1: %% Blank ms. for AASTEX 5.2, edited down from sample.tex 1/31/07
2:
3: \documentclass{emulateapj}
4: %\begin{document}
5: \usepackage{apjfonts}
6:
7: %\documentclass[12pt,preprint]{aastex}
8:
9: %% \documentclass[manuscript]{aastex}
10:
11: %% preprint2 produces a double-column, single-spaced document:
12:
13: %% \documentclass[preprint2]{aastex}
14:
15: \newcommand{\msun}{M_\odot}
16: \catcode`\@=11
17: \newcommand{\gapprox}{\mathrel{\mathpalette\@versim>}}
18: \newcommand{\lapprox}{\mathrel{\mathpalette\@versim<}}
19: \newcommand{\propapprox}{\mathrel{\mathpalette\@versim\propto}}
20: \newcommand{\@versim}[2]
21: {\lower3.1truept\vbox{\baselineskip0pt\lineskip0.5truept
22: \ialign{$\m@th#1\hfil##\hfil$\crcr#2\crcr\sim\crcr}}}
23: \catcode`\@=12
24:
25:
26: \shorttitle{CHANDRA OBSERVATION OF KEPLER'S SNR}
27: %\shortauthors{REYNOLDS ET AL. }
28:
29: \begin{document}
30:
31:
32: \title{A Deep {\sl Chandra} Observation of Kepler's Supernova Remnant:
33: A Type Ia Event with Circumstellar Interaction
34: }
35:
36: \author{Stephen P. Reynolds,\altaffilmark{1}
37: Kazimierz J. Borkowski,\altaffilmark{1}
38: Una Hwang, \altaffilmark{2}
39: John P.~Hughes,\altaffilmark{3}
40: Carles Badenes,\altaffilmark{3,4}
41: J.M.~Laming,\altaffilmark{5}
42: and J.M.~Blondin\altaffilmark{1}}
43:
44: \altaffiltext{1}{Department of Physics, North Carolina State University,
45: Raleigh NC 27695-8202; stephen\_reynolds@ncsu.edu}
46: \altaffiltext{2}{NASA/GSFC, Code 660, Greenbelt, MD 20771}
47: \altaffiltext{3}{Department of Physics and Astronomy, Rutgers University,
48: Piscataway, NJ 08854-8019}
49: \altaffiltext{4}{{\it Chandra} Fellow.}
50: \altaffiltext{5}{NRL, Code 7674 L, Washington, DC 20375-5320}
51:
52:
53: \begin{abstract}
54:
55: We present initial results of a 750 ks Chandra observation of the
56: remnant of Kepler's supernova of AD 1604. The strength and prominence
57: of iron emission, together with the absence of O-rich ejecta,
58: demonstrate that Kepler resulted from a thermonuclear supernova, even
59: though evidence for circumstellar interaction is also strong. We have
60: analyzed spectra of over 100 small regions, and find that they fall
61: into three classes. (1) The vast majority show Fe L emission between
62: 0.7 and 1 keV and Si and S K$\alpha$ emission; we associate these with
63: shocked ejecta. A few of these are found at or beyond the mean blast
64: wave radius. (2) A very few regions show solar O/Fe abundance
65: rations; these we associate with shocked circumstellar medium (CSM).
66: Otherwise O is scarce. (3) A few regions are dominated by continuum,
67: probably synchrotron radiation. Finally, we find no central point
68: source, with a limit $\sim 100$ times fainter than the central object
69: in Cas A. The evidence that the blast wave is interacting with CSM
70: may indicate a Ia explosion in a more massive progenitor.
71:
72: \end{abstract}
73:
74: \keywords{ISM: individual (\objectname{G4.5+6.8} (SN 1604)) ---
75: supernova remnants --- X-rays: ISM --- supernovae: general}
76:
77: \section{Introduction}
78:
79: Kepler's supernova of 1604 \citep{kepler1606} was the most recent
80: well-observed Galactic supernova. The supernova type has been
81: controversial for decades; the Galactic latitude ($b = 6\fdg8$)
82: implies a distance above the Galactic plane of 470 pc for a distance
83: of 4 kpc \citep{sankrit05}, arguing for a Type Ia (Population II)
84: progenitor, but optical observations \citep{blair91} show radiative
85: shocks in dense knots, suggestive of circumstellar material (CSM) shed
86: by the progenitor. Bandiera (1987) proposed that the progenitor was a
87: massive runaway star with a strong wind; Kepler's bright N rim
88: suggests a bow shock directed away from the Galactic plane.
89:
90: X-ray observations of young supernova remnants can give powerful
91: constraints on progenitor models and explosion mechanisms; whether
92: Kepler had a Type Ia or core-collapse origin, the detailed distribution
93: of ejected material and the composition of CSM could be used to test
94: SN models. We undertook a long observation of Kepler with the {\sl
95: Chandra} X-ray Observatory to study the composition and distribution
96: of ejecta and CSM, to search for any compact object, and to search
97: for convincing evidence of the SN type.
98:
99:
100: \section{Observations and Results}
101:
102:
103: We observed Kepler's SNR for 741 ks with the {\sl Chandra} X-ray
104: Observatory ACIS-S CCD camera as a Large Project in six installments
105: between April and July 2006. In all six, Kepler was positioned
106: (slightly differently) on the S3 back-illuminated chip. Data were
107: processed using CIAO v3.4 and calibrated using CALDB v3.1.0. A large
108: background region to the N of the remnant (covering most of the
109: remaining area on the S3 chip) was used for almost all spectra except
110: when local backgrounds were called for, as described below. Spectral
111: analysis was performed with XSPEC v.12 \citep{arnaud96}. We used
112: the nonequilibrium-ionization (NEI) v2.0 thermal models, based on the
113: APEC/APED spectral codes \citep{smith01} and augmented by addition of
114: inner-shell processes \citep{badenes06}.
115:
116: We obtained about $3 \times 10^7$ source counts, with fewer than 3\%
117: background (though some of those may be dust-scattered source
118: photons). S/N permitting, spectra were extracted between 0.3 and 8
119: keV. Images were smoothed using platelets with a default smoothing
120: parameter $\gamma = 1/2$ \citep{willett06}.
121:
122: Figure~\ref{totx} shows the full observation, with red for energies
123: between 0.3 and 0.72 keV (including primarily O), green for 0.72 --
124: 1.7 keV (primarily Fe L-shell transitions, also including Ne and Mg
125: K$\alpha$ complexes), and blue for 1.7 -- 8 keV (with a strong
126: contribution from Si K$\alpha$). Immediately evident is the absence
127: of dominantly soft-spectrum regions, and the hardness of emission at
128: the periphery. A few regions do show reddish colors; representative
129: spectra are discussed below.
130:
131: No central point source is obvious. The bright central knot is
132: resolved, and several fainter central knots, though perhaps
133: unresolved, also have spectra of shocked gas. A faint source could be
134: concealed under bright extended emission. Our conservative estimate
135: assumes a high total background of 860 cts per ACIS S3 pixel,
136: appropriate for the bright central band of emission. A 5$\sigma$
137: detection of a point source within the 80\%\ encircled power radius of
138: $0\farcs685$ (6 px area) would require 400 cts, or $6 \times 10^{-4}$
139: cts s$^{-1}$. The compact central object in Cas A \citep{pavlov00}, if
140: placed within Kepler and using the absorbing column density determined
141: below, would produce 0.18 cts s$^{-1}$. We conclude that we would
142: detect such a source even if its luminosity were $\sim 100$ times
143: fainter.
144:
145:
146: \begin{figure}
147: \centering
148: %\epsscale{1.15}
149: \epsscale{1.1}
150: %\plotone{f1p.png}
151: \includegraphics[width=3truein]{f1p.png}
152: \caption{Merged image between 0.3 and 8 keV. Red: 0.3 -- 0.72 keV;
153: green, 0.72 -- 1.7 keV; blue, 1.7 -- 8 keV. All three images were
154: smoothed using platelet smoothing \citep{willett06}. Image size
155: is $4\farcm7 \times 3\farcm9$. Regions from which spectra of Figs.~\ref{spectra1}
156: and \ref{spectra2}
157: were extracted are shown.
158: \vspace{-0.1truein}
159: \label{totx}
160: }
161: \end{figure}
162:
163:
164: For spectral analysis, we fixed the absorbing column density at $N_H =
165: 5.2 \times 10^{21}$ cm$^{-2}$, a value we found by fitting lineless
166: regions in the SE with an absorbed synchrotron model {\tt srcut} using
167: \citet{wilms00} abundances. We used the three-color image of
168: Fig.~\ref{totx} to aid in identifying small, relatively homogeneous
169: regions of contrasting spectral character. We identified and examined
170: about 100 regions. The vast majority show prominent lines of Fe
171: L-shell transitions, He-like Si, S, Ar, and Ca, and Fe K$\alpha$, and
172: a handful had contrasting properties. We found that the spectra can
173: be separated into three main classes. Over 90\% of the regions showed
174: Fe- and Si-rich spectra; we identify these with SN ejecta. A few
175: regions showed spectra with O and Mg emission consistent with solar-
176: or near-solar abundances; we identify these with CSM. Finally, a few
177: regions showed continuum-dominated spectra.
178:
179: Examples of ejecta spectra (regions 1 -- 3 on Fig.~\ref{totx})
180: are shown in Fig.~\ref{spectra1} and the
181: top spectrum (Fe L region, 6) in Fig.~\ref{spectra2}. In particular,
182: Fig.~\ref{spectra1} shows spectra of three outer knots, where we might
183: expect to find the lightest nucleosynthetic products at the largest
184: distances, specifically the O that should indicate a CC explosion.
185: All of them, however, show strong Si, S, and Fe L emission; prominent
186: Ar, Ca, and Fe K$\alpha$ lines are also present in the bright N
187: knot 1. The western and southern knots 2 and 3 have comparable Si and S
188: emission, but they differ in their Fe L-shell emission.
189:
190: Knot 3 is well modeled by a heavy-element plane shock
191: \citep{hendrick} with velocity $v_s$ of 1010 km s$^{-1}$ ([889, 1120]
192: km s$^{-1}$ 90\%\ confidence interval), shock ionization age $\tau =
193: 2.69 (2.32,3.18) \times 10^{10}$ cm$^{-3}$ s, Si/Fe and S/Fe ratios
194: (by number) of 2.64 (2.35,2.92) and 2.24 (1.59,2.87), respectively.
195: The knot is clearly composed of ejecta. A trace amount of Mg (Mg/Fe
196: $=0.052$ [0.003,0,10]) has been included, because it formally improves
197: the fit, but its presence is much in doubt in view of deficiencies in
198: Fe L atomic data. A synchrotron component (constrained by the absence
199: of radio flux from the knot [DeLaney et al.~2002]) improves the fit
200: quality.
201: %since there is no measurable radio flux at
202: %the location of the southern knot, we used a $3\sigma$ radio upper
203: %limit of 0.23 mJy in XSPEC {\tt srcut} model to arrive at a lower
204: %limit to the cut-off frequency $\nu_c$ of $2.40 (1.28,3.95) \times
205: %10^{17}$ Hz.
206: We assumed no collisionless heating at the shock; if such heating were
207: present, $v_s$ could be substantially less. Si and S are the primary
208: constituents of the southern knot, but a significant ($1/4$ by mass)
209: amount of Fe is present.
210:
211: In the western knot 2, the shock speed is lower ($v_s = 500 [360,830]$
212: km s$^{-1}$) and the ionization age is longer ($\tau = 4.8 [2.7,14]
213: \times 10^{10}$ cm $^{-3}$ s) but most significant are much higher
214: Si/Fe and S/Fe ratios, 11.6 (8.9,14.5) and 14.7 (10.2,21.1),
215: respectively. There is noticeably less Fe (6\%\ by mass) in this
216: knot; different Fe contents of knots 2 and 3
217: partially account for their different Fe L-shell emission strengths.
218: Finally, in knot 1 the relative strengths of Si, S, and Fe
219: L-shell emission are similar, so this is also a Si- and S-rich knot
220: containing Fe. (Because a simple plane shock model failed to provide a
221: satisfactory fit at high energies, we have not attempted a detailed
222: abundance analysis for knot 1).
223:
224: \begin{figure}
225: \centering
226:
227: %\includegraphics[height=3.7truein,width=2.3truein,angle=270]{f2.pdf}
228: \includegraphics[width=3truein]{f2.pdf}
229: \caption{Spectra of three outer knots shown on Fig.~\ref{totx}. Top
230: (red): Bright Si- and S-rich knot 1 in the N. Fe L shell, Fe K$\alpha$,
231: Ar, and Ca emission are also strong. Middle (blue) and
232: bottom (green): Knots in the S (3) and W (2), with plane shock model fits
233: shown.
234: %Different Fe abundances partially account for different Fe L
235: %shell strengths seen in their spectra.
236: %\label{spectraouterknots}
237: \label{spectra1}
238: }
239: \end{figure}
240:
241:
242:
243: The striking lack of obvious O features in the vast majority of the
244: regions we examined led us to search specifically for soft-spectrum
245: regions that might be expected to have a different composition. We
246: found a few (all showing a pink color in Fig.~\ref{totx}) with
247: discernible O features; two are shown in Fig.~\ref{spectra2}. They
248: constitute our second class of spectral region. The central knot has
249: optical and IR counterparts \citep*{blair91,blair07}, implying that it
250: is composed of ambient CSM. We obtained the spectrum shown in
251: Fig.~\ref{spectra2} after subtracting a local background from the
252: bright E-W band nearby. In the plane-shock model fit shown in
253: Fig.~\ref{spectra2}, we allowed the abundances of Ne, Mg, Si, and Fe
254: to vary (with S and Ni tied to Si and Fe, respectively). The poor
255: spectral resolution of CCD's and the uncertain calibration at low
256: energies, combined with deficiencies in Fe L atomic data, do not allow
257: us to determine reliable absolute abundances with respect to H and He,
258: or even a relative N abundance, but supersolar abundances for all
259: elements except N are ruled out. We assumed near solar composition by
260: setting the O abundance to its solar value, and determined relative
261: abundances with respect to O. We assumed a supersolar \citep[by a
262: factor of 3;][]{blair07} N abundance. We obtained a temperature $kT =
263: 0.64$ keV (0.59,0.70), $\tau = 2.7 (1.8, 3.6) \times 10^{11}$
264: cm$^{-3}$ s, and abundances with respect to solar of Ne = 0.64 (0.41,
265: 0.91), Mg = 0.83 (0.61, 1.1), Si (tied to S) = 1.4(1.0, 1.9), and Fe
266: (tied to Ni) = 1.1 (0.94, 1.4). When N was allowed to vary, the
267: best-fit N abundance is 3.2 (1.9, 4.6), where errors do not include
268: systematic uncertainties particularly important at low photon
269: energies. The model underestimates the continuum near the Si and S
270: K$\alpha$ lines, most likely because of the presence of
271: multitemperature plasma in the central knot, so abundances of these
272: elements may be somewhat overestimated. The central knot spectrum is
273: consistent with the same N-enhanced near-solar abundance composition
274: found from optical spectra of the CSM in Kepler.
275: Fig.~\ref{spectra2} also shows another soft-spectrum knot, 5; its
276: spectrum is consistent with that of the central knot.
277:
278: Finally, a few regions, primarily in the SE, showed featureless or
279: almost featureless spectra (one example, region 4, is shown in
280: Fig.~\ref{spectra2}). These are well-described by models of
281: synchrotron emission extending from the radio \citep{reynolds07}; they
282: form our third class of region, and will be described more fully in a
283: future publication.
284:
285:
286: \begin{figure}
287: \centering
288:
289: %\includegraphics[height=3.7truein,width=2.3truein,angle=270]{f3.pdf}
290: \includegraphics[width=3truein]{f3.pdf}
291: \caption{Spectra of four knots shown on Fig.~\ref{totx}.
292: Top (blue): Fe L region 6. No O or Mg features are evident.
293: Second from top (green): Soft-spectrum knot 5.
294: An inflection at 0.6 keV is
295: O Ly$\alpha$, and a feature at Mg K$\alpha$ is labeled.
296: Third from top at 0.8 keV (red): bright central knot. Note O and Mg features.
297: Bottom at 0.8 keV (magenta): synchrotron-dominated region 4 at rim in SE.
298: Only a faint Si feature is apparent.
299: \label{spectra2}
300: }
301: \end{figure}
302:
303:
304: \section{Kepler's progenitor type}
305:
306: Spectral regions 1, 2, 3, and 6 in our first, Si- and Fe-rich class
307: are almost certainly ejecta, of a composition completely inconsistent
308: with a CC event, but expected for a thermonuclear explosion. In fact,
309: they resemble Si- and Fe-rich compact ejecta knots in the vicinity of
310: the eastern blast wave in Tycho \citep{decourchelle01}. Ejecta knots
311: far ahead of the blast wave with such an advanced nuclear burning
312: composition have not been found so far in any CC SNRs, even Cas A;
313: except for the narrow jet, O is instead the most common element in the
314: outermost ejecta \citep{fesen06}. Only about 5\% of our regions fell
315: into the second class of roughly solar O/Fe and perhaps elevated N, as
316: expected for CSM. We can state with confidence that nowhere in Kepler
317: are there spectral regions consistent with the expected CC O/Fe ratio
318: of 70 \citep[e.g.,][]{iwamoto99}; it is very difficult to locate O
319: spectral features at all. Where strong Fe and Si indicate ejecta, O
320: is absent. However, the dense CSM shown in optical observations and
321: now identified in a few regions in X-rays is not expected for a Type
322: Ia progenitor. In view of the unusual nature of our result, we
323: collect the evidence on Kepler's progenitor type.
324:
325: {\bf 1. Integrated X-ray spectrum and ejecta.} The integrated spectrum
326: (which resembles region 6, Fig.~\ref{spectra2} top), as well as most
327: of our spectral regions, shows very strong Fe L, no apparent O or Mg,
328: strong Si and often strong Fe K emission. These elements are all
329: identified with ejecta. (O and Mg lines can be seen in the RGS
330: spectrum; Ballet 2004.) Even knots 1, 2, and 3 at the farthest
331: periphery show evidence of strong nuclear processing. The spectra
332: resemble the integrated spectrum of the Ia remnant of Tycho's SNR
333: (Badenes et al. 2006), but are totally unlike that of the young O-rich
334: CC remnant, 1E0102-70.4, which is dominated by O and Ne emission lines
335: (Sasaki et al.~2006). Previous work with {\it ASCA} and {\it
336: XMM-Newton} (Kinugasa \& Tsunemi 1999, Cassam-Chena\"{i} et al. 2004),
337: derived substantial masses of shocked Fe, 0.1-0.3 M$_\odot$
338: \citep{kinugasa99}.
339:
340: {\bf 2. Scarcity of O.} By contrast, there is little evidence for O
341: in Kepler's SNR. It is seen in only a few regions with spectra quite
342: different from the ejecta knots in having relatively weak Fe L
343: emission. Most of these regions are coincident with optical knots
344: which have been identified with CSM of roughly solar abundances except
345: for elevated N. The X-ray knots do not require enhanced N abundance,
346: but are consistent with the optical abundances. In particular, O/Fe
347: is roughly solar. In SN ejecta of Type Ia, O/Fe is expected to be
348: only 0.3-0.7 by number, whereas in CC explosions, it is about 70
349: (e.g., Iwamoto et al. 1999). None of the spectra we examined had the
350: high O/Fe ratios required for CC ejecta. (Several poorly-understood
351: luminous Type Ib/c SNe listed by \cite{richardson06} could be an
352: exception, because their O/Fe ratios are much less than typical for CC
353: explosions. Without detailed explosion models for these SNe confirming
354: their CC origin, we cannot exclude such possible extreme CC events on
355: the basis of the O/Fe ratio alone, but they are ruled out by arguments
356: below.) We conclude that regions with detectable O are CSM.
357:
358: {\bf 3. Central source.} No central point source is seen, with a
359: limit at least 100 times fainter than the source in Cas A.
360:
361: We find the X-ray evidence for Kepler's origin in a thermonuclear
362: event to be compelling. However, additional evidence can be found
363: from optical observations.
364:
365: {\bf 4. Balmer-dominated shocks.} \cite{blair91} discovered
366: substantial emission from nonradiative, Balmer-dominated shocks. Such
367: shocks demand the presence of partially neutral upstream material,
368: ruling out most CC scenarios. A hot (i.e., compact) progenitor of a
369: stripped CC Type Ib/c SN would have a strong ionizing flux before the
370: explosion, while a cooler star would be larger at the time of the
371: explosion, resulting in a large ionizing flux at shock breakout. The
372: presence of neutral H sets an upper limit to the progenitor radius
373: $R_*$ of $26 \alpha^{0.575} \left(E/10^{51} \ {\rm erg}\right)^{-0.23}
374: \left(M_{CSM}/M_\odot\right)^{0.575}
375: \left(M_{ej}/M_\odot\right)^{0.25} R_\odot$ \citep{chevalier05}, where
376: $13.6\alpha$ eV ($\alpha>1$) is the mean energy of an ionizing photon,
377: and $M_{CSM}$ and $M_{ej}$ are the shocked CSM and ejecta
378: masses. Since $M_{CSM}$ is only $\sim 1 M_\odot$ \citep{blair07},
379: explosions of red supergiants with $R_* \sim 600 R_\odot$ are ruled
380: out, including not only Type IIP and Type IIL SNe with full and
381: partially-stripped H envelopes, respectively, but also Type IIb SNe
382: like 1993J with only a residual ($0.2 M_\odot$) H envelope left on its
383: He core \citep{woosley94}. While an exotic stripped CC event,
384: resulting from a star with a residual hydrogen envelope of low ($\ll
385: 0.2 M_\odot$) mass but large radius, might avoid ionizing the CSM
386: before or during the explosion, such events become even less probable
387: than a Type Ia with CSM interaction. An additional problem with a
388: stripped CC progenitor is a large expected mechanical luminosity of
389: its stellar wind, since radio observations of circumstellar
390: interactions in Type Ib/c SNe are consistent with WR winds
391: \citep[e.g.,][]{chevalier07}. A powerful WR-like wind would have
392: accelerated and swept away any dense CSM to distances of many pc from
393: the SN, in obvious conflict with observations.
394:
395: {\bf 5. Light curve.} \cite{stephenson02} summarize the historical
396: data. Kepler's observed peak magnitude of $m \sim -3$ gives $M_V
397: \cong -18.8$ (using the observed $E_{B - V} = 0.9$ to obtain $A_V =
398: 2.8$), brighter than most CC events but typical for SNe Ia
399: \citep[e.g.,][]{hamuy96}. The light curve, though
400: resembling a normal Ia, could also describe a SN IIL (but see above).
401: SN 1987A-like events (fairly compact blue stars with massive H envelopes)
402: produce light curves that are far too broad, and are subluminous as
403: well.
404:
405:
406: \section{Discussion and conclusions}
407:
408: We believe that X-ray and optical evidence points compellingly to the
409: conclusion that Kepler resulted from a thermonuclear event. However,
410: our deep {\it Chandra} observations also confirm that Kepler's blast
411: wave is encountering material with at least solar metallicity,
412: coincident with optically emitting N-enhanced material -- most likely
413: circumstellar material lost by a fairly massive star, either the
414: progenitor of the white dwarf or the companion, or both. Interaction
415: with CSM is unexpected for Type Ia events. Some Type Ia SNe have
416: shown signs of CSM interaction in the form of narrow H lines, like SN
417: 2002ic \citep{hamuy03} and SN 2005gj \citep{aldering06}, but the
418: interpretation of these objects remains controversial (Benetti et
419: al.~2006; but see Prieto et al.~2007). Patat et al.~(2007) give
420: spectroscopic evidence that SN 2006X is a normal Type Ia SN with CSM
421: interaction. Evidence for CSM interaction in SNRs appears to be
422: relatively more widespread, and in some cases can be associated with
423: Type Ia events. Prominent Fe emission in DEM L238 and DEM L249, two
424: LMC SNRs, indicates a Type Ia origin. However, the brightness and
425: location of that Fe, at small radii with a large ionization age,
426: cannot be explained with standard Type Ia explosions into a uniform
427: ISM (Borkowski et al 2006); interaction with a nonuniform CSM seems to
428: be required. Relatively bright central emission dominated by Fe is
429: common in LMC SNRs \citep{nishiuchi01,vanderheyden04}, suggesting that
430: this class of SNRs is not a negligible population. N103B
431: \citep{lewis03} also shows strong Fe and Si emission and evidence for
432: CSM interaction. Kepler's SNR could just be the nearest example of a
433: class of thermonuclear supernovae interacting with dense CSM from the
434: progenitor system.
435:
436: The required CSM might indicate an origin in a younger, more massive
437: population. Three recent studies of extragalactic SNe Ia
438: \citep{mannucci06,sullivan06,aubourg07} propose the presence of a
439: ``prompt'' population (delay $\lapprox 5 \times 10^8$ yr) in addition
440: to an ``old'' population (delay $\sim$ Gyr) in later-type galaxies.
441: If stars with initial masses less than about 8 $M_\odot$ produce white
442: dwarfs, the minimum latency for a thermonuclear supernova would be
443: only about 50 Myr. A substantial number of remnants of Type Ia
444: supernovae should come from the ``prompt'' population.
445:
446: However, Kepler's origin remains problematic. The elevated CSM
447: abundances rule out a very old population, but the systemic velocity
448: \citep[$> 200$ km s$^{-1}$;][]{blair91} seems to indicate either a
449: halo population, inconsistent with the abundances, or a runaway,
450: probably inconsistent with a binary system. (The Galactic bulge
451: population would not extend to Kepler's Galactocentric distance of 4.5
452: kpc.) The total CSM mass of $\gapprox 1 M_\odot$ \citep{blair07}
453: requires too large an initial mass for an extreme halo population.
454: Could Kepler have resulted from the thermonuclear explosion of a
455: single star? Current conventional wisdom requires that all AGB stars
456: eject most of their envelopes to leave a sub-Chandrasekhar white
457: dwarf; if that ejection were occasionally impeded, a single-star
458: Type Ia supernova might result \citep[e.g.,][]{tout05}. If Kepler's
459: progenitor were a single star liberated from the Galactic plane by the
460: core-collapse explosion of its companion about 3 Myr ago, its maximum
461: lifetime would only be of order 50 Myr, also arguing for a more
462: massive progenitor.
463:
464: Our conclusion that Kepler resulted from a Type Ia event with CSM
465: interaction has important implications. Models of SN Ia progenitor
466: systems make widely varying predictions for the nature of any CSM
467: \citep[e.g.,][]{badenes07}; these predictions may now be testable.
468: Similarly, SN Ia explosion models can now be tested in detail against
469: the abundances and spatial distribution of elements in Kepler, and can
470: be contrasted with Tycho's SNR, apparently the result of a
471: conventional SN Ia encountering only low-density, more or less uniform
472: ISM \citep{badenes06}. The existence of a more massive subclass of SNe
473: Ia with properties that may be systematically different has obvious
474: important implications for cosmological use of SNe Ia, as it predicts
475: a redshift-dependent mix of the two types. Since the use of SNe Ia
476: for cosmology requires correction of peak magnitudes from light-curve
477: shapes with a relation calibrated locally, the application of that
478: same relation to more distant SNe Ia may raise concerns.
479:
480: \vspace{-0.18truein}
481: \acknowledgments
482: We gratefully acknowledge NASA support through {\it Chandra} Guest
483: Observer awards GO6-7061A, B, and C. CB was supported by NASA through
484: Chandra Postdoctoral Fellowship Award PF6-70046 issued by the CXC,
485: operated by SAO for and on behalf of NASA under contract NAS8-03060.
486:
487: %\appendix
488:
489:
490: \begin{thebibliography}{}
491:
492:
493: \bibitem[Aldering et al.(2006)]{aldering06}
494: Aldering et al. 2006, ApJ, 650, 510
495:
496: \bibitem[Arnaud(1996)]
497: {arnaud96}
498: Arnaud, K.~A. 1996, in Astronomical Data Analysis and Systems V, eds.
499: G.Jacoby \& J.Barnes, ASP Conf. Series, v.101, 17
500:
501: \bibitem[Aubourg et al.(2007)]{aubourg07}
502: Aubourg, E., et al.~2007, ApJ, submitted
503: (arXiv:0707.1328v2)
504:
505: \bibitem[Badenes et al.(2007)]{badenes07}
506: Badenes, C., Hughes, J. P., Bravo, E., \& Langer, N. 2007, ApJ, 662, 472
507:
508: \bibitem[Badenes et al.(2006)] {badenes06}
509: Badenes, C., et al.
510: 2006, ApJ, 645, 1373
511:
512: \bibitem[Ballet(2004)]{ballet04}
513: Ballet, J. 2004, in High-Resolution X-ray Spectroscopy with {\it XMM-Newton}
514: and {\it Chandra} (London: MSSL)
515:
516: \bibitem[Bandiera(1987)]{bandiera87}
517: Bandiera, R. 1987, ApJ, 319, 885
518:
519: \bibitem[Benetti et al.(2006)]{benetti06}
520: Benetti, S., et al.
521: 2006, ApJ, 653, L129
522:
523: \bibitem[Blair et al.(1991)]
524: {blair91}
525: Blair, W. P., Long, K., \& Vancura, O. 1991, ApJ, 366, 484
526:
527: \bibitem[Blair et al.(2007)]{blair07}
528: Blair, W.P., et al.~2007, ApJ, 662, 998
529:
530: \bibitem[Borkowski et al.(2006)]{borkowski06}
531: Borkowski, K.J., Hendrick, S.P., \& Reynolds, S.P.
532: 2006, ApJ, 652, 1259
533:
534: \bibitem[Cassam-Chena\"\i\ et al.(2004)]{cassam04}
535: Cassam-Chena\"\i, G., et al. 2004, A\&A, 414, 545
536:
537: \bibitem[Chevalier(2005)]{chevalier05}
538: Chevalier, R. A. 2005, ApJ, 619, 839
539:
540: \bibitem[Chevalier(2007)]{chevalier07}
541: Chevalier, R. A. 2007, RevMexAA (Serie de Conferencias), 30, 41
542:
543: \bibitem[Decourchelle et al.(2001)]{decourchelle01}
544: Decourchelle, A., et al. 2001, A\&A, 365, L218
545:
546: \bibitem[DeLaney et al.(2002)]{delaney02}
547: DeLaney, T., et al. 2002, ApJ, 580, 914
548:
549: \bibitem[Fesen et al.(2006)]{fesen06}
550: Fesen, R. A., et al. 2006, ApJ, 636, 859
551:
552: \bibitem[Hamuy et al.~(1996)]{hamuy96}
553: Hamuy, M., et al. 1996, AJ, 112, 2391
554:
555: \bibitem[Hamuy et al.(2003)]{hamuy03}
556: Hamuy, M., et al. 2003, Nature, 424, 651
557:
558: \bibitem[Hendrick(2003)]{hendrick}
559: Hendrick, S. P. 2003, Ph.D. Thesis, North Carolina State Univ.
560:
561: \bibitem[Iwamoto et al.(1999)]{iwamoto99}
562: Iwamoto, K., et al.
563: 1999, ApJS, 125, 439
564:
565: \bibitem[Kepler(1606)]{kepler1606}
566: Kepler, J. 1606, De Stella Nova in Pede Serpentarii
567: (Prague: Paul Sessi)
568:
569: \bibitem[Kinugasa \& Tsunemi(1999)]
570: {kinugasa99}
571: Kinugasa, K., \& Tsunemi, H. 1999, PASJ, 51, 239
572:
573: \bibitem[Lewis et al.(2003)]{lewis03}
574: Lewis, K.T., et al.
575: 2003, ApJ, 582, 770
576:
577: \bibitem[Mannucci et al.(2006)]{mannucci06}
578: Mannucci, F., Della Valle, M., \& Panagia, N. 2006, MNRAS, 370, 773
579:
580: \bibitem[Nishiuchi et al.~(2001)]{nishiuchi01}
581: Nishiuchi, M., et al.
582: 2001, PASJ, 53, 99
583:
584: \bibitem[Patat et al.(2007)]{patat07}
585: Patat, F., et al.~2007, Science, 317, 924
586:
587: \bibitem[Pavlov et al.~(2000)]{pavlov00}
588: Pavlov, G.G., et al.
589: 2000, ApJ, 531, L53
590:
591: \bibitem[Prieto et al.~(2007)]{prieto07}
592: Prieto, J.L., et al. 2007, AJ, submitted
593: (arXiv:0706.4088v1)
594:
595: \bibitem[Reynolds et al.~(2007)]{reynolds07}
596: Reynolds, S.P., et al.
597: 2007, BAAS, 39, 114
598:
599: \bibitem[Richardson et al.(2006)]{richardson06}
600: Richardson, D., Branch, D., \& Baron, E. 2006, AJ, 131, 2233
601:
602: \bibitem[Sankrit et al.(2005)]{sankrit05}
603: Sankrit, R., et al.
604: 2005, Adv. Sp. Res., 35, 1027
605:
606: \bibitem[Sasaki et al.(2006)]{sasaki06}
607: Sasaki, M., et al.
608: 2006, ApJ, 642, 260
609:
610: \bibitem[Smith et al.(2001)]{smith01}
611: Smith, R. K., et al.
612: J. C. 2001, ApJ, 556, L91
613:
614: \bibitem[Stephenson \& Green(2002)]{stephenson02}
615: Stephenson, F.R., \& Green, D.A.
616: 2002, Historical Supernovae and their Remnants (Oxford: Clarendon),
617: p.78
618:
619: \bibitem[Sullivan et al.(2006)]{sullivan06}
620: Sullivan, M., et al. 2006, ApJ, 648, 868
621:
622: \bibitem[Tout(2005)]{tout05}
623: Tout, C. A. 2005, in White Dwarfs: Cosmological and Galactic Probes,
624: ed. E.M.Sion, S.Vennes \& H.L.Shipman (Dordrecht:Springer), p.135
625:
626: \bibitem[van der Heyden et al.~(2004)]{vanderheyden04}
627: van der Heyden, K. J., et al.
628: 2004, A\&A, 421, 1031
629:
630: \bibitem[Willett(2007)]{willett06}
631: Willett, R. 2007, in Statistical Challenges in Modern
632: Astronomy IV, ed. G.J. Babu \& E.D. Feigelson (APS Conf. Ser. 371, in press);
633: http://www.ee.duke.edu/~willett/pubs.html
634:
635: \bibitem[Wilms et al.(2000)]{wilms00}
636: Wilms, J., Allen, A., \& McCray, R. 2000, ApJ, 542, 914
637:
638: \bibitem[Woosley et al.(1994)]{woosley94}
639: Woosley, S., et al.
640: 1994, ApJ, 429, 300
641:
642: \end{thebibliography}
643:
644: \end{document}
645:
646: \clearpage
647:
648: \figcaption{Merged image between 0.3 and 8 keV. Red: 0.3 -- 0.72 keV;
649: green, 0.72 -- 1.7 keV; blue, 1.7 -- 8 keV. All three images were
650: smoothed using platelet smoothing \citep{willett06}. Image size
651: is $4\farcm7 \times 3\farcm9$. Regions from which spectra of Figs.~\ref{spectra1}
652: and \ref{spectra2}
653: were extracted are shown.
654: \label{totx}
655: }
656:
657:
658: \figcaption{Spectra of three outer knots shown on Fig.~\ref{totx}. Top
659: (red): Bright Si- and S-rich knot 1 in the N. Fe L shell, Fe K$\alpha$,
660: Ar, and Ca emission are also strong. Middle (blue) and
661: bottom (green): Knots in the S (3) and W (2), with plane shock model fits
662: shown.
663: \label{spectra1}
664: }
665:
666: \figcaption{Spectra of four knots shown on Fig.~\ref{totx}.
667: Top (blue): Fe L region 6. No O or Mg features are evident.
668: Second from top (green): Soft-spectrum knot 5.
669: An inflection at 0.6 keV is
670: O Ly$\alpha$, and a feature at Mg K$\alpha$ is labeled.
671: Third from top at 0.8 keV (red): bright central knot. Note O and Mg features.
672: Bottom at 0.8 keV (magenta): synchrotron-dominated region 4 at rim in SE.
673: Only a faint Si feature is apparent.
674: \label{spectra2}
675: }
676:
677:
678:
679: \end{document}
680:
681:
682:
683:
684:
685:
686:
687: The integrated spectrum (which resembles the top L spectrum at the top
688: of Figure 3), as well as most of our spectral regions, show very
689: strong Fe L, no apparent O or Mg, strong Si and often strong Fe K
690: emission. These are all identified with ejecta. Even knots at the
691: farthest periphery show evidence of strong nuclear processing. The
692: spectra resemble the integrated spectrum of the Ia remnant of Tycho's
693: SNR (Badenes et al. 2006), but are totally unlike that of the young
694: O-rich CC remnant, E002-70.4, which is dominated by O and Ne emission
695: lines (Sasaki et al. 2006). The Fe K emission in Kepler is pervasive
696: throughout the remnant, as is expected for Ia remnants. Previous work
697: with ASCA and XMM-Newton derived substantial masses of Fe in Kepler's
698: SNR (Kinugasa \& Tsunemi 1999, Cassam-Chena\"{i} et al. 2004), at
699: 0.1-0.3 M$_\odot$.
700:
701:
702: The integrated X-ray spectrum of Kepler has been observed with {\sl
703: ASCA} \citep{kinugasa99} and {\sl XMM-Newton} \citep{cassam04}; it
704: shows prominent lines of Fe L, Si, S, and Fe K. Weaker lines of N, O,
705: and Mg can be seen in the {\sl XMM-Newton} RGS spectrum
706: \citep{ballet04}. \cite{kinugasa99} derived an iron mass of 0.1 --
707: 0.3 $M_\odot$ and asserted a Type Ia origin. Our long observation
708: with arcsecond spatial resolution allows us to identify and examine
709: small regions to try to separate and analyze shocked ejecta and CSM.
710: If ejecta can be clearly separated, their abundances should identify
711: the SN type: the expected Type Ia O/Fe ratio is 0.3 -- 0.7 by number,
712: in contrast to the CC ratio of about 70 \citep[e.g.][]{iwamoto99}.
713: