxy and supernova redshifts}
507: \tablehead{
508: \colhead{SN} &
509: \colhead{$z_{\rm GAL}$} &
510: \colhead{$z_{\rm SN}$} &
511: \colhead{$z_{\rm SNID}$} \\
512: \colhead{(1)} &
513: \colhead{(2)} &
514: \colhead{(3)} &
515: \colhead{(4)}
516: }
517: \startdata
518: 1996bj & 0.574 & 0.581 (0.005) & 0.580 (0.008) \\
519: & & & 0.582 (0.008) \\
520: \hline
521: 1997ex & 0.361 & 0.362 (0.002) & 0.362 (0.005) \\
522: & & & 0.361 (0.004) \\
523: & & & 0.362 (0.004) \\
524: \hline
525: 2001go & 0.552 & 0.552 (0.005) & 0.552 (0.008) \\
526: & & & 0.556 (0.008) \\
527: & & & 0.550 (0.009) \\
528: \hline
529: 2002iz & 0.427 & 0.425 (0.004) & 0.422 (0.006) \\
530: & & & 0.428 (0.006) \\
531: \hline
532: b027 & \nodata & 0.315 (0.003) & 0.315 (0.006) \\
533: & & & 0.315 (0.004) \\
534: \hline
535: 2003js & 0.363 & 0.361 (0.003) & 0.359 (0.004) \\
536: & & & 0.363 (0.006) \\
537: \hline
538: 04D2an & 0.621 & 0.614 (0.006) & 0.608 (0.007) \\
539: & & & 0.625 (0.011) \\
540: \hline
541: 2006mk & 0.475 & 0.477 (0.003) & 0.479 (0.005) \\
542: & & & 0.478 (0.007) \\
543: & & & 0.474 (0.008) \\
544: & & & 0.476 (0.006) \\
545: \hline
546: 2006sc & 0.357 & 0.356 (0.004) & 0.355 (0.007) \\
547: & & & 0.357 (0.007) \\
548: & & & 0.356 (0.006) \\
549: \hline
550: 2006tk & \nodata & 0.312 (0.003) & 0.312 (0.006) \\
551: & & & 0.310 (0.003) \\
552: & & & 0.315 (0.006) \\
553: \hline
554: 2007tg & \nodata & 0.502 (0.004) & 0.503 (0.009) \\
555: & & & 0.503 (0.008) \\
556: & & & 0.502 (0.007) \\
557: \hline
558: 2007tt & 0.374 & 0.376 (0.004) & 0.367 (0.008) \\
559: & & & 0.379 (0.007) \\
560: & & & 0.377 (0.005) \\
561: \hline
562: 2007un & 0.283 & 0.285 (0.004) & 0.287 (0.006) \\
563: & & & 0.285 (0.007) \\
564: & & & 0.285 (0.005)
565: \enddata
566: \tablenotetext{\ }{{\it Column headings:}
567: (1) SN name.
568: (2) Galaxy redshift (the typical error is $<0.001$).
569: (3) SN redshift, quoted as the error-weighted mean of the individual
570: redshifts for each epoch.
571: (4) SNID redshift for each epoch, in order of increasing age.
572: }
573: \end{deluxetable}
574:
575:
576:
577: %-----------------------------------------------------------------
578: \subsection{\label{sect:tdiff}Accuracy of Relative Age Determination}
579:
580: An accurate determination of the rate of aging involves accurate
581: knowledge of age {\it differences}. In what follows we test how well
582: SNID determines differential ages using the subsample of
583: low-redshift \sneia\ presented in Section~\ref{sect:snid}. While
584: the determination of {\it absolute} ages has no impact on the
585: main result of this paper, we discuss their accuracy in
586: Appendix~\ref{sect:tcomp}.
587:
588: We determine the rest-frame age ($t_{\rm spec}$) of each of the 145
589: spectra in the low-redshift subsample of 22 \sneia. We then compute
590: the absolute age difference ($\Delta t_{\rm spec}$) between each
591: unique pair of spectra corresponding to a given supernova. This
592: amounts to 631 pairs. This age difference is then compared with the
593: absolute observer-frame age difference ($\Delta t_{\rm obs}$) for each
594: spectrum pair. Since $z \approx 0$ for this subsample, $\Delta
595: t_{\rm spec}$ can be directly compared to $\Delta t_{\rm obs}$ with
596: no correction for time dilation. Given the restriction to ages between
597: $-10$ and +30~d from maximum light in the low-redshift subsample,
598: $\Delta t_{\rm spec}$ (and hence $\Delta t_{\rm obs}$) is at most 40 d.
599:
600: The results are displayed in the upper panel of
601: Fig.~\ref{fig:tdiff}. There is good agreement between $\Delta t_{\rm
602: spec}$ and $\Delta t_{\rm obs}$, with a dispersion of only 2.0 d
603: about the one-to-one correspondence. For $\Delta t_{\rm obs} \gtrsim
604: 30$ d, however, SNID systematically underestimates the age
605: difference by $\sim 1.5$ d. This is more apparent in the plot of
606: residuals in the middle panel. it is mainly due to a systematic
607: overestimate of rest-frame ages $t_{\rm spec}\lesssim-7$~d from
608: maximum light, due to the lack of spectral templates in the SNID
609: database with similar ages (see Fig.~\ref{fig:thist} and
610: \citealt{SNID}).
611:
612: %%% Fig. 5
613: \begin{figure}
614: \epsscale{1.1}
615: \plotone{f5.eps}
616: \caption{\label{fig:tdiff}
617: Upper panel: Rest-frame age difference ($\Delta t_{\rm spec}$)
618: vs. observer-frame age difference ($\Delta t_{\rm obs}$) for each
619: spectrum pair for a given supernova. There are 631 such pairs, with a
620: dispersion of 2.0~d about the one-to-one correspondence.
621: Middle panel: Residuals in the upper panel vs. $\Delta t_{\rm obs}$.
622: Lower panel: Ratio of $\Delta t_{\rm spec}$ to $\Delta t_{\rm obs}$,
623: again vs. $\Delta t_{\rm obs}$. For $\Delta t_{\rm obs} > 6$ d, the
624: fractional difference is less than 20\%.
625: }
626: \end{figure}
627:
628: The lower panel of Fig.~\ref{fig:tdiff} shows the absolute fractional
629: age difference vs. $\Delta t_{\rm obs}$. The quantity $|\Delta t_{\rm
630: obs}/\Delta t_{\rm spec}|$ is a direct measure of the accuracy we can
631: achieve for the aging rate determination. As expected, the fractional
632: age difference decreases with increasing age difference. For $\Delta
633: t_{\rm obs} > 6$ d, this difference drops below 20\%. The
634: high-redshift data presented in the previous section span a sufficient
635: range of observer-frame age difference that the aging rate
636: determination is accurate. Note that the systematic underestimate of
637: the age difference for $\Delta t_{\rm obs} \gtrsim 30$~d results in
638: a negligible fractional difference.
639:
640:
641: %-----------------------------------------------------------------
642: \subsection{\label{sect:trate}Aging Rate Determination}
643:
644: The rest-frame age of each high-redshift \snia\ spectrum ($t_{\rm
645: spec}$) is determined as outlined in Section~\ref{sect:snid}. In each
646: case, we fix the redshift to that determined in the previous section.
647: The results are displayed in Table~\ref{table:tdiff}, along with the
648: corresponding observed date of each spectrum ($t_{\rm obs}$). However,
649: since the aging rate determination depends on age {\it differences} (see
650: previous Section), we also report the observer-frame and rest-frame
651: age from the first spectrum, respectively denoted $\Delta t_{\rm obs}$
652: and $\Delta t_{\rm spec}$ in Table~\ref{table:tdiff}.
653:
654: We can then trivially compute the aging rate for each supernova. This
655: is simply done through a least-squares fit of a line to $\Delta t_{\rm
656: spec}$ versus $\Delta t_{\rm obs}$. The slope of the line is a
657: measure of the aging rate, which should equal $1/(1+z)$ in an
658: expanding universe (see Appendix~\ref{sect:math}). Were there no time
659: dilation, the aging rate would equal one. The results are displayed in
660: Fig.~\ref{fig:tdilfig}.
661:
662: %%% Table 2
663: % SNID ages and differences
664: \begin{deluxetable}{lrrrr}
665: %\tabletypesize{\scriptsize}
666: \tablewidth{0pt}
667: \tablecaption{\label{table:tdiff}Observer-frame and rest-frame age differences}
668: \tablehead{
669: \colhead{SN} &
670: \colhead{$t_{\rm obs}$} &
671: \colhead{$t_{\rm spec}$} &
672: \colhead{$\Delta t_{\rm obs}$} &
673: \colhead{$\Delta t_{\rm spec}$} \\
674: \colhead{(1)} &
675: \colhead{(2)} &
676: \colhead{(3)} &
677: \colhead{(4)} &
678: \colhead{(5)}
679: }
680: \startdata
681: 1996bj & 367.99 &$-$2.2 (3.0) & 0.00 & 0.0 (3.0) \\
682: & 378.04 & 3.1 (2.2) & 10.05 & 5.3 (2.2) \\
683: \hline
684: 1997ex & 815.08 &$-$1.6 (1.6) & 0.00 & 0.0 (1.6) \\
685: & 839.96 & 17.4 (2.1) & 24.88 & 19.0 (2.1) \\
686: & 846.03 & 21.2 (2.0) & 30.95 & 22.8 (2.0) \\
687: \hline
688: 2001go & 2021.70 & 7.9 (2.3) & 0.00 & 0.0 (2.3) \\
689: & 2027.69 & 9.8 (1.7) & 5.99 & 1.9 (1.7) \\
690: & 2059.17 & 31.2 (1.6) & 37.47 & 23.3 (1.6) \\
691: \hline
692: 2002iz & 2586.95 &$-$0.5 (2.2) & 0.00 & 0.0 (2.2) \\
693: & 2614.57 & 17.6 (1.2) & 27.62 & 18.1 (1.2) \\
694: \hline
695: b027 & 2589.95 &$-$3.5 (1.8) & 0.00 & 0.0 (1.8) \\
696: & 2616.57 & 18.4 (1.6) & 26.62 & 21.9 (1.6) \\
697: \hline
698: 2003js & 2942.46 &$-$4.9 (1.6) & 0.00 & 0.0 (1.6) \\
699: & 2966.71 & 12.5 (1.2) & 24.25 & 17.4 (1.2) \\
700: \hline
701: 04D2an & 3026.20 &$-$2.5 (1.6) & 0.00 & 0.0 (1.6) \\
702: & 3032.20 & 0.9 (1.3) & 6.00 & 3.4 (1.3) \\
703: \hline
704: 2006mk & 4031.71 &$-$6.2 (1.0) & 0.00 & 0.0 (1.0) \\
705: & 4040.72 &$-$0.6 (2.2) & 9.01 & 5.6 (2.2) \\
706: & 4051.77 & 7.3 (1.9) & 20.06 & 13.5 (1.9) \\
707: & 4063.78 & 18.5 (1.8) & 32.07 & 24.7 (1.8) \\
708: \hline
709: 2006sc & 4063.58 & 0.9 (1.6) & 0.00 & 0.0 (1.6) \\
710: & 4076.65 & 9.8 (1.4) & 13.07 & 8.9 (1.4) \\
711: & 4084.68 & 13.4 (2.2) & 21.10 & 12.5 (2.2) \\
712: \hline
713: 2006tk & 4089.57 &$-$8.8 (2.4) & 0.00 & 0.0 (2.4) \\
714: & 4100.57 & 0.3 (2.0) & 11.00 & 9.1 (2.0) \\
715: & 4103.59 & 2.9 (0.9) & 14.02 & 11.7 (0.9) \\
716: \hline
717: 2007tg & 4381.75 &$-$6.1 (2.0) & 0.00 & 0.0 (2.0) \\
718: & 4391.65 &$-$0.5 (1.8) & 9.90 & 5.6 (1.8) \\
719: & 4405.61 & 10.0 (1.5) & 23.86 & 16.1 (1.5) \\
720: \hline
721: 2007tt & 4415.81 &$-$5.0 (2.1) & 0.00 & 0.0 (2.1) \\
722: & 4430.65 & 6.1 (1.4) & 14.84 & 11.1 (1.4) \\
723: & 4443.58 & 14.9 (2.2) & 27.77 & 19.9 (2.2) \\
724: \hline
725: 2007un & 4441.61 & 3.2 (2.2) & 0.00 & 0.0 (2.2) \\
726: & 4451.60 & 11.1 (1.3) & 9.99 & 7.9 (1.3) \\
727: & 4460.58 & 17.7 (1.4) & 18.97 & 14.5 (1.4)
728: \enddata
729: \tablenotetext{\ }{{\it Column headings:}
730: (1) SN name.
731: (2) Julian date ($JD$) minus 2,450,000 at midpoint of observation.
732: (3) SN rest-frame age in days from maximum light, derived from the
733: cross-correlation with spectral templates using SNID.
734: (4) Observer-frame days from first spectrum.
735: (5) Rest-frame days from first spectrum.
736: }
737: \end{deluxetable}
738:
739: %%% Fig. 6
740: \begin{figure*}
741: \epsscale{1.1}
742: \plotone{f6.eps}
743: \caption{\label{fig:tdilfig}
744: Comparison of rest-frame ($\Delta t_{\rm spec}$) and observer-frame
745: ($\Delta t_{\rm obs}$) time from the first spectrum, for each of the
746: 13 high-redshift \sneia\ in our sample. The abscissa and ordinate
747: ranges are both set to [$-$3,+40]~d in all cases. The slope of
748: the best-fit line (solid line) gives a measurement of the
749: apparent aging rate of the supernova, which is compared to the
750: expected $1/(1+z)$ value. The dotted line in each plot
751: corresponds to $\Delta t_{\rm spec}=\Delta t_{\rm obs}$.
752: }
753: \end{figure*}
754:
755: We note that comparing the inverse of the slope in
756: Fig.~\ref{fig:tdilfig} (denoted ``age factor'' by
757: \citealt{Foley/etal:2005}) and $(1+z)$ leads to asymmetric errors.
758: The errors on the ``age
759: factor'' [$\equiv (1+z)$] become highly non-Gaussian when the uncertainties of
760: the individual age measurements are large ($\gtrsim 1$ d, as is the
761: case in this paper, and in \citealt{Foley/etal:2005} for SN 1997ex), whereas the errors
762: on the aging rate [$\equiv 1/(1+z)$] are always Gaussian. This is
763: illustrated in Fig.~\ref{fig:slopemc} using a Monte Carlo simulation
764: of the age measurements for SN~1997ex presented by
765: \cite{Foley/etal:2005}. Using the same errors on the individual age
766: measurements, the distribution of the slope measurements is highly
767: non-Gaussian in $(1+z)$ space, while it is normally distributed in
768: $1/(1+z)$ space.
769:
770: %%% Fig. 7
771: \begin{figure}
772: \epsscale{1.1}
773: \plotone{f7.eps}
774: \caption{\label{fig:slopemc}
775: Monte Carlo results illustrating the advantage of working in
776: $1/(1+z)$ space for time-dilation measurements. Solid lines:
777: recovered slope (in standard deviations from the mean, $\mu$) in
778: $(1+z)$ (left) and $1/(1+z)$ space (right), using
779: the SN age errors reported by \cite{Foley/etal:2005}. The
780: distribution is highly non-Gaussian in the former case.
781: }
782: \end{figure}
783:
784: The individual aging rate measurements presented here alone reject
785: models that predict no time dilation at a high significance (up to
786: $\sim6\sigma$), and all (except for SN~2006mk) are within $1\sigma$ of
787: the expected $1/(1+z)$ factor. In the next section we combine all
788: aging rate measurements (including those for the low-redshift sample)
789: to test each hypothesis more thoroughly.
790:
791:
792:
793:
794: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
795: %%
796: %% Time-dilation
797: %%
798: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
799:
800: \section{Testing the $1/(1+\lowercase{z})$ Time-Dilation Hypothesis}\label{sect:tdil}
801:
802: We have determined the aging rate for the subsample of 22 low-redshift
803: \sneia. We combine these aging rates with those determined for the 13
804: high-redshift \sneia\ of our sample (see Table~\ref{table:tdil}) to
805: test the $1/(1+z)$ time-dilation hypothesis. As noted in
806: Section~\ref{sect:tdiff} and Appendix~\ref{sect:tcomp},
807: these measurements rely on a database of
808: \snia\ spectra whose ages have already been corrected for the expected
809: $1/(1+z)$ time-dilation factor. However, since all \sneia\ in the SNID
810: database are at redshifts $z\le0.05$, the correction is small
811: ($\lesssim 1$ d) and has a negligible impact on the aging rate
812: measurements.
813:
814: %%% Table 3
815: % tdil results
816: \begin{deluxetable}{lccc}
817: %\tabletypesize{\scriptsize}
818: \tablewidth{0pt}
819: \tablecaption{\label{table:tdil}Aging rate measurements}
820: \tablehead{
821: \colhead{SN} &
822: \colhead{$z$} &
823: \colhead{$1/(1+z)$} &
824: \colhead{Aging rate}
825: }
826: \startdata
827: \multicolumn{4}{c}{Low redshift ($z < 0.04$)} \\
828: \hline
829: 1981B & 0.006 & 0.994 & 1.099 (0.071) \\
830: 1989B & 0.002 & 0.998 & 1.036 (0.079) \\
831: 1992A & 0.006 & 0.994 & 1.040 (0.077) \\
832: 1994D & 0.002 & 0.998 & 1.011 (0.152) \\
833: 1996X & 0.007 & 0.993 & 0.886 (0.116) \\
834: 1997br & 0.007 & 0.993 & 1.120 (0.149) \\
835: 1998V & 0.018 & 0.983 & 1.047 (0.166) \\
836: 1998ab & 0.027 & 0.974 & 0.987 (0.151) \\
837: 1998dm & 0.007 & 0.993 & 0.778 (0.119) \\
838: 1998eg & 0.025 & 0.976 & 0.883 (0.085) \\
839: 1999cl & 0.008 & 0.992 & 0.870 (0.122) \\
840: 1999dq & 0.014 & 0.986 & 0.928 (0.055) \\
841: 1999ej & 0.014 & 0.986 & 1.147 (0.159) \\
842: 1999gp & 0.027 & 0.974 & 0.839 (0.060) \\
843: 2000fa & 0.021 & 0.979 & 0.884 (0.076) \\
844: 2001V & 0.015 & 0.985 & 0.973 (0.072) \\
845: 2001eh & 0.037 & 0.964 & 1.148 (0.107) \\
846: 2001ep & 0.013 & 0.987 & 0.925 (0.086) \\
847: 2002ha & 0.014 & 0.986 & 0.973 (0.134) \\
848: 2003cg & 0.004 & 0.996 & 1.006 (0.131) \\
849: 2003du & 0.006 & 0.994 & 0.909 (0.091) \\
850: 2006lf & 0.013 & 0.987 & 0.941 (0.110) \\
851: \hline
852: \multicolumn{4}{c}{High redshift ($z > 0.2$)} \\
853: \hline
854: 1996bj & 0.574 & 0.635 & 0.527 (0.369) \\
855: 1997ex & 0.361 & 0.735 & 0.745 (0.076) \\
856: 2001go & 0.552 & 0.644 & 0.652 (0.062) \\
857: 2002iz & 0.427 & 0.701 & 0.655 (0.089) \\
858: b027 & 0.315 & 0.760 & 0.823 (0.092) \\
859: 2003js & 0.363 & 0.734 & 0.718 (0.082) \\
860: 04D2an & 0.621 & 0.617 & 0.567 (0.341) \\
861: 2006mk & 0.475 & 0.678 & 0.753 (0.060) \\
862: 2006sc & 0.357 & 0.737 & 0.619 (0.121) \\
863: 2006tk & 0.312 & 0.762 & 0.835 (0.181) \\
864: 2007tg & 0.502 & 0.666 & 0.687 (0.102) \\
865: 2007tt & 0.374 & 0.728 & 0.718 (0.108) \\
866: 2007un & 0.283 & 0.779 & 0.759 (0.135)
867: \enddata
868: \end{deluxetable}
869:
870: All aging rate measurements are shown in Fig.~\ref{fig:tdil}. The
871: solid line shows the expected $1/(1+z)$ time-dilation factor, while
872: the dashed line represents the ``tired light'' hypothesis of
873: \citet{Zwicky:1929}. According to this hypothesis, photons lose energy
874: as they interact with matter and other photons in a static
875: universe. The energy loss is proportional to the distance from the
876: source, and causes a redshift in spectra as in an expanding
877: universe. However, this hypothesis does not predict a
878: time-dilation effect, and so the aging rate should equal one for all
879: redshifts.
880:
881: %%% Fig. 8
882: \begin{figure*}
883: \epsscale{1.1}
884: \plotone{f8.eps}
885: \caption{\label{fig:tdil}
886: Apparent aging rate versus $1/(1+z)$ for the 13 high-redshift ($0.28 \le
887: z \le 0.62$) and 22 low-redshift ($z < 0.04$) \sneia\ in our
888: sample. Overplotted are the expected $1/(1+z)$ time dilation (solid
889: line) and the best-fit $1/(1+z)^b$ model (with $b=0.97\pm0.10$; dotted
890: line and gray area). The dashed line corresponds to no time dilation,
891: as expected in the tired-light model --- clearly inconsistent with the
892: data. The inset shows a close-up view of the low-redshift
893: sample. These data are summarized in Table~\ref{table:tdil}.
894: }
895: \end{figure*}
896:
897: As expected, the measurement of a time-dilation effect is more obvious
898: at larger redshift, and the precision improves as the number and time
899: span of spectra for each supernova increases. This latter effect
900: explains why the aging rate measurements for SN~1996bj ($z=0.574$) and
901: SN~04D2an ($z=0.621$) have a large associated error --- despite being
902: the two highest-redshift \sneia\ in our sample, since
903: only two spectra separated by $\sim10$ (for SN~1996bj) and $\sim6$
904: (for SN~04D2an) observer-frame days are available.
905:
906: A simple $\chi^2$ analysis is sufficient to confirm what the eye sees:
907: the hypothesis of no time dilation is not a good fit to the data
908: ($\chi^2=150.3$ for 35 degrees of freedom; see
909: Table~\ref{table:modelcomp}), with a goodness-of-fit of $\sim0\%$
910: (defined as GoF$=\Gamma(\nu/2,\chi^2/2)/\Gamma(\nu/2)$, where
911: $\Gamma(\nu/2,\chi^2/2)$ is the incomplete gamma function and $\nu$ is
912: the number of degrees of freedom) --- namely, a null probability of
913: obtaining data that are a worse fit to the model, assuming that the
914: model is indeed correct. The expected $1/(1+z)$ time-dilation factor,
915: on the other hand, yields a good fit to the data ($\chi^2=27.0$
916: for 35 degrees of freedom), with GoF$=83.2\%$, and is largely
917: favored over the null hypothesis of no time dilation ($\Delta \chi^2
918: \approx 123$).
919:
920: %%% Table 4
921: % model comparisons
922: \begin{deluxetable*}{lrrrrrrrrr}
923: %\tabletypesize{\scriptsize}
924: \tablewidth{0pt}
925: \tablecaption{\label{table:modelcomp}Time-dilation model comparison}
926: \tablehead{
927: &
928: \multicolumn{3}{c}{All SNe} &
929: \multicolumn{3}{c}{High-redshift SNe only} &
930: \multicolumn{3}{c}{Low-redshift SNe only} \\
931: \cline{2-4} \cline{5-7} \cline{8-10}
932: Model\tablenotemark{a} &
933: $\chi^2$/dof & GoF & $\Delta$AIC &
934: $\chi^2$/dof & GoF & $\Delta$AIC &
935: $\chi^2$/dof & GoF & $\Delta$AIC \\
936: & & (\%) & & & (\%) & & & (\%) &
937: }
938: \startdata
939: $1/(1+z)$ & 27.0/35 & 83.2 & 0 & 3.6/13 & 99.5 & 0 & 23.4/22 & 38.2 & 0 \\
940: $1/(1+z)^b$ & 26.9/34 & 80.2 & 1 & 3.4/12 & 99.2 & 1 & 20.3/21 & 50.0 &$-$1\\
941: tired light & 150.3/35 & 0.0 & 123 & 123.4/13 & 0.0 & 119 & 26.9/22 & 21.4 & 3
942: \enddata
943: \tablenotetext{a}{The best-fit values for the $b$ exponent in the
944: second model are as follows.
945: All SNe: $b=0.97\pm0.10$;
946: high-redshift SNe only: $b=0.95\pm0.10$;
947: low-redshift SNe only: $b=3.18\pm1.28$}
948: \end{deluxetable*}
949:
950: This result holds (and in fact improves)
951: when we consider only the high-redshift sample (see
952: Table~\ref{table:modelcomp}).
953: This works because the $z=0$ end of the aging rate versus redshift
954: relation (Fig.~\ref{fig:tdil}) is fixed to unity by theory, so the
955: low-redshift sample is not needed to anchor the theoretical curve at
956: $z\approx 0$ (although it is still used to calibrate the $t_{\rm spec}$
957: measurement).
958: The low-redshift data alone do not enable us to distinguish between
959: the two hypotheses, since the impact of time dilation is small at such
960: low redshifts.
961:
962: In Fig.~\ref{fig:tdil_ratio} a different view of Fig.~\ref{fig:tdil}
963: shows the distributions of the ratio between the aging rate and
964: $1/(1+z)$ for both the low-redshift (open histogram) and high-redshift
965: (hatched histogram) samples. Both distributions are within $\sim20\%$
966: of a unit ratio, again validating the hypothesis of time dilation over
967: a large redshift range. The apparent bias to lower values of the
968: ratio for the low-redshift sample is not statistically significant, as
969: the mean error on the aging rate is of order one bin size
970: ($\lesssim 0.1$; see Table~\ref{table:tdil}).
971:
972: %%% Fig. 9
973: \begin{figure}
974: \epsscale{1.1}
975: \plotone{f9.eps}
976: \caption{\label{fig:tdil_ratio}
977: Ratio of the aging rate to $1/(1+z)$ for all \sneia\ in
978: Fig.~\ref{fig:tdil}. Both the low-redshift (open histogram) and
979: high-redshift (hatched histogram) samples are shown.
980: }
981: \end{figure}
982:
983: In what follows we test whether the data favor a nonlinear dependence
984: of the aging rate on redshift, namely
985:
986: \begin{equation}
987: \label{eqn:power}
988: {\rm aging\ rate} = \frac{1}{(1+z)^b},
989: \end{equation}
990:
991: \noindent
992: where $b$ is a free parameter. While Eq.~[\ref{eqn:power}]
993: satisfies the same zero point as the two previous hypotheses (aging
994: rate equal to 1 at $z=0$), no model actually predicts such a
995: dependence of the aging rate on redshift. Nonetheless, small
996: deviations from the expected $1/(1+z)$ factor would have profound
997: implications for our assumption of FLRW cosmology.
998:
999: Again, we performed a least-squares fit to the entire sample, and also
1000: to the individual high- and low-redshift samples (see
1001: Table~\ref{table:modelcomp}).
1002: The data constrain the $b$ exponent to 10\%
1003: ($1\sigma$), and yield $b=0.97\pm0.10$ for the entire sample
1004: (Fig.~\ref{fig:tdil}; dotted line and gray region) and
1005: $b=0.95\pm0.10$ for the high-redshift sample. As expected, the
1006: low-redshift sample alone is insufficient to constrain the free
1007: parameter ($b=3.18\pm1.28$). Nonetheless, the samples that include the
1008: high-redshift objects have a best-fit value for $b$ that is consistent
1009: with $b=1$, and thus with the expected $1/(1+z)$ time-dilation
1010: factor.
1011:
1012: Since this model has an additional free parameter, it is
1013: instructive to use information criteria to compare it to the simple
1014: $1/(1+z)$ prediction.
1015: These model comparison statistics favor models that yield a good fit
1016: to the data with fewer parameters.
1017: As in \cite{Davis/etal:2007}, we use the
1018: Akaike information criterion (AIC; \citealt{AIC}). For Gaussian errors
1019: (which is the case here, see Section~\ref{sect:trate}),
1020: this criterion can be expressed as
1021:
1022: \begin{equation}
1023: \label{eqn:aic}
1024: {\rm AIC} = \chi^2 + 2k,
1025: \end{equation}
1026:
1027: \noindent
1028: where $k$ is the number of free parameters
1029: \citep{Davis/etal:2007}. Comparison of
1030: models simply involves computing the difference in
1031: AIC ($\Delta$AIC) with respect to the model with the lowest value for
1032: this criterion. A difference in
1033: AIC of 2 is considered positive evidence against the model with the
1034: higher AIC, whereas a difference of 6 is considered strong evidence
1035: \citep{Liddle:2004,Davis/etal:2007}. In the models considered here (see
1036: Table~\ref{table:modelcomp}), the expected $1/(1+z)$ time-dilation
1037: model has the lowest
1038: AIC (although this is not true for the low-redshift sample), and we
1039: compute AIC differences with respect to that model.
1040:
1041: With $\Delta {\rm AIC}=1$, we conclude that the information
1042: criteria do not provide positive evidence against a $1/(1+z)^b$
1043: dependence of the aging rate. The $\chi^2$ per degree of freedom is
1044: also satisfactory for the samples that include the high-redshift
1045: objects.
1046:
1047: The two other models considered previously have no free parameters
1048: ($k=0$ in Eq.~[\ref{eqn:aic}]), hence $\Delta {\rm
1049: AIC}=\Delta\chi^2$, and the information criterion is reduced to a
1050: simple $\chi^2$ test.
1051:
1052:
1053:
1054: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1055: %%
1056: %% Conclusion
1057: %%
1058: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1059:
1060: \section{Conclusion}\label{sect:ccl}
1061:
1062: We have presented 35 spectra of 13 high-redshift ($0.28 \le z \le
1063: 0.62$) \sneia, which include previously unpublished data from the
1064: ESSENCE and SNLS projects and from our own dedicated program at the
1065: ESO Very Large Telescope. Given the rapid and predictable evolution of
1066: \snia\ spectral features with age, as well as the relative homogeneity
1067: of \snia\ spectra at a given age, one is able to determine the
1068: (rest-frame) age of a single spectrum with a typical accuracy of 1--3
1069: d
1070: \citep{Riess/etal:1997a,Foley/etal:2005,Hook/etal:2005,Howell/etal:2005,SNID}.
1071:
1072: Using the Supernova Identification (SNID) code of \citet{SNID}, we
1073: determine the ages of each spectrum in the supernova
1074: rest frame. Comparison with the observed time difference between the
1075: spectra yields an apparent aging rate consistent
1076: with $1/(1+z)$, as expected in a homogeneous and isotropic expanding
1077: universe.
1078: Moreover, the data unambiguously rule out the ``tired
1079: light'' hypothesis \citep{Zwicky:1929} in which photons lose energy as
1080: they interact with matter and other photons in a static universe.
1081:
1082: The fact that the age determination is so accurate over
1083: a large redshift range also shows that the deviations between
1084: spectra of low- and high-redshift \sneia\ in our sample are small.
1085:
1086: We also test for alternate dependencies of the
1087: aging rate on redshift, namely $1/(1+z)^b$, although these are not
1088: predicted by any model. Whether we consider the entire sample or only
1089: the high-redshift sample, the best-fit value for the $b$
1090: exponent is consistent with $b=1$, and thus with the expected
1091: $1/(1+z)$ factor.
1092:
1093: That these data provide a confirmation of the time-dilation factor
1094: expected in an expanding universe should be of no
1095: surprise. Nonetheless, previous use of \snia\ light curves to test
1096: this hypothesis \citep{SN1995K,Goldhaber/etal:2001} are
1097: prone to the spread in intrinsic light-curve widths and its
1098: possible variation with redshift (which includes selection effects;
1099: see Section~\ref{sect:intro}).
1100:
1101: The data presented here are unique in that they enable the most direct
1102: test of the $1/(1+z)$ time-dilation hypothesis over a larger redshift
1103: range than has yet been performed. This hypothesis is favored
1104: beyond doubt over models that predict no time dilation. With more
1105: data, the focus will shift to testing more thoroughly the alternative
1106: $1/(1+z)^b$ dependence of the aging rate on redshift. Any significant
1107: deviation from $b=1$ would have a profound impact on our assumption of
1108: a FLRW cosmology to describe the universal expansion.
1109:
1110:
1111:
1112: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1113: %%
1114: %% Acknowledgments
1115: %%
1116: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1117:
1118: \begin{acknowledgments}
1119: The authors wish to thank the Supernova Legacy Survey (SNLS)
1120: collaboration, and in particular St\'ephane Basa and Tianmeng Zhang,
1121: for providing spectra of supernova 04D2an prior to publication.
1122: The ESSENCE survey is supported by the US National Science Foundation
1123: (NSF) through grants AST 04-43378 and AST 05-07475. Support for
1124: supernova research at Harvard University, including the CfA Supernova
1125: Archive, is provided in part by NSF
1126: grant AST 06-06772. The Dark Cosmology Centre is funded by the Danish
1127: National Research Foundation.
1128: A.~C. acknowledges the support of CONICYT, Chile, under grants FONDECYT
1129: 1051061 and FONDAP 15010003.
1130: T.~M.~D. appreciates the support of the Villum Kann Rasmussen Fonden.
1131: A.~V.~F. is grateful for the support of NSF grant AST 06-07894.
1132: G.~P. acknowledges the support of the Proyecto FONDECYT 3070034.
1133: \end{acknowledgments}
1134:
1135: {\it Facilities:}
1136: \facility{VLT:Kueyen (FORS1)},
1137: \facility{VLT:Antu (FORS2)},
1138: \facility{Gemini:South (GMOS)},
1139: \facility{Gemini:Gillete (GMOS)},
1140: \facility{Keck:I (LRIS)},
1141: \facility{Keck:II (DEIMOS, ESI)},
1142: \facility{Magellan:Baade (IMACS)},
1143: \facility{Magellan:Clay (LDSS3)}.
1144:
1145:
1146:
1147: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1148: %%
1149: %% Appendix A
1150: %%
1151: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1152:
1153: \begin{appendix}
1154: \section{Time Dilation in an Expanding Universe}\label{sect:math}
1155:
1156: In a homogeneous, isotropic expanding universe, the interval $d\tau$
1157: between two space-time events is given by the Robertson-Walker (RW)
1158: metric
1159: \citep{Robertson:1935,Robertson:1936a,Robertson:1936b,Walker:1936},
1160:
1161: \begin{equation}
1162: \label{eqn:flrw}
1163: d\tau^2 = c^2dt^2 - a^2(t)
1164: \left[ \frac{dr^2}{1-kr^2} + r^2(d\theta^2+\sin^2\theta d\phi^2) \right],
1165: \end{equation}
1166:
1167: \noindent
1168: where $c$ is the speed of light in vacuum, $t$ is the cosmic time,
1169: $(r,\theta,\phi)$ are the comoving spatial coordinates, $k$ is the
1170: curvature parameter, and $a(t)$ is the dimensionless scale factor. In
1171: what follows we assume the present-day value of the dimensionless
1172: scale factor $a_0 = 1$.
1173:
1174: Photons travel along null geodesics ($d\tau^2=0$). In what follows we
1175: consider radial null rays only ($d\theta=d\phi=0$). For a photon
1176: emitted at time $t_1$ from an object located at
1177: $(r_1,\theta_1,\phi_1)$ and observed at time $t_0$,
1178: Eq.~[\ref{eqn:flrw}] implies
1179:
1180: \begin{equation}
1181: \label{eqn:t}
1182: \int_{t_1}^{t_0} \frac{cdt}{a(t)} = \int_0^{r_1} \frac{dr}{\sqrt{1-kr^2}}
1183: \equiv f(r_1).
1184: \end{equation}
1185:
1186: \noindent
1187: Here we assume that the object from which the photon was emitted
1188: has constant coordinates $(r_1,\theta_1,\phi_1)$ such that $f(r_1)$,
1189: also known as the comoving distance, is time independent. Thus, for a
1190: photon emitted at time $t_1+\delta t_1$ and observed at time
1191: $t_0+\delta t_0$, Eq.~[\ref{eqn:flrw}] also implies
1192:
1193: \begin{equation}
1194: \label{eqn:dt}
1195: \int_{t_1+\delta t_1}^{t_0 + \delta t_0} \frac{cdt}{a(t)} = f(r_1).
1196: \end{equation}
1197:
1198: \noindent
1199: For small $\delta t_1$ (and hence small $\delta t_0$), the rate of
1200: change of the scale factor remains roughly constant and
1201: Eqs.~[\ref{eqn:t}] and [\ref{eqn:dt}] imply
1202:
1203: \begin{equation}
1204: \frac{\delta t_0}{a_0} = \frac{\delta t_1}{a(t_1)}.
1205: \end{equation}
1206:
1207: \noindent
1208: Hence, a light signal emitted with frequency $\nu_1$ will reach us with
1209: frequency $\nu_0$ such that
1210:
1211: \begin{equation}
1212: \frac{\nu_0}{\nu_1} = \frac{\delta t_1}{\delta t_0} = \frac{a(t_1)}{a_0}.
1213: \end{equation}
1214:
1215: Using the standard definition of redshift, $z = (\l_0-\l_1)/\l_1 =
1216: \nu_1/\nu_0 - 1$, we obtain a relationship between observed and
1217: rest-frame time intervals in a RW metric as a function of redshift $z$:
1218:
1219: \begin{equation}
1220: \label{eqn:dtfrac}
1221: \frac{\delta t_0}{\delta t_1} = 1+z.
1222: \end{equation}
1223:
1224: \noindent
1225: A supernova at redshift $z$ will thus appear to age $(1+z)$ times more
1226: slowly with respect to a local event at $z \approx 0$.
1227:
1228: The prediction of time dilation proportional to $(1+z)$ is generic to
1229: expanding universe models, whether the underlying theory be general
1230: relativity (e.g., the Friedmann-Lema\^itre-Robertson-Walker universe),
1231: special relativity (e.g., the Milne Universe), or Newtonian expansion.
1232: A point of confusion can occur in the special relativistic case for
1233: which the well-known time-dilation factor is given by
1234:
1235: \bea
1236: \gamma_{\rm SR} & = & \left[1-\left(\frac{v}{c}\right)^2\right]^{-1/2} \\
1237: & = & \frac{1}{2} \left(1+z+\frac{1}{1+z}\right),
1238: \label{eq:SRtimedilation}
1239: \eea
1240:
1241: \noindent
1242: which evidently differs from $(1+z)$. Thus it might be assumed that a
1243: special relativistic expansion can be distinguished from the FLRW
1244: universe using a time-dilation test\footnote{In fact, such an
1245: erroneous assumption was made by one of the current authors of
1246: \cite{Davis/Lineweaver:2004}.}.
1247:
1248: This is not the case.
1249: Special relativistic expansion of the
1250: universe assumes there is an inertial frame that extends to infinity
1251: (impossible in the non-empty
1252: general relativistic picture) and that the expansion involves objects
1253: moving through this inertial frame. The time-dilation factor from
1254: Eq.~[\ref{eq:SRtimedilation}] relates the proper time in the moving
1255: emitter's inertial frame ($\delta t_1$) to the proper time in the
1256: observer's inertial frame ($\delta t_0$). To measure this time
1257: dilation the observer has to set up a set of synchronized clocks (each
1258: at rest in the observer's inertial frame) and take readings of the
1259: emitter's proper time as the emitter moves past each synchronized
1260: clock. The readings show that the emitter's clock is time dilated
1261: such that $\delta t_0 = \gamma_{\rm SR}\delta t_1$.
1262:
1263: We do not have this set of synchronized clocks at our disposal when
1264: we measure time dilation of supernovae in an expanding universe
1265: and therefore Eq.~[\ref{eq:SRtimedilation}] is not the time dilation we
1266: observe. We must also take into account an extra time-dilation factor
1267: that occurs because the distance to the emitter (and thus the distance
1268: light has to propagate to reach us) is increasing. In the time
1269: $\delta t_0$ the emitter moves a distance $v\delta t_0$ away from us.
1270: The total proper time we observe, $\delta t_{0,\rm tot}$, is $\delta
1271: t_0$ plus an extra factor describing how long light takes to traverse
1272: this extra distance ($v \delta t_0 /c$),
1273:
1274: \beq
1275: \delta t_{0,\rm tot} = \delta t_0 (1+v/c).
1276: \eeq
1277:
1278: \noindent
1279: The relationship between proper time at the emitter and proper time at
1280: the observer is thus
1281:
1282: \bea
1283: \delta t_{0,\rm tot} &=& \gamma_{\rm SR} \delta t_1 (1+v/c)\\
1284: &=& \delta t_1 \sqrt{\frac{1+v/c}{1-v/c}} \\
1285: &=& \delta t_1(1+z),
1286: \eea
1287:
1288: \noindent
1289: which is identical to the GR time-dilation equation.
1290:
1291: Non-cosmological redshifts (i.e., not due to universal expansion) also
1292: cause a time-dilation effect described by
1293: Eq.~[\ref{eqn:dtfrac}]. However, these additional effects from
1294: peculiar velocities and gravitational redshifts contribute random
1295: error only, and do not bias the measurements presented here.
1296:
1297:
1298:
1299: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1300: %%
1301: %% Appendix B
1302: %%
1303: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1304:
1305: %-----------------------------------------------------------------
1306: \section{Comparison of Spectral and Light-Curve Ages}\label{sect:tcomp}
1307:
1308: To test the accuracy of the age determination using SNID, we select
1309: the \sneia\ for which a well-sampled light curve is available around
1310: maximum light. Only the ESSENCE and SNLS \sneia\ in our sample have
1311: associated light curves for which we could determine the date of
1312: maximum brightness ($t_{\rm max}$). To do so we used the MLCS2k2 light-curve
1313: fitting code of \citet{MLCS2k2}, as done by
1314: \cite{Wood-Vasey/etal:2007}. This way we can determine the time
1315: difference (in the {\it observer} frame) between maximum light
1316: ($t_{\rm max}$) and the time the spectrum was obtained ($t_{\rm obs}$).
1317: We compare this time interval with the {\it rest-frame} age
1318: determined through cross-correlation with local \snia\ spectral
1319: templates using SNID ($t_{\rm spec}$). We expect a one-to-one
1320: correspondence between
1321:
1322: \begin{equation}
1323: t_{\rm LC} = \frac{t_{\rm obs}-t_{\rm max}}{1+z}
1324: \end{equation}
1325:
1326: \noindent
1327: and $t_{\rm spec}$.
1328:
1329: The result is shown as black points in Fig.~\ref{fig:tcomp}. While the
1330: agreement is good, there is a mean systematic offset of $-1.6$ d
1331: between $t_{\rm spec}$ and $t_{\rm LC}$, as shown in the middle
1332: panel. If this offset were to affect only a subset of age measurements
1333: for a given supernova, the impact on the aging rate determination
1334: would be severe. To check this, we correct the spectral ages of a
1335: given supernova for the mean difference between $t_{\rm LC}$ and
1336: $t_{\rm spec}$. This ``corrected'' age residual, $\Delta t_{\rm corr}$
1337: is plotted in the lower panel of Fig.~\ref{fig:tcomp}. The mean
1338: residual drops to $-0.1$~d and the scatter decreases slightly.
1339:
1340: %%% Fig. 10
1341: \begin{figure}
1342: \epsscale{0.8}
1343: \plotone{f10.eps}
1344: \caption{\label{fig:tcomp}
1345: Upper panel: comparison of supernova {\it rest-frame} ages (in days
1346: from maximum light) obtained from cross-correlation with spectral
1347: templates ($t_{\rm spec}$) and from fits to the light curve
1348: ($t_{\rm LC}$). 145 age measurements for the subsample
1349: of 22 low-redshift \sneia\ are shown in gray. The dashed line
1350: represents the one-to-one correspondence between $t_{\rm LC}$ and
1351: $t_{\rm spec}$.
1352: Middle panel: Age residuals, $\Delta t = t_{\rm spec} - t_{\rm
1353: LC}$. We also indicate the standard deviation ($\sigma$) and mean
1354: residual ($\mu$).
1355: Lower panel: Same as above, where each point has been corrected for
1356: the mean offset between $t_{\rm spec}$ and $t_{\rm LC}$ for a given
1357: supernova.
1358: }
1359: \end{figure}
1360:
1361: Since there are 2 to 4 $t_{\rm spec}$ measurements for a given
1362: supernova, and only one measurement of $t_{\rm max}$, the source of
1363: the discrepancy between the spectral and light-curve ages is most
1364: likely due to the determination of the date of maximum using the
1365: light-curve fitter. Indeed, using a different light-curve fitter
1366: (SALT2; \citealt{SALT2}) yields values for $t_{\rm max}$ that differ from
1367: the MLCS2k2 measurements by more than one day in 9 out of 10 cases,
1368: and by more than two days for three objects (SNe~2003js, 2007tg, and
1369: 2007un). These discrepancies are due to a combination of differences
1370: in light-curve fitter algorithms and data quality (light-curve
1371: sampling around maximum light and signal-to-noise ratio of each
1372: light-curve measurement; see \citealt{Miknaitis/etal:2007}).
1373:
1374: Therefore, while there is a systematic offset between part of these
1375: different age determinations, this offset affects all measurements in
1376: a similar fashion and has no impact on the determination of the {\it
1377: rate} of aging. In fact, the main result of this paper (see
1378: Section~\ref{sect:tdil}) is completely independent of $t_{\rm LC}$, and
1379: hence of $t_{\rm max}$. Nonetheless, the comparison between spectral
1380: and light-curve ages confirms the accuracy of age determination using
1381: spectra alone \citep{SNID}. The age measurements for all the
1382: high-redshift \sneia\ in our sample are reported in
1383: Table~\ref{table:tcomp}.
1384:
1385: %%% Table 5
1386: % SNID ages
1387: \begin{deluxetable}{lrrrr}
1388: %\tabletypesize{\scriptsize}
1389: \tablewidth{0pt}
1390: \tablecaption{\label{table:tcomp}Comparison of rest-frame light-curve and
1391: spectral ages}
1392: \tablehead{
1393: \colhead{SN} &
1394: \colhead{$t_{\rm LC}$} &
1395: \colhead{$t_{\rm spec}$} &
1396: \colhead{$\Delta t$} &
1397: \colhead{$\Delta t_{\rm corr}$} \\
1398: \colhead{(1)} &
1399: \colhead{(2)} &
1400: \colhead{(3)} &
1401: \colhead{(4)} &
1402: \colhead{(5)}
1403: }
1404: \startdata
1405: 2002iz & 0.1 (1.1) &$-$0.5 (2.2) &$-$0.6 (2.4) & 0.8 (1.5) \\
1406: & 19.4 (1.1) & 17.6 (1.2) &$-$1.8 (1.6) &$-$0.4 (2.3) \\
1407: \hline
1408: b027 & $-$2.4 (0.5) &$-$3.5 (1.8) &$-$1.1 (1.9) &$-$0.9 (1.7) \\
1409: & 17.9 (0.5) & 18.4 (1.6) & 0.5 (1.7) & 0.7 (1.4) \\
1410: \hline
1411: 2003js & $-$3.2 (0.3) &$-$4.9 (1.6) &$-$1.7 (1.6) & 0.3 (2.0) \\
1412: & 14.6 (0.3) & 12.5 (1.2) &$-$2.1 (1.2) &$-$0.1 (2.3) \\
1413: \hline
1414: 04D2an & $-$3.2 (0.9) &$-$2.5 (1.6) & 0.7 (1.8) & 0.2 (1.4) \\
1415: & 0.5 (0.9) & 0.9 (1.3) & 0.4 (1.6) &$-$0.1 (1.3) \\
1416: \hline
1417: 2006mk & $-$3.6 (0.7) &$-$6.2 (1.0) &$-$2.6 (1.3) &$-$0.5 (2.7) \\
1418: & 2.6 (0.7) &$-$0.6 (2.2) &$-$3.2 (2.3) &$-$1.1 (3.3) \\
1419: & 10.0 (0.7) & 7.3 (1.9) &$-$2.7 (2.0) &$-$0.6 (2.8) \\
1420: & 18.2 (0.7) & 18.5 (1.8) & 0.3 (1.9) & 2.4 (0.9) \\
1421: \hline
1422: 2006sc & 0.9 (0.5) & 0.9 (1.6) & 0.0 (1.7) & 0.9 (1.0) \\
1423: & 10.5 (0.5) & 9.8 (1.4) &$-$0.7 (1.5) & 0.2 (1.2) \\
1424: & 16.4 (0.5) & 13.4 (2.2) &$-$3.0 (2.3) &$-$2.1 (3.2) \\
1425: \hline
1426: 2006tk & $-$6.1 (0.5) &$-$8.8 (2.4) &$-$2.7 (2.5) &$-$0.8 (2.8) \\
1427: & 2.3 (0.5) & 0.3 (2.0) &$-$2.0 (2.1) &$-$0.1 (2.2) \\
1428: & 4.6 (0.5) & 2.9 (0.9) &$-$1.7 (1.0) & 0.2 (1.9) \\
1429: \hline
1430: 2007tg & $-$6.6 (0.9) &$-$6.1 (2.0) & 0.5 (2.2) & 0.2 (1.2) \\
1431: & 0.0 (0.9) &$-$0.5 (1.8) &$-$0.5 (2.0) &$-$0.8 (1.2) \\
1432: & 9.3 (0.9) & 10.0 (1.5) & 0.7 (1.8) & 0.4 (1.3) \\
1433: \hline
1434: 2007tt & $-$2.5 (0.6) &$-$5.0 (2.1) &$-$2.5 (2.2) &$-$0.1 (2.7) \\
1435: & 8.3 (0.6) & 6.1 (1.4) &$-$2.2 (1.6) & 0.2 (2.5) \\
1436: & 17.7 (0.6) & 14.9 (2.2) &$-$2.8 (2.2) &$-$0.4 (3.0) \\
1437: \hline
1438: 2007un & 4.3 (0.4) & 3.2 (2.2) &$-$1.1 (2.3) & 0.1 (1.4) \\
1439: & 12.1 (0.4) & 11.1 (1.3) &$-$1.0 (1.3) & 0.2 (1.4) \\
1440: & 19.1 (0.4) & 17.7 (1.4) &$-$1.4 (1.5) &$-$0.2 (1.7)
1441: \enddata
1442: \tablenotetext{\ }{{\it Column headings:}
1443: (1) SN name.
1444: (2) SN rest-frame age in days from maximum light, derived from the
1445: light curve.
1446: (3) SN rest-frame age in days from maximum light,
1447: derived from the cross-correlation with spectral templates using
1448: SNID.
1449: (4) $\Delta t = t_{\rm spec} - t_{\rm LC}$.
1450: (5) $\Delta t$ corrected for the mean offset between $t_{\rm spec}$
1451: and $t_{\rm LC}$.
1452: }
1453: \end{deluxetable}
1454:
1455: In making the comparison we have implicitly assumed what we are
1456: trying to show, namely a time-dilation factor of $(1+z)$. Accordingly, we
1457: also make the same comparison for our subsample of 22 low-redshift
1458: \sneia\ ($0.002 \le z \le 0.04$). At such low redshifts, the $(1+z)$
1459: correction present in $t_{\rm LC}$ is negligible (the mean correction
1460: is $\sim 0.06$ d). The result is shown as gray points in
1461: Fig.~\ref{fig:tcomp}. The mean residual between $t_{\rm LC}$ and
1462: $t_{\rm spec}$ for this low-redshift sample is close to zero with a
1463: small scatter ($\sigma \approx 1.5$ d), and unlike the
1464: high-redshift sample there is no significant systematic offset between
1465: the two age measurements.
1466:
1467: The age measurements presented in Table~\ref{table:tdiff} also
1468: enable us to infer the
1469: date of maximum light for each supernova using spectra alone
1470: (corresponding to $t_{\rm spec}=0$).
1471: This way we are able to determine the time of
1472: maximum for the \sneia\ in our sample for which a well-sampled
1473: light curve was unavailable (SNe~1996bj, 1997ex, and 2001go; see
1474: Table~\ref{table:tmax}). We can also compare the dates of maximum as
1475: inferred from a fit to the light curve ($t_{\rm max}^{\rm
1476: LC}$) with those determined from the spectra alone ($t_{\rm max}^{\rm
1477: spec}$). The results are also shown in Table~\ref{table:tmax}. For
1478: four objects (SNe~2003js, 2006mk, 2006tk, and 2007tt) the disagreement is
1479: larger than $1\sigma$, and explains the systematic negative offset
1480: between $t_{\rm spec}$ and $t_{\rm LC}$ seen in Fig.~\ref{fig:tcomp}.
1481:
1482: %%% Table 6
1483: % tmax comparison
1484: \begin{deluxetable}{lrrr}
1485: %\tabletypesize{\scriptsize}
1486: \tablewidth{0pt}
1487: \tablecaption{\label{table:tmax}Comparison of dates of maximum light}
1488: \tablehead{
1489: \colhead{SN} &
1490: \colhead{$t_{\rm max}^{\rm LC}$} &
1491: \colhead{$t_{\rm max}^{\rm spec}$} &
1492: \colhead{$\Delta t_{\rm max}$} \\
1493: \colhead{(1)} &
1494: \colhead{(2)} &
1495: \colhead{(3)} &
1496: \colhead{(4)}
1497: }
1498: \startdata
1499: 1996bj & \nodata & 372.16 (3.73) & \nodata \\
1500: 1997ex & \nodata & 817.16 (1.98) & \nodata \\
1501: 2001go & \nodata & 2011.47 (3.12) & \nodata \\
1502: 2002iz & 2586.83 (1.51) & 2587.71 (3.21) & 0.88 (3.54) \\
1503: b027 & 2593.09 (0.65) & 2594.20 (1.90) & 1.11 (2.01) \\
1504: 2003js & 2946.80 (0.47) & 2949.29 (1.67) & 2.49 (1.73) \\
1505: 04D2an & 3031.36 (1.50) & 3030.61 (1.82) & $-$0.75 (2.36) \\
1506: 2006mk & 4036.95 (0.96) & 4040.32 (1.02) & 3.37 (1.40) \\
1507: 2006sc & 4062.39 (0.64) & 4061.71 (2.77) & $-$0.68 (2.84) \\
1508: 2006tk & 4097.56 (0.66) & 4100.13 (0.98) & 2.57 (1.18) \\
1509: 2007tg & 4391.62 (1.34) & 4391.34 (1.58) & $-$0.28 (2.08) \\
1510: 2007tt & 4419.29 (0.83) & 4422.47 (1.83) & 3.18 (2.01) \\
1511: 2007un & 4436.12 (0.47) & 4437.15 (3.12) & 1.03 (3.16)
1512: \enddata
1513: \tablenotetext{\ }{{\it Column headings:}
1514: (1) SN name.
1515: (2) JD $-$ 2,450,000 of maximum light, derived from the light curve.
1516: (3) JD $-$ 2,450,000 of maximum light, derived from the spectra.
1517: (4) $\Delta t_{\rm max} = t_{\rm max}^{\rm spec} - t_{\rm max}^{\rm LC}$.
1518: }
1519: \end{deluxetable}
1520:
1521: \end{appendix}
1522:
1523:
1524: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1525: %%
1526: %% Bibliography
1527: %%
1528: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1529:
1530: % Create the reference section using BibTeX:
1531: \bibliographystyle{apj}
1532: \bibliography{ms}
1533:
1534:
1535:
1536: \end{document}
1537:
1538:
1539: % LocalWords: Blondin Centre Maries Vej DK Stromlo Leibundgut Schwarzschild de
1540: % LocalWords: Strasse Garching Sollerman Spyromilio Stritzinger Clocchiatti ia
1541: % LocalWords: Pontificia Universidad olica Departamento Astronom Astrof isica
1542: % LocalWords: Casilla Nieuwland Jha Kavli Krisciunas Matheson Miknaitis San ds
1543: % LocalWords: Fermilab Pignata Cerro Tololo Vasey Zenteno multiepoch ESO SNID
1544: % LocalWords: Tonry restframe APS RW dt dr kr eqn cdt comoving eqns aa MLCS bj
1545: % LocalWords: Overplotted timedilation ccccccccc IAU Lidman Exptime Riess LRIS
1546: % LocalWords: KeckII lidman VLT FORS iz ESI KeckI GMOS js mk sc Baade IMACS tk
1547: % LocalWords: lcccc lccrrrr lrrrr lccc br dm eg dq ej gp ep dp er du lf SNe ee
1548: % LocalWords: lrrrrrrrrrrrr dof GoF BIC AIC Campanas Frelinghuysen Piscataway
1549: % LocalWords: Zwicky's Bolometric bolometric Eq Eqs blueshifted Stehle Benetti
1550: % LocalWords: blueshift subluminous overluminous Supernova Identification CfA
1551: % LocalWords: redward Jesper lrrrrr Minkowski lrrr gauge Akaike pre Lema tre
1552: % LocalWords: FLWR Malmquist Goldhaber cx ic fg gz Damke Narayan SNLS itre tg
1553: % LocalWords: Friedmann FLRW Basa tt un Laboratoire d'Astrophysique Marseille
1554: % LocalWords: BP Cedex Tmax Organisation bg bol Cygni TDIFF lrrrrrrrrr AST QLD
1555: % LocalWords: MODELCOMP NOAO Paranal Astier ephane Tianmeng Zhang CONICYT Kann
1556: % LocalWords: FONDECYT FONDAP Villum Fonden Proyecto Hamuy bb Mazzali UZC Antu
1557: % LocalWords: Zwicky Liddle Kueyen Gillete LDSS eps UBVRI
1558: