1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4:
5: % PACKAGES TO USE
6: \usepackage{psfig}
7: \usepackage{amsmath}
8: \usepackage{amssymb}
9: \usepackage{graphicx}
10:
11:
12: % NEW COMMANDS
13: \newcommand{\MCcode}{\texttt{SEDONA}}
14: \newcommand{\kms}{\ensuremath{\mathrm{km~s}^{-1}}}
15: \newcommand{\msun}{\ensuremath{M_\odot}}
16: \newcommand{\zsun}{\ensuremath{Z_\odot}}
17: \newcommand{\texp}{\ensuremath{t_{\mathrm{exp}}}}
18: \newcommand{\Mni}{\ensuremath{M_{\mathrm{Ni}}}}
19: \newcommand{\Nifs}{\ensuremath{^{56}\mathrm{Ni}}}
20: \newcommand{\Cofs}{\ensuremath{^{56}\mathrm{Co}}}
21: \newcommand{\Fefs}{\ensuremath{^{56}\mathrm{Fe}}}
22: \newcommand{\Feff}{\ensuremath{^{54}\mathrm{Fe}}}
23: \newcommand{\Mch}{{\ensuremath{\mathrm{M}_\mathrm{ch}}}}
24: \newcommand{\dmf}{\ensuremath{\Delta M_{15}}}
25: \newcommand{\dmfb}{\ensuremath{\Delta M_{15}(B)}}
26: \newcommand{\dmfv}{\ensuremath{\Delta M_{15}(V)}}
27: \newcommand{\dmfbol}{\ensuremath{\Delta M_{15}(\mathrm{bol})}}
28: \newcommand{\dmfbolt}{\ensuremath{\Delta M_{30}(\mathrm{bol})}}
29: \newcommand{\dmfbt}{\ensuremath{\Delta M_{30}(B)}}
30: \newcommand{\dmfubv}{\ensuremath{\Delta M_{15}(UBVRI)}}
31: \newcommand{\Mfe}{\ensuremath{M_{\mathrm{Fe}}}}
32: \newcommand{\Msun}{{\ensuremath{\mathrm{M}_{\odot}}}}
33: \newcommand{\phiNi}{{\ensuremath{\phi_\mathrm{Ni}}}}
34: \newcommand{\NiCent}{\ensuremath{M^c_{\mathrm{Ni}}}}
35: \newcommand{\KE}{\ensuremath{E_K}}
36: \newcommand{\Mb}{\ensuremath{M_B}}
37: \newcommand{\Mv}{\ensuremath{M_V}}
38: \newcommand{\Mbol}{\ensuremath{M_{\rm bol}}}
39: \newcommand{\bmvmax}{\ensuremath{(B-V)_{\rm max}}}
40: \newcommand{\bmvf}{\ensuremath{(B-V)_{15}}}
41: \newcommand{\tbol}{\ensuremath{t_{\rm bol}}}
42: \newcommand{\tb}{\ensuremath{t_{\rm B}}}
43: \newcommand{\td}{\ensuremath{t_{\rm d}}}
44: \newcommand{\Bband}{\ensuremath{B}-band}
45: \newcommand{\bmv}{\ensuremath{B-V}}
46: \newcommand{\bmr}{\ensuremath{B-R}}
47:
48: \shortauthors{Kasen \& Woosley}
49: \shorttitle{Type~Ia Supernova Width Luminosity Relation}
50:
51: \submitted{}
52:
53: \bibliographystyle{apj}
54:
55: \begin{document}
56:
57: \title{On the Origin of the Type~Ia Supernova Width-Luminosity Relation}
58:
59: \author{Daniel Kasen\altaffilmark{1,2}\email{kasen@pha.jhu.edu} and S.E. Woosley\altaffilmark{3}}
60: \altaffiltext{1}{Allan C. Davis Fellow, Department of Physics and Astronomy, Johns Hopkins University, Baltimore, MD 21218}
61: \altaffiltext{2}{Space Telescope Science Institute, Baltimore, MD 21218}
62: \altaffiltext{3}{University of California, Santa Cruz}
63:
64:
65: \begin{abstract}
66: Brighter Type~Ia supernovae (SNe~Ia) have broader, more slowly
67: declining \Bband\ light curves than dimmer SNe~Ia. We study the
68: physical origin of this width-luminosity relation (WLR) using detailed
69: radiative transfer calculations of Chandrasekhar mass SN~Ia models. We
70: find that the luminosity dependence of the diffusion time (emphasized
71: in previous studies) is in fact of secondary relevance in
72: understanding the model WLR. Instead, the essential physics involves
73: the luminosity dependence of the
74: \emph{spectroscopic/color} evolution of SNe~Ia. Following
75: maximum-light, the SN colors are increasingly affected by the
76: development of numerous Fe~II/Co~II lines which blanket the \Bband\
77: and, at the same time, increase the emissivity at longer wavelengths.
78: Because dimmer SNe~Ia are generally cooler, they experience an earlier
79: onset of Fe~III to Fe~II recombination in the iron-group rich layers
80: of ejecta, resulting in a more rapid evolution of the SN colors to the
81: red. The faster \Bband\ decline rate of dimmer SNe~Ia thus reflects
82: their faster ionization evolution.
83: \end{abstract}
84:
85: \keywords{radiative transfer -- supernovae; cosmology}
86:
87: \section{Introduction}
88:
89: Although normal Type~Ia supernovae (SNe~Ia) are generally considered a
90: homogeneous class, they nevertheless show substantial ($\sim 1$~mag)
91: variations in luminosity at peak. The use of SNe~Ia for cosmology
92: measurements thus relies on empirical calibration techniques. Most
93: common among these is the width-luminosity relation
94: \citep{Phillips_1999}. Bright SNe~Ia generally
95: have broad \Bband\ light curves (LCs) which decline slowly after
96: peak. Dimmer SNe~Ia have narrower, more quickly declining LCs.
97:
98: Understanding the physical origin of the width-luminosity relation
99: (WLR) is a primary goal of the theory and modeling of SNe~Ia.
100: Unfortunately, a full theoretical description of the SN~Ia explosion
101: mechanism is still lacking. Normal SNe~Ia are widely believed to be
102: the thermonuclear disruption of carbon-oxygen white dwarfs near the
103: Chandrasekhar limit, however a number of uncertainties remain
104: regarding the structure of the progenitor, the precise ignition
105: conditions, and the physics of the turbulent nuclear combustion that
106: unbinds the star.
107:
108: Fortunately, these uncertainties need not be fully resolved in order
109: to study the WLR. In SN~Ia explosions, hydrodynamic and nuclear
110: burning processes last only $\sim 1$~minute. The subsequent
111: luminosity is powered entirely by the decay of radioactive elements
112: synthesized in the explosion, in particular
113: \Nifs\ in the decay chain $\Nifs \rightarrow\ \Cofs \rightarrow\
114: \Fefs$. The LCs of SNe~Ia are therefore fully determined by the
115: composition and density structure of the material ejected in the
116: explosion. Naturally, the mass of \Nifs\ produced (\Mni) is the
117: primary determinate of the peak brightness of the event. Observations
118: indicate that for normal objects, \Mni\ spans the range
119: $0.4-0.9$~\Msun, with a typical value $\Mni \approx 0.6~\Msun$.
120:
121: The challenge thus falls to radiative transfer theory to explain why
122: SNe~Ia with larger \Mni\ also have broader \Bband\ LCs. Most previous
123: transfer studies have emphasized diffusion arguments of one sort or
124: another \citep{Hoeflich_WLR, Hoeflich_99by, Pinto-Eastman_WLR,
125: Mazzali_WLR}. In SNe~Ia, the ejecta remain optically thick for the
126: first several months after explosion. The width of the bolometric LC
127: is related to the time scale for photons to escape the ejecta by
128: diffusion. In principle, the WLR could be explained if brighter
129: SNe~Ia have higher effective opacities and hence a longer diffusion
130: time. Different authors have invoked different physical arguments to
131: motivate this sort of opacity dependence (see
132: \S\ref{Sec:Others}).
133:
134: In this paper, we stress that the WLR does not in fact hinge on the
135: bolometric diffusion time, but is instead principally a
136: \emph{broadband} phenomenon. In particular, the \Bband\ LC depends
137: sensitively on the rate at which the SN colors evolve progressively
138: redward following maximum light. Dimmer SNe~Ia exhibit a faster color
139: evolution than brighter SNe~Ia; this turns out to be a primary reason
140: for their relatively faster
141: \Bband\ decline rates. A relevant theoretical explanation of the
142: WLR should therefore focus on the effect \Mni\ has on the
143: \emph{spectroscopic/color evolution} of SNe~Ia, rather than on the
144: overall diffusion time-scale.
145:
146: In what follows, we use realistic time-dependent multi-group radiative
147: transfer calculations to explain the physical origin of the WLR in a
148: set of Chandrasekhar-mass SN~Ia models. We demonstrate that the faster
149: color evolution of dimmer models reflects their faster
150: \emph{ionization evolution}. Following
151: maximum-light, the SN colors are increasingly affected by the
152: development of numerous Fe~II/Co~II lines which blanket the
153: \Bband\ and, at the same time,
154: increase the emissivity at longer wavelengths. Because dimmer SNe~Ia
155: are generally cooler, they experience an earlier onset of Fe~III to
156: Fe~II recombination in the iron-group rich layers of ejecta, resulting
157: in a more rapid evolution of the SN colors to the red. This explains
158: their faster \Bband\ decline rate.
159:
160: \section{Previous Theoretical Studies}
161: \label{Sec:Others}
162:
163:
164:
165:
166: Previous radiative transfer studies of SNe~Ia have occasionally
167: reproduced the basic trend of the WLR among sets of models with
168: varying \Mni. In these studies, the model relation is typically
169: explained by claiming a strong luminosity dependence of the diffusion
170: time. To the extent that the transfer calculations include the
171: relevant multi-wavelength processes, color evolution effects are also
172: captured in the simulations as a matter of course. Sometimes their
173: relevance to the WLR is explicitly recognized and quantified
174: \citep[e.g.,][]{Mazzali_WLR}.
175:
176: Using simple dimensional arguments, one can show that the photon
177: diffusion time (and hence bolometric LC width) scales with mean
178: opacity as $t_d \propto \kappa^{1/2}$. In SNe~Ia, the dominant
179: opacities are due to electron scattering and bound-bound line
180: transitions, the latter of which is enhanced by Doppler broadening in
181: the differentially expanding ejecta. Iron group species (iron,
182: cobalt, and nickel) have the largest line opacity, due to their
183: complex atomic structure. The iron group line opacities increases
184: sharply to the blue, due to the larger number of lines at shorter
185: wavelengths.
186:
187: At least three different (though related) arguments have been given as
188: to why $\kappa$ (and hence the diffusion time) should increase with
189: \Mni, each emphasizing a different physical mechanism.
190:
191: \citet[][see also \cite{Khokhlov_93,Hoeflich_WLR}]{Hoeflich_99by}
192: emphasize the temperature dependence of the diffusion time. Because
193: SNe~Ia with larger \Mni\ have higher temperatures, the bulk of the
194: radiation energy is concentrated at shorter wavelengths. Since the
195: line opacity increases sharply to the blue, the effective mean opacity
196: (and hence the diffusion time) increases with \Mni.
197:
198: \citet[][see also \cite{Pinto-Eastman_I, Pinto-Eastman_II}]{Pinto-Eastman_WLR}
199: emphasize the effect of the ionization state on the diffusion time.
200: They argue that \td\ is largely determined by the rate at which
201: blue/ultraviolet photons fluoresce to longer wavelengths via
202: interaction with iron group elements, thereby enhancing photon escape.
203: They suggest that fluorescence is more efficient in singly, as opposed
204: to doubly ionized species. Because SNe~Ia with lower \Mni\ are cooler
205: and less ionized, the diffusion time is shorter.
206:
207: \cite{Mazzali_WLR} emphasize the effect of the ejecta composition on
208: the diffusion time. SNe~Ia with large \Mni\ have a greater abundance
209: of iron group elements, and thus presumably larger opacities.
210:
211: In fact, all three of the above mentioned diffusion effects must be
212: operative at some level in the LCs of SNe~Ia. The magnitude of the
213: effects (in sum) can be quantified by direct examination of the model
214: bolometric LCs. If diffusion time arguments are indeed of prime
215: relevance in explaining the WLR, it follows that the bolometric LCs of
216: brighter models must be significantly broader than those of dimmer
217: models. However, in the previous studies this bolometric behavior is
218: not explicitly demonstrated using detailed multi-group transfer
219: calculations and realistic opacities.
220:
221: In the following, we present a set of model calculations in which the
222: shape of the bolometric LCs in fact shows very little direct
223: dependence on \Mni. Variations in the diffusion time are therefore of
224: secondary relevance in addressing the \Bband\ WLR. Instead, the model
225: WLR primarily reflects the faster spectroscopic/color evolution of
226: SNe~Ia with lower \Mni. It is therefore of fundamental importance
227: (and a primary goal of this paper) to explain in detail the physics
228: controlling the rate of color evolution in SNe~Ia.
229:
230: Because the flux mean opacity itself depends on the spectral energy
231: distribution of the radiation field, the physics controlling the color
232: evolution of SNe~Ia is of course deeply intertwined with that
233: determining the diffusion time. In the end, however, the relative
234: importance of each can be quantified by comparing the strength of the
235: WLR in the broadband versus the bolometric LCs of SNe~Ia. We suggest
236: below that such an analysis may be very useful in furthering our
237: understanding of SNe~Ia.
238:
239:
240: \begin{deluxetable*}{llllllllllllll}
241: \tablecaption{Observable properties of the models}
242: \tablehead{
243: \Mni & $t_{bol}$ & $t_B$ & $M_{\rm bol}$ & $M_B$ & \Mv\ & \dmfbol\ & \dmfb\ & \dmfv\ & \dmfubv & \bmvmax & \bmvf\ }
244: \startdata
245: 0.70 & 15.56 & 19.34 & -19.46 & -19.34 & -19.35 & 0.81 & 1.09 & 0.72 & 0.83 & -0.017 & 0.568 \\
246: 0.63 & 15.55 & 19.00 & -19.36 & -19.25 & -19.27 & 0.82 & 1.17 & 0.77 & 0.84 & -0.007 & 0.621 \\
247: 0.56 & 15.54 & 18.70 & -19.24 & -19.14 & -19.18 & 0.83 & 1.28 & 0.82 & 0.88 & 0.011 & 0.693 \\
248: 0.49 & 15.56 & 18.36 & -19.11 & -19.02 & -19.08 & 0.84 & 1.38 & 0.86 & 0.91 & 0.034 & 0.750 \\
249: 0.42 & 15.52 & 17.85 & -18.96 & -18.87 & -18.97 & 0.84 & 1.45 & 0.92 & 0.94 & 0.057 & 0.802 \\
250: 0.35 & 15.51 & 17.22 & -18.78 & -18.71 & -18.83 & 0.84 & 1.53 & 0.98 & 0.96 & 0.089 & 0.851 \\
251: \enddata
252: \end{deluxetable*}
253:
254:
255:
256: \section{Models}
257:
258:
259: In a companion paper \citep[][hereafter Paper~I]{Woosley_WLR}, we
260: perform an extensive parameter study of SN~Ia LCs using 1-dimensional
261: (1D) hydrodynamical models of the explosion. Here we focus on a
262: restricted set of those models in order to demonstrate the basic
263: transfer physics underlying the WLR. Each model is a derivative of
264: the spherical Chandrasekhar mass explosion model, M070103, which has a
265: four zone stratified compositional structure. From the center out,
266: these zones consist of: 0.1~\Msun\ of stable iron group elements;
267: 0.7~\Msun\ of \Nifs; 0.3~\Msun\ of intermediate mass elements (IMEs;
268: silicon,sulfur, calcium, argon); and 0.3~\Msun\ of unburned
269: carbon/oxygen. Mild mixing is applied at the zone interfaces. The
270: synthetic LCs and spectral time-series of model M070103 are in good
271: agreement with observed normal SNe~Ia (Paper~I).
272:
273: A simple prescription is used to vary the \Nifs\ abundance in model
274: M070103, allowing us to isolate the effect of \Mni\ on the LCs. New
275: models are constructed by reducing the abundance of \Nifs\ everywhere
276: by a fractional amount (viz., $10,20,30,40,$ and $50$\%) while
277: increasing the abundance of IMEs by a corresponding amount. This
278: amounts to a variation of \Mni\ from 0.35 to 0.7~\Msun\ while
279: maintaining a fixed velocity distribution of \Nifs. Because the
280: energy released in burning to IMEs is comparable to burning to \Nifs,
281: the dynamics of the explosion will not be greatly affected by the
282: compositional change. The ejecta structures bear resemblance to some
283: 1D delayed-detonation or pulsed-detonation models
284: \citep{Hoeflich_DD}.
285:
286: We compute synthetic LCs of the models using the multi-dimensional
287: time-dependent Monte Carlo radiative transfer code \MCcode\
288: \citep{Kasen_MC}. Given a homologously expanding SN ejecta structure,
289: \MCcode\ calculates the emergent spectral time series at high
290: wavelength resolution. Broadband LCs are then constructed by
291: convolving the spectra with the appropriate filter transmission
292: functions. The code includes a detailed gamma-ray transfer procedure
293: to determine the rate of radioactive energy deposition, and a
294: radiative equilibrium solution of the temperature structure. No
295: ad-hoc inner boundary condition is employed. Ionization and
296: excitation are computed assuming local thermodynamic equilibrium, and
297: bound-bound line transitions are treated using the expansion opacity
298: formalism and an approximate two-level atom approach to wavelength
299: redistribution, assuming a constant redistribution probability
300: $\epsilon = 0.8$ for all lines. We have used
301: \MCcode\ to recalculate the broadband LCs of the scaled models used to
302: study the WLR in
303: \cite{Pinto-Eastman_WLR} and found excellent agreement with the
304: results presented there. See \cite{Kasen_MC} for a detailed code
305: description and verification.
306:
307:
308: \section{Origin of the Width-Luminosity Relation}
309:
310: \begin{figure}
311: \begin{center}
312: \includegraphics[width=8.5cm]{f1.eps}
313: \caption{\Bband\ light curves (left panel) and width-luminosity relation
314: (right panel) of the models discussed in the text. The shaded region
315: in the right panel is the empirical WLR of \cite{Phillips_1999} with
316: a magnitude dispersion of 0.15~mag and absolute calibration $\Mb =
317: -19.3$~mag at $\dmfb = 1.1$~mag.
318: \label{Fig:wlr}}
319: \end{center}
320: \end{figure}
321:
322: The WLR is often quantified as a relation between peak \Bband\
323: magnitude, \Mb, and the drop in \Bband\ magnitude 15 days after peak,
324: \dmfb. Table~1 lists the relevant observable parameters of the
325: synthetic LC models discussed in the previous section. In the models,
326: the variation of \Mni\ from 0.35 to 0.70~\Msun\ results in a variation
327: in
328: \Mb\ from -18.71 to -19.34~mag, and in \dmfb\ from 1.53 to 1.09~mag.
329: As seen in Figure~1, the models obey a clear WLR, with a slope that
330: agrees reasonably (though not exactly) with the observed one of
331: \cite{Phillips_1999}.
332:
333: The model WLR obtained in Figure~\ref{Fig:wlr} is not attributable to
334: a strong systematic dependence of the diffusion time on \Mni, as
335: emphasized in previous studies. On the contrary, the bolometric rise
336: time to maximum and the bolometric decline rate are similar among all
337: models: $\tbol = 15.51-15.56$~days, $\dmfbol = 0.81-0.84$~mag. As
338: seen in Figure~\ref{Fig:bol}, differences in the bolometric LCs are
339: maximal (though still modest) when measured 30 days after $B$-maximum:
340: $\dmfbolt = 1.43-1.61$~mag. Thus, in this set of models, the
341: luminosity dependence of the diffusion time is in fact of secondary
342: relevance in understanding the WLR.
343:
344: Instead, we identify the crucial physical effect to be that of \Mni\
345: on the spectroscopic/color evolution of the models. As seen in
346: Figure~\ref{Fig:colors}, all models have very similar colors at
347: \Bband\ maximum, with \bmv\ and \bmr\ within $0$ and $0.15$~mag.
348: Thereafter, the color curves diverge, with the dimmer models evolving
349: redward more rapidly. By 15 days after $B$-maximum, differences in
350: \bmv\ and \bmr\ color have grown to nearly 0.3 and 0.6~mag
351: respectively. It is this faster color evolution of the dimmer models
352: that accounts for their larger \dmfb.
353:
354: The origin of this color evolution behavior can be traced directly to
355: the impact of iron-group lines on the post-maximum spectra. Soon after
356: maximum light, the \Bband\ is increasingly affected by the development
357: of Fe~II/Co~II lines (Figure~\ref{Fig:evolution}). Especially
358: prominent are the strong absorption blends at wavelengths
359: 4000-4600~\AA\ and 4800-5300~\AA, however a pseudo-continuum of
360: numerous weaker lines also contribute to the overall line blanketing
361: in the blue. At the same time, a large number of blended Fe~II/Co~II
362: lines emerge in the far red and near-infrared, enhancing the
363: pseudo-continuum emissivity at these wavelengths
364: \citep{Kasen_NIR}. The net effect of the onset of the
365: Fe~II/Co~II lines is thus to redistribute flux absorbed in the blue
366: part of the spectrum to longer wavelengths. Naturally, the rate at
367: which these lines develop plays a critical role in determining the
368: color evolution (and hence
369: \Bband\ decline rate) of SNe~Ia.
370:
371: In general, Fe~II/Co~II lines become prominent once the layers of
372: ejecta rich in \Nifs\ (since decayed to iron/cobalt) begin to
373: recombine from doubly to singly ionized, a process which occurs quite
374: suddenly at temperature $T \approx 7000$~K. For the epochs at and
375: before \Bband\ maximum, the iron/cobalt-rich layers of ejecta ($v \la
376: 9000$~\kms) have $T > 7000$~K and Fe~II/Co~II lines are weak or absent
377: in all model spectra (see Figure~\ref{Fig:spectra}; top panel). At
378: these times, the optical continuum is pseudo-blackbody and the
379: spectral and color differences among the models are minor, even when
380: \Mni\ is varied by a factor of two. The brighter/broader SNe~Ia have
381: a bluer
382: \bmvmax\ color, but by a relatively small amount, as noted in
383: observations \citep{Phillips_1999}.
384:
385: In the weeks following maximum light, however, the layers of iron-rich
386: ejecta progressively cool to $T \la 7000$~K and, as these layers
387: recombine, the Fe~II/Co~II lines become steadily stronger. The exact
388: rate at which the lines develop depends sensitively on the overall
389: ejecta temperature scale. Models with lower \Mni\ have generally
390: lower temperatures and hence an earlier onset of Fe~III to Fe~II
391: recombination. As seen in Figure~\ref{Fig:spectra} (bottom panel) by
392: 15 days after maximum the line blanketing is much stronger (and the
393: colors much redder) in the lower \Mni\ model. The larger
394: \dmfb\ of the dimmer model thus reflects the
395: more rapid development of the Fe~II/Co~II lines.
396:
397:
398: To be clear, the WLR is not merely a consequence of the generally
399: redder colors of dimmer SNe~Ia, but rather of the faster rate at which
400: the colors \emph{evolve} in these objects. In considering this
401: spectroscopic effect, it is useful to keep in mind that the
402: temperature differences among normal SNe~Ia are expected to be rather
403: small. The ejecta temperature (as determined by the balance of
404: radiative heating and cooling) scales with luminosity as $T
405: \propto L^{1/4}$ and hence $T \propto
406: \Mni^{1/4}$. A factor of two difference in \Mni\ therefore amounts to
407: only a $\sim 20\%$ difference in temperature. If the SN spectrum is
408: reasonably characterized by a blackbody, the corresponding variations
409: in \bmv\ or \bmr\ color would be relatively small ($\la 0.15$~mag).
410: Indeed, this is essentially why the pre-maximum and maximum light
411: colors of normal SNe~Ia show only a mild dependence on \Mni.
412:
413: In the post-maximum epochs, however, the SN~Ia spectrum increasingly
414: deviates from blackbody due to the burgeoning strength of the
415: Fe~II/Co~II lines. The strength of these lines depends very
416: sensitively on the ionization state. The ionization in turn obeys an
417: extremely non-linear dependence on temperature, undergoing sudden
418: recombination at $T \approx 7000$~K. Thus, in the post-maximum
419: epochs, small differences in temperature translate to large
420: differences in \bmv\ and \bmr\ color. The onset of this non-linear
421: behavior is why the decline rate \dmfb\ offers a good diagnostic of
422: the SN temperature scale, and hence luminosity. Other approaches to
423: calibrating SNe~Ia, such as the CMAGIC method \citep{Wang_Cmagic} and
424: the $C_{12}$ correlation
425: \citep{Wang_C12} are likewise explained by the same strong dependence
426: of the SN colors on the ionization state during the weeks following
427: maximum light.
428:
429: At around 30 days after $B$-maximum, the iron group layers of ejecta
430: finish completely recombining from doubly to singly ionized. From
431: this time on, differences in the ionization state among the models
432: cease to be of great significance. The color curves reconverge and
433: the models enter a phase of nearly identical color evolution (see
434: Figure~\ref{Fig:colors}). The same general behavior is noted in the
435: observations \citep{Lira_1996, Phillips_1999}. For this reason, the
436: decline rate measured over longer time periods (e.g., \dmfbt) provides
437: a less sensitive measure of the model peak magnitude.
438:
439: \section{Relevance To Observations}
440:
441:
442: \begin{figure}[t]
443: \begin{center}
444: \includegraphics[width=8.5cm]{f2.eps}
445: \caption{\emph{Left:}
446: Bolometric LCs of the models, normalized at peak and shown relative to
447: time of bolometric maximum. The LC shapes are all very similar over
448: the first three weeks after maximum light. Moderate differences of
449: order 0.15~mag appear around 30 days after maximum, when the brighter
450: models reach a small secondary maximum.
451: \emph{Right:} \Bband\ of the models, normalized at peak and shown
452: relative to time of $B$-maximum. The LC decline rate just after maximum
453: depends strongly on
454: \Mni.
455: \label{Fig:bol}}
456: \end{center}
457: \end{figure}
458:
459:
460: \begin{figure}[t]
461: \begin{center}
462: \includegraphics[width=8.5cm]{f3.eps}
463: \caption{Color evolution of the models discussed in the text. In all panels,
464: the dimmer models are the topmost lines. All models have similar
465: colors near maximum light; thereafter the dimmer ones (lower \Mni)
466: become redder more rapidly. The differing rates of color evolution
467: account for differences in \dmfb.
468: \label{Fig:colors}}
469: \end{center}
470: \end{figure}
471:
472:
473:
474:
475: \begin{figure}
476: \begin{center}
477: \includegraphics[width=8.5cm]{f4.eps}
478: \caption{Spectral evolution from \Bband\
479: maximum to twenty days later for the $\Mni = 0.7~\Msun$ model. The
480: figure illustrates how the drop in $B$-band magnitude depends
481: sensitively on color changes due, in large part, to the development of
482: spectral line blanketing from iron group elements. The shaded bands in
483: the plot highlight two prominent spectral absorption features due to
484: Fe~II/Co~II line blends.
485: \label{Fig:evolution}}
486: \end{center}
487: \end{figure}
488:
489:
490:
491: \begin{figure}
492: \begin{center}
493: \includegraphics[width=8.5cm]{f5.eps}
494: \caption{Demonstration of the crucial effect of Fe~II/Co~II
495: line blanketing on \dmfb. The figure compares synthetic
496: spectra of the \Mni = 0.35 and \Mni = 0.70~\Msun\ models. All spectra
497: have been normalized by dividing by the bolometric luminosity at the
498: time. The blue, green, and red color shadings show the regions of the
499: B,V and R bands respectively. Near maximum light (top panel) both
500: models have similar, pseudo-blackbody coninuua. Fifteen days later
501: (bottom panel) broad absorption features from Fe~II/Co~II lines
502: dominate the \Bband. Because the cooler \Mni = 0.35~\Msun\ model is
503: less ionized, the post-maximum line blanketing develops more quickly,
504: leading to the faster \Bband\ decline rate.
505: \label{Fig:spectra}}
506: \end{center}
507: \end{figure}
508:
509:
510:
511: The observed color evolution of SNe~Ia bears a striking similarity to
512: the model behavior of Figure~\ref{Fig:colors}
513: \citep[e.g.,][Figure~1]{Wang_C12}. Normal SNe~Ia display fairly
514: uniform \bmv\ colors at maximum light; thereafter the dimmer objects
515: become redder more quickly. The observed differences in color at day
516: $+15$ are comparable to the differences in \dmfb.
517:
518: \cite{Gaston_EW}
519: quantifies the spectral evolution of SNe~Ia directly using equivalent
520: width measures. His approach nicely illustrates the sudden growth of
521: Fe~II/Co~II features (there labeled ``Mg~II'') in the $B$-band
522: beginning just after maximum light. The rate at which these features
523: develop is shown to correlate tightly with \dmfb. Such analyses
524: clearly point to the spectroscopic nature of the WLR.
525:
526: The construction of bolometric LCs from $uvoir$ observations of SNe~Ia
527: can be used to further explore this issue. If the luminosity
528: dependence of the diffusion time is in fact central to explaining the
529: WLR, one expects to observe a strong width luminosity relation in the
530: \emph{bolometric} LCs of SNe~Ia. It is currently unclear from the limited
531: observational sample whether such a relation exists at all.
532: \cite{Phillips_99ac}, for example, construct bolometric LCs
533: from $UBVRI$- and (when available) $JHK$-band observations of 15
534: SNe~Ia. The bolometric decline rates \dmfbol\ of the sample show some
535: degree of scatter, but no clear correlation with the peak bolometric
536: magnitude $M_{\rm bol}$.
537:
538: The bolometric LC constructions of
539: \cite{Contardo_bol}, which use primarily $BVRI$-band observations, do show a
540: noisy correlation between
541: \dmfbol\ and \Mbol\ \citep[see also][]{Stritz}.
542: This correlation, however, may in large part reflect the lack of
543: ultraviolet and near-infrared (NIR) data used in the bolometric
544: constructions. At day 15 after $B$-maximum, for example, 20\% or more
545: of the SN~Ia flux may occur at NIR wavelengths
546: \citep{Suntzeff_bol}. As the NIR contribution at day~15
547: is relatively greater in the dimmer SNe~Ia, $UBVRI$ constructions may
548: systematically overestimate the value of \dmfbol\ for dimmer objects
549: by as much as $0.1-0.2$~mag. This behavior is seen in our models
550: (Table~1) where one notes that
551: \dmfubv\ exhibits a much stronger dependence on
552: \Mni\ than does the true bolometric decline rate
553: \dmfbol. In fact, the slope of our
554: $UBVRI$ model WLR agrees fairly well with the empirical one of
555: \cite{Contardo_bol}, suggesting that the
556: observed correlation reflects, in large part, the role of color
557: evolution.
558:
559: Clearly the construction of true bolometric LCs for a sample of SNe~Ia
560: using UV through $K$-band observations would be of great value in
561: constraining the radiative transfer in SNe~Ia. In the specific models
562: studied in this paper, differences in the bolometric LCs are
563: relatively small, and attributable to the reduction in flux mean
564: opacity which accompanies the onset of Fe~II/Co~II lines
565: \citep{Pinto-Eastman_WLR,Kasen_NIR}. The bolometric differences
566: become most prominent around 30 days after $B$-maximum, when the
567: brighter models show a slight rise to a secondary bolometric maximum.
568: Differences in diffusion time may therefore be somewhat more relevant
569: in formulations of the WLR employing later time data.
570:
571: The bolometric LCs of SNe~Ia may also depend sensitively on the ejecta
572: kinetic energy and the radial distribution of \Nifs\ (see Paper~I).
573: By varying these two other parameters, it is possible to construct
574: sets of models that show greater or lesser degrees of scatter and/or
575: correlation in the bolometric WLR. In addition, NLTE effects (not
576: included here) may also contribute to a stronger luminosity dependence
577: of the bolometric LCs, especially if the probability of wavelength
578: redistribution in lines is strongly temperature dependent. Such
579: issues will be the subject of future studies. However, because the
580: rate of color evolution dominates the $B$-band decline rate, the WLR
581: should remain a robust luminosity calibration even when the bolometric
582: LCs of SNe~Ia show significant random variations.
583:
584:
585: The bolometric rise time to maximum is the same for all models in this
586: paper: $\tbol = 15.5$~days. The \Bband\ rise time (\tb), on the other
587: hand, depends on the color evolution. Because iron group recombination
588: occurs earlier in lower \Mni\ models, their colors begin evolving
589: redward at earlier epochs. Thus, in both models and observations, the
590: dimmer SNe~Ia exhibit a shorter \tb.
591:
592: Because Fe~II/Co~II line blanketing generally increases to the blue,
593: the WLR will usually be stronger in bluer wavelength bands. The
594: $V$-band, for instance, is impacted by the development of Fe~II/Co~II
595: lines to a lesser extent than the \Bband. This leads, in both models
596: and observations, to a slower $V$-band decline rate \dmfv\ and a weaker
597: dependence of \dmfv\ on the peak $V$-band magnitude \Mv.
598:
599: The \Bband\ decline rate also has an intimate connection to the
600: double-peaked morphology of the far red and NIR LCs of SNe~Ia. The
601: flux absorbed in Fe~II/Co~II lines at blue wavelengths is
602: redistributed to the red, eventually leading to a NIR secondary
603: maximum
604: \citep{Pinto-Eastman_II, Kasen_NIR}.
605: Naturally, in both models and observations, SNe~Ia with faster \Bband\
606: decline rates have earlier NIR secondary maxima.
607:
608: On the whole, one recognizes that \emph{the ionization evolution of
609: iron group elements is the most significant physical factor in
610: understanding SN~Ia LCs and the WLR}. The faster overall
611: spectroscopic and photometric evolution of dimmer SNe~Ia reflects,
612: primarily, their faster ionization evolution.
613:
614: In addition, the distribution of iron group elements in the ejecta can
615: be identified as a crucial secondary parameter in the WLR. All models
616: in this paper have substantial iron group abundance out to $v \approx
617: 9000~\kms$. If the iron group elements are mixed to yet higher
618: velocities, the development of Fe~II/Co~II lines will be hastened and
619: intensified. Thus, for given \Mni, enhancing the iron group abundance
620: in the outer layers of ejecta
621: \emph{increases} the \Bband\ decline rate. The relatively low dispersion
622: in the observed WLR suggests that normal SNe~Ia exhibit limited
623: variability in their iron group distributions. The physical origin of
624: this regularity, however, can only be addressed by detailed first
625: principle explosion modeling of SNe~Ia.
626:
627: Fortunately for future precision cosmology experiments, the radiative
628: transfer effects identified in this paper can be used to monitor and
629: limit deviations in the WLR. Variations in the velocity distribution
630: of iron group elements, for example, can be constrained by measuring
631: the Doppler-shifts of iron features in the SN~Ia spectra. NIR
632: observations provide additional indicators of the ionization evolution
633: and iron group distribution in the ejecta \citep{Kasen_NIR}. Close
634: inspection of well sampled color curves (or spectral time series)
635: should reveal equivalent or improved calibration techniques and may be
636: useful in quantifying subtle deviations from the standard
637: evolution. These and related tests should be feasible with the data
638: from ongoing observational campaigns, and will help to limit the
639: intrinsic scatter or evolutionary biases potentially affecting SNe~Ia
640: cosmology experiments. With the aid of detailed transfer models such
641: as those presented here, refinements to the calibration techniques
642: need not rely on the fortuitous discovery of empirical correlations,
643: but can be devised to reflect the essential physics of SN~Ia light
644: curves.
645:
646: \acknowledgements
647: The authors would like to thank Phil Pinto and Ron Eastman for very
648: helpful discussions regarding the behavior of the bolometric light
649: curves, both in their prior theoretical studies and in the
650: observations. We also thank Alex Conley, David Jeffery, and Adam
651: Riess for helpful comments and discussions. DK is supported by the
652: Allan C. Davis fellowship at Johns Hopkins University and the Space
653: Telescope Science Institute. SEW acknowledges support from NASA
654: (NNG05GG08G), and the DOE Program for Scientific Discovery through
655: Advanced Computing (SciDAC; DE-FC02-01ER41176) This research used
656: resources of the National Energy Research Scientific Computing Center,
657: which is supported by the Office of Science of the U.S. Department of
658: Energy under Contract No. DE-AC03-76SF00098.
659:
660:
661: \begin{thebibliography}{16}
662: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
663:
664: \bibitem[{{Contardo} {et~al.}(2000){Contardo}, {Leibundgut}, \&
665: {Vacca}}]{Contardo_bol}
666: {Contardo}, G., {Leibundgut}, B., \& {Vacca}, W.~D. 2000, A\&A, 359, 876
667:
668: \bibitem[{{Folatelli}(2004)}]{Gaston_EW}
669: {Folatelli}, G. 2004, New Astronomy Review, 48, 623
670:
671: \bibitem[{{H{\"o}flich} {et~al.}(2002){H{\" o}flich}, {Gerardy}, {Fesen}, \&
672: {Sakai}}]{Hoeflich_99by}
673: {H{\" o}flich}, P., {Gerardy}, C.~L., {Fesen}, R.~A., \& {Sakai}, S. 2002, ApJ,
674: 568, 791
675:
676: \bibitem[{{H{\"o}flich} {et~al.}(1996){H{\"o}flich}, {Khokhlov}, {Wheeler},
677: {Phillips}, {Suntzeff}, \& {Hamuy}}]{Hoeflich_WLR}
678: {H{\"o}flich}, P., {Khokhlov}, A., {Wheeler}, J.~C., {Phillips}, M.~M.,
679: {Suntzeff}, N.~B., \& {Hamuy}, M. 1996, ApJ, 472, L81+
680:
681: \bibitem[{{H{\"o}flich} {et~al.}(1995){H{\"o}flich}, {Khokhlov}, \&
682: {Wheeler}}]{Hoeflich_DD}
683: {Hoflich}, P., {Khokhlov}, A.~M., \& {Wheeler}, J.~C. 1995, ApJ, 444, 831
684:
685: \bibitem[{{Kasen}(2006)}]{Kasen_NIR}
686: {Kasen}, D. 2006, ApJ, 649, 939, (astro-ph/0606449)
687:
688: \bibitem[{{Kasen} {et~al.}(2006)}]{Kasen_MC}
689: {Kasen}, D., {et~al.} 2006, ApJ, accepted (astro-ph/0606111)
690:
691: \bibitem[Khokhlov et al.(1993)]{Khokhlov_93} Khokhlov, A., Mueller,
692: E., \& Hoeflich, P.\ 1993, \aap, 270, 223
693:
694: \bibitem[Lira(1996)]{Lira_1996} Lira, P.\ 1996, Masters Thesis, 3
695:
696:
697: \bibitem[{{Mazzali} {et~al.}(2001){Mazzali}, {Nomoto}, {Cappellaro},
698: {Nakamura}, {Umeda}, \& {Iwamoto}}]{Mazzali_WLR}
699: {Mazzali}, P.~A., {Nomoto}, K., {Cappellaro}, E., {Nakamura}, T., {Umeda}, H.,
700: \& {Iwamoto}, K. 2001, ApJ, 547, 988
701:
702: \bibitem[{{Phillips} {et~al.}(1999){Phillips}, {Lira}, {Suntzeff}, {Schommer},
703: {Hamuy}, \& {Maza}}]{Phillips_1999}
704: {Phillips}, M.~M., {Lira}, P., {Suntzeff}, N.~B., {Schommer}, R.~A., {Hamuy},
705: M., \& {Maza}, J. 1999, AJ, 118, 1766
706:
707:
708: \bibitem[Phillips et al.(2006)]{Phillips_99ac} Phillips, M.~M., et
709: al.\ 2006, \aj, 131, 2615
710:
711: \bibitem[{{Pinto} \& {Eastman}(2000{\natexlab{a}})}]{Pinto-Eastman_I}
712: {Pinto}, P.~A., \& {Eastman}, R.~G. 2000{\natexlab{a}}, ApJ, 530, 744
713:
714: \bibitem[{{Pinto} \& {Eastman}(2000{\natexlab{b}})}]{Pinto-Eastman_II}
715: ---. 2000{\natexlab{b}}, ApJ, 530, 757
716:
717: \bibitem[{{Pinto} \& {Eastman}(2001)}]{Pinto-Eastman_WLR}
718: ---. 2001, New Astronomy, 6, 307
719:
720: \bibitem[Suntzeff(2003)]{Suntzeff_bol} Suntzeff, N.~B.\ 2003, From
721: Twilight to Highlight: The Physics of Supernovae, 183
722:
723: \bibitem[Stritzinger et al.(2006)]{Stritz} Stritzinger, M., et
724: al.\ 2006, A\&A, 131, 2615
725:
726: \bibitem[{{Wang} {et~al.}(2003){Wang}, {Goldhaber}, {Aldering}, \&
727: {Perlmutter}}]{Wang_Cmagic}
728: {Wang}, L., {Goldhaber}, G., {Aldering}, G., \& {Perlmutter}, S. 2003, ApJ,
729: 590, 944
730:
731: \bibitem[{{Wang} {et~al.}(2005){Wang}, {Wang}, {Zhou}, {Lou}, \&
732: {Li}}]{Wang_C12}
733: {Wang}, X., {Wang}, L., {Zhou}, X., {Lou}, Y.-Q., \& {Li}, Z. 2005, ApJ, 620,
734: L87
735:
736: \bibitem[{{Woosley} {et~al.}(2006){Woosley}, {Kasen}, {Blinnikov}, \&
737: {Sorokina}}]{Woosley_WLR}
738: {Woosley}, S., {Kasen}, D., {Blinnikov}, S., \& {Sorokina}, E. 2006, ApJ, submitted.
739:
740: \end{thebibliography}
741:
742: %\bibliography{../../BIBIN/MAIN.bib}
743:
744: \clearpage
745:
746:
747:
748:
749: \end{document}
750:
751:
752:
753: