0812.0339/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[preprint2]{emulateapj}
3: 
4: \newcommand{\vdag}{(v)^\dagger}
5: \newcommand{\myemail}{katsuda@ess.sci.osaka-u.ac.jp}
6: 
7: 
8: 
9: %% You can insert a short comment on the title page using the command below.
10: 
11: %\slugcomment{Not to appear in Nonlearned J., 45.}
12: 
13: 
14: \shorttitle{Forward Shock Proper-Motions of Kepler's SNR}
15: \shortauthors{S. Katsuda, et al.}
16: 
17: 
18: \begin{document}
19: 
20: 
21: \title{Forward Shock Proper Motions of Kepler's Supernova Remnant} 
22: 
23: 
24: \author{S. Katsuda\altaffilmark{1}, H. Tsunemi\altaffilmark{1},
25:   H. Uchida\altaffilmark{1}, and M. Kimura\altaffilmark{1}
26: }
27: 
28: 
29: \altaffiltext{1}{Department of Earth and Space Science, Graduate School
30: of Science, Osaka University,\\ 1-1 Machikaneyama, Toyonaka, Osaka,
31: 560-0043, Japan; katsuda@ess.sci.osaka-u.ac.jp}
32: 
33: %\altaffiltext{2}{}
34: 
35: 
36: \begin{abstract}
37: 
38: The X-ray structure of Kepler's supernova remnant shows a rounded
39: shape delineated by forward shocks.  We measure proper-motions of the
40: forward shocks on overall rims of the remnant, by using archival {\it
41:   Chandra} data taken in two epochs with time difference of 6.09 yr.
42: The proper-motions of the forward shocks on the northern rim are
43: measured to be 0$^{\prime\prime}$.076 ($\pm$0$^{\prime\prime}$.032 $\pm$0$^{\prime\prime}$.016) -- 0$^{\prime\prime}$.11 ($\pm$0$^{\prime\prime}$.014
44: $\pm$0$^{\prime\prime}$.016) yr$^{-1}$, while those on the rest of the rims are
45: measured to
46: be 0$^{\prime\prime}$.15 ($\pm$0$^{\prime\prime}$.017
47: $\pm$0$^{\prime\prime}$.016) -- 0$^{\prime\prime}$.30
48: ($\pm$0$^{\prime\prime}$.048 $\pm$0$^{\prime\prime}$.016) 
49: yr$^{-1}$, here the first-term errors are statistical uncertainties
50: and the second-term errors are systematic uncertainties.  Combining
51: the best-estimated shock velocity of 1660$\pm$120\,km\,sec$^{-1}$
52: measured for Balmer-dominated filaments in the northern and central
53: portions of the remnant (Sankrit et al.\ 2005) with the proper-motions
54: derived for the forward shocks on the northern rim, we estimate the
55: distance of 3.3$^{+1.6}_{-0.4}$\,kpc to the remnant.  We measure
56: the expansion indices, $m$, (defined as $R \varpropto t^m$) to be
57: 0.47--0.82 for most of the rims.  These values are consistent
58: with those expected in Type-{\scshape I}a SN explosion models, in
59: which the ejecta and the circumstellar medium have power-law density
60: profiles whose indices are 5--7 and 0--2, respectively.  Also, we
61: shuold note
62: the slower expansion on the northern rim than that on the southern
63: rim.  This is likely caused by the inhomogeneous circumstellar medium;
64: the density of the circumstellar medium is higher in the north than
65: that in
66: the south of the remnant.  The newly estimated geometric center,
67: around which we believe the explosion point exists, is located at
68: $\sim$5$^{\prime\prime}$ offset in the north of the radio center.
69: 
70: \end{abstract}
71: \keywords{ISM: individual (Kepler's Supernova) --- shock waves ---
72:   supernova remnants --- X-rays: ISM} 
73: 
74: 
75: \section{Introduction}
76: 
77: Kepler's supernova remnant (SNR; SN 1604) is one of the historical
78: SNRs.  The X-ray structure of the remnant shows a clear circular shape
79: with a radius of about 100$^{\prime\prime}$.  The distance estimations
80: to the remnant have been scattered in a range of 3--5\,kpc (e.g.,
81: Green 1984; Schaefer 1994).  Although the far side of 5\,kpc supported
82: by H {\scshape I} observations (Reynoso \& Goss 1999) has been adopted
83: in recent literature, optical studies combining proper-motions with 
84: shock velocities of Balmer-dominated filaments have preferred a near
85: side of 2.9$\pm$0.4\,kpc (Blair et al.\ 1991) or
86: 3.9$^{+1.4}_{-0.9}$\,kpc (Sankrit et al.\ 2005).   
87: 
88: The SN type of Kepler's SNR is very interesting.  Based on
89: the light curve of the Kepler's SN, Baade (1943) classified this remnant
90: as a result of Type-{\scshape I}a SN.  {\it Ginga} spectrum from this 
91: remnant showed a strong Fe K line, supporting that the Kepler's SN was
92: a Type-{\scshape I}a SN (Hatsukade et al.\ 1990).  {\it ASCA} spectrum 
93: of this remnant showed strong K-shell lines from Si, S, and Fe,
94: and the measured relative abundances supported Type-{\scshape I}a
95: origin (Kinugasa \& Tsunemi 1999).  Recent deep {\it Chandra} 
96: observations detected strong emission lines from heavy elements almost
97: everywhere in the remnant, again supporting Type-{\scshape I}a
98: origin (Reynolds et al.\ 2007).  On the other hand, the density of the
99: ambient medium around the remnant was estimated to be so high (at
100: least 1\,cm$^{-3}$; 
101: e.g., Hughes \& Helfand 1985) at a large height of $\sim$470\,pc
102: (assuming the distance of 4\,kpc to the remnant) from the Galactic
103: plane.  The high density of the ambient medium was considered to be a
104: signature of circumstellar material (CSM) which was blown off from a
105: progenitor star as a stellar wind.  In addition to the existence of
106: the dense surroundings, Nitrogen-rich materials as a result of
107: CNO-processing in a massive progenitor star were detected in some
108: optical knots (van den Bergh \& Kamper 1977; Dennefeld 1982).  These
109: facts suggested that the Kepler's remnant was core-collapse in origin.
110: Also, Bandiera (1987) suggested a massive runaway star as a possible
111: progenitor of Kepler's SNR, since it was able to naturally account for
112: the presence of the CSM and the asymmetric structure of the remnant in
113: optical wavelength.  Recently, Blair et al.\ (2007) proposed that the
114: Kepler's SN was categorized as a Type-{\scshape I}a explosion in a
115: region with significant CSM, which was a small but growing class of
116: Type-{\scshape  I}a SNe named as Type-{\scshape I}a/IIn by Kotak et
117: al.\ (2004).  Due to the large expansion of Kepler's SNR, it is the
118: best target from which we can investigate the detailed preexisting
119: structures of the CSM for this rare kind of SNe. 
120: 
121: Here, we report proper-motions of the forward shocks on the overall
122: rims of Kepler's SNR, by using archival {\it Chandra} data.  We derive
123: the distance to the remnant, combining the proper-motions we measure
124: with the optically determined forward shock velocity (Sankrit et al.\ 
125: 2005).  Also, we present evolutional states in various portions of the
126: remnant, which gives us critical information on structures of the
127: surrounding CSM.
128: 
129: \section{Observations}
130: 
131: Kepler's SNR was observed on three epochs in 2000 (PI: Holt, S.), 2004
132: (PI: Rudnick, L.), and 2006 (PI: Reynolds, S.).  We use the first
133: (ObsID.\ 116) and the last (ObsID.\ 6715) observation to measure
134: proper-motions of the forward shocks of this remnant.  These
135: observations were, respectively, done in June 30th 2000 and August
136: 4--6th 2006, resulting in the time difference of 6.09 yr.  The entire
137: remnant was covered on the ACIS-S3 back-illuminated chip in both
138: observations.  We start our analysis from level 2 event files
139: processed with calibration data files in CALDB ver.\ 3.4.0 for ObsID.\
140: 116, ver.\ 3.2.2 for ObsID.\ 6715.  We exclude the high-background
141: periods for data from ObsID.\ 116, whereas there seems no significant
142: background flares for the data from ObsID.\ 6715 so that we reject no
143: data from the level 2 event file for this data set.  The resulting net 
144: exposure times for ObsID.\ 116 and 6715 are 37.8\,ks and 159.1\,ks,
145: respectively. 
146: 
147: To measure proper-motions of the forward shocks, we use an energy range
148: of 1.0--8.0\,keV, i.e., we do not use an energy range below 1.0\,keV
149: where building up contaminants on the detector significantly reduce
150: count rates for the second-epoch observation.  Figure~\ref{fig:image}
151: {\it left} shows the second-epoch image of the entire Kepler's SNR in
152: the energy range of 1.0--8.0\,keV.  
153: 
154: \section{Analysis and Results}
155: 
156: We register the two images by aligning them on four point sources that
157: are obviously seen in the vicinity of the remnant.  We determine the
158: positions of the four sources in both observations, employing {\tt
159:   wavdetect} software included in CIAO ver.\ 3.4.  We find a systematic
160: offset between the two images; the sense of the difference is that
161: the first-epoch image is south and west of the second.  We calculate
162: the error-weighted mean offset of the four point sources to be 
163: 0$^{\prime\prime}$.14 in right ascension and 0$^{\prime\prime}$.19
164: in declination.  Then, the first-epoch image is shifted in right
165: ascension by 0$^{\prime\prime}$.14 and in declination by
166: 0$^{\prime\prime}$.19 with respect to the second-epoch image.  Once
167: the two images are registered, we do not need to consider the absolute 
168: astrometric uncertainty of 0$^{\prime\prime}$.6 reported by the 
169: {\it Chandra} calibration team as systematic errors in our analysis.
170: We take account of 0$^{\prime\prime}$.1, i.e., the relative
171: astrometric uncertainty reported by the {\it Chandra} calibration
172: team, as systematic errors in the following proper-motion
173: measurements.  
174: 
175: Figure~\ref{fig:image} {\it right} shows a difference image between
176: two epochs after being registered and normalized to match the
177: count rates in the two epochs.  We clearly see a signature of the
178: expansion as positive emission (seen as white in the figure) with
179: negative emission (seen as black in the figure).  Note that the
180: horizontal stripes seen in the northern rim of the remnant are due
181: to bad columns on the ACIS-S3 chip.
182: 
183: To measure proper-motions of the forward shocks of the remnant, we
184: focus on 14 rectangular regions at the edge of the X-ray image
185: (see, Fig.~\ref{fig:image} {\it left}).  We then project the image
186: into one dimension so that we generate one-dimensional profile along
187: perpendicular direction to the shock motion (hereafter, radial
188: profile).  Each bin of the radial profile is spaced by
189: 0$^{\prime\prime}$.25.  We adjust the position angle for each
190: rectangular region as tangential as possible to the shock front in the
191: following way.  We generate radial profiles for position angle with
192: trial angles of 2$^\circ$ steps.  For each trial position angle, we
193: calculate 
194: \[I = \sum_{i} I_{i}^2,
195: \]
196: where $I_{i}$ is the intensity in each bin, $i$.  The position angle
197: with the largest value of $I$ represents the most tangential angle to the
198: shock front.  The 14 regions with the ``best'' position angles are
199: shown in Fig.~\ref{fig:image} {\it left}.  We plot example radial
200: profiles from Reg-4 and Reg-14 in Fig.~\ref{fig:profiles}, from which 
201: we can see apparent shifts between the two epochs.  To quantitatively
202: measure the shifts, the first-epoch profile is shifted in radius with
203: respect to the second-epoch profile, and $\chi^2$ is calculated from
204: the difference between the two profiles at each shift position (e.g.,
205: Katsuda et al.\ 2008).  In
206: this way, we obtain $\chi^2$ profiles as a function of shift
207: positions.  By applying a quadratic function for the $\chi^2$ profile,
208: we measure the best-shift position where the minimum $\chi^2$-value
209: ($\chi^2_\mathrm{ min}$) occurs.  90\% statistical uncertainties on
210: the best-shift are calculated using a criteria of $\chi^2_\mathrm{min}
211: + 2.7$.  Table.~\ref{tab:summary} summarizes the best-shift positions
212: between the two observations, $\chi^2_\mathrm{min}$ per degrees of
213: freedom, and proper motions for all the regions indicated in
214: Fig.~\ref{fig:image} {\it left}.  The azimuthal angle for each region, 
215: which is measured counterclockwisely from the radio center
216: at $\alpha = $ 17$^{\mathrm h}$30$^{\mathrm m}$41$^{\mathrm s}$.5,
217: $\delta = -21^{\circ}29^{\prime}23^{\prime\prime}$ (J2000; Matsui et  
218: al.\ 1984), is also listed in the table.  
219: 
220: 
221: We find that the proper-motions vary from location to location.
222: Plotting the proper-motion derived in each region as a function of
223: azimuthal angle in Fig.~\ref{fig:azimuth}, we find a trend that the
224: proper-motions derived in the northern regions (i.e., Regs-1, 2, 13,
225: and 14) are slower than those derived in the rest of the regions.  
226: 
227: \section{Discussion}
228: 
229: \subsection{Distance to the Remnant}
230: 
231: We have measured proper-motions of the forward shocks on overall
232: rims of Kepler's SNR.  If we combine them with the shock
233: velocities, the distance to the remnant can be determined.  The
234: velocities of the forward shocks associated with Balmer-dominated
235: filaments, which were located at the northern rim and the central
236: portion of the remnant, were determined to be
237: 1670--2800\,km\,sec$^{-1}$ (Fesen et al.\ 1989) or
238: 1550--2000\,km\,sec$^{-1}$ (Blair et al.\ 1991) from their H$\alpha$  
239: emission line widths.  Under the assumption of little or no
240: temperature equilibration between electrons and protons, Sankrit et
241: al.\ (2005) determined the 
242: best-estimated shock velocity to be 1660$\pm$120\,km\,sec$^{-1}$.
243: Since the best-estimated shock velocity represents the average
244: velocity of the forward shocks with Balmer-dominated filaments seen in
245: the northern and the central portion of the remnant, we should combine it
246: with an error-weighted mean proper-motion measured in the northern
247: regions (i.e., Regs-1, 2, 13, 14).  Then, the best-estimated distance
248: to the remnant is determined to be 3.3
249: ($v$/1660\,km\,sec$^{-1}$)($\mu$/0$^{\prime\prime}$.107
250: yr$^{-1}$)$^{-1}$ kpc.  We can estimate a distance range of
251: 2.9--4.9\,kpc to the remnant, considering the variation of the
252: proper-motions derived in the four northern regions of 
253: 0$^{\prime\prime}$.076--0$^{\prime\prime}$.11 yr$^{-1}$ as well as the
254: uncertainty of the shock velocity.  While the best-estimated distance
255: determined here is less than $\sim$5\,kpc (Reynoso \& Goss 1999), it
256: is well within the values previously measured based on the combination
257: of the proper-motion and the shock-velocity of the Balmer-dominated
258: filaments: from 2.9\,kpc (Blair et al.\ 1991) to 3.9\,kpc (Sankrit et
259: al.\ 2005).  
260: 
261: \subsection{Asymmetry of the Forward Shock Velocity}
262: 
263: We find velocity asymmetries of the forward shocks of Kepler's SNR:
264: shock velocities on the northern rim are 1.5--3 times slower than
265: those on the other rims.  If we assume the pressure equilibrium in the
266: remnant, the velocity contrast of 1.5--3 requires a density contrast
267: of $\sim$2--9.  Recently, Blair et al.\ (2007) estimated that the
268: preshock density in the north of the remnant is $\sim$4--9 times
269: higher than that in the south, based on brightness variations observed
270: between the northern and southern rims of the remnant at 24\,$\mu$m.
271: Therefore, the slower expansion observed on the northern rim than
272: those on the rest of the rims seems to be well explained by the
273: density contrast suggested from the observation at 24\,$\mu$m.   
274: 
275: It is interesting that Kepler's SNR appears quite round in spite of
276: the asymmetric velocities of the forward shocks.  One possibility to
277: explain this feature is that the forward shock encountered a dense gas
278: on the northern rim so recently that we can hardly see apparent
279: deceleration of the shock from morphological point of view.  
280: However, there have been no reports of such a very recent shock
281: deceleration so far.  (In addition, we can not obtain at least
282: strong evidence that the forward shock on the northwestern rim shows
283: recent significant deceleration; Sankrit et al.\ [2005] 
284: measured a proper-motion of a Balmer-dominated filament positionally
285: coincident with the forward shock in Reg-13 from observations
286: performed on two epochs in 1987 and 2003 to be
287: 0$^{\prime\prime}$.089$\pm$0$^{\prime\prime}$.009 yr$^{-1}$ that is
288: consistent with our estimation of the proper-motion of
289: 0$^{\prime\prime}$.076 [$\pm$0$^{\prime\prime}$.032
290: $\pm$0$^{\prime\prime}$.016] yr$^{-1}$ determined in between 2000 and
291: 2006).  We propose another possibility that the expansion center of
292: the remnant is located at a relatively northern position compared to
293: the geometric center 
294: determined by the entire remnant.  In fact, the outer edges of the
295: X-ray structure seem to be outlined by two (ideally concentric)
296: circles whose centers are located at the north of the radio center: one with
297: a relatively small radius outlines the northern-half edge of the
298: remnant, whereas the other with a relatively large radius outlines
299: the southern-half edge.  We estimate the best-fit circle for either
300: the northern-half edge or the southern-half edge independently.  Here, 
301: we define the northern half in a range of azimuthal angles from
302: $-$50$^\circ$ (or 310$^\circ$) to 45$^\circ$, whereas the
303: southern half in a range of azimuthal angles from 95$^\circ$ to
304: 258$^\circ$, such that the northern half covers the regions with
305: proper-motions of $\sim$0$^{\prime\prime}$.1 yr$^{-1}$, 
306: whereas the southern half covers the other regions with proper-motions
307: of above $\sim$0$^{\prime\prime}$.2 yr$^{-1}$.  We define the edge
308: of the X-ray extent (hereafter, X-ray boundary) as contours of
309: 5 counts per 0$^{\prime\prime}$.492-sided pixel in the second-epoch
310: 0.3--8.0\,keV band image.  We ignore X-ray boundaries showing
311: apparent deviations from circular curvatures, in order not to yield
312: misleading geometric centers and radii.  The X-ray
313: boundaries used to estimate the best-fit circles are drawn by
314: white curves (contours) in Fig.~\ref{fig:center}.  Assuming 
315: the geometric center, we calculate the distance, $R_i$, between
316: each pixel, $i$, on the X-ray boundaries and the geometric center.
317: For various trial geometric centers, we calculate 
318: \[K = \sum_{i} (R - R_{i})^2,
319: \]
320: where $R$ is a variable parameter representing a radius.  We can
321: derive the best-fit radius, $R$, as well as the geometric center
322: at the minimum of the $K$-value.  The best-fit circles representing the
323: X-ray boundaries for the northern and southern halves are centered at
324: $\alpha = $ 17$^{\mathrm h}$30$^{\mathrm m}$41$^{\mathrm s}$.4, $\delta = 
325: -21^{\circ}29^{\prime}19^{\prime\prime}$ (J2000) with a radius of
326: 93$^{\prime\prime}$, and at $\alpha = $ 17$^{\mathrm h}$30$^{\mathrm
327:   m}$41$^{\mathrm s}$.9, $\delta =
328: -21^{\circ}29^{\prime}16^{\prime\prime}$ (J2000) 
329: with a radius of 120$^{\prime\prime}$, respectively.  Parts of these
330: two circles are indicated in Fig.~\ref{fig:center} as white dashed
331: pie. We see that both geometric centers estimated are
332: shifted by about 5$^{\prime\prime}$ toward the north of the remnant
333: from the radio center (see, Fig.~\ref{fig:center}), which
334: results in a smaller radius for the circle representing the
335: northern-half X-ray boundary than that for the southern-half X-ray
336: boundary; the ratio of the radius is derived to be about 3:4.  
337: Although we cannot derive uncertainties on the center positions and
338: radii in the least square method used above, we might estimate
339: uncertainties on these parameters from the $\chi^2$ method by
340: introducing fake errors on the data ($R_i$).  We introduce the fake 
341: errors of 1\,\% for the northern half and 3.5\,\% for the southern
342: half, respectively, so that we can derive the reduced $\chi^2$-value 
343: of $\sim$1.  The 90\% uncertainties, i.e., $\chi^2 <
344: \chi^2_\mathrm{min} + 6.25 $ as appropriate for three interesting
345: parameters, are estimated to be $\pm$0$^{\prime\prime}$.5 or
346: $\pm$1$^{\prime\prime}$.5 (in right ascension), and
347: $\pm$1$^{\prime\prime}$.5 or $\pm$2$^{\prime\prime}$ (in declination)
348: for the northern or southern halves, respectively.  Therefore, 
349: the derived offset of $\sim$5$^{\prime\prime}$ from
350: the radio center seems to be significant, although we cannot strongly
351: state the significance without knowing the uncertainty of the radio center.
352: Hydrodynamic simulations of a remnant expanding into a medium with a
353: density gradient (Dohm-Palmer \& Jones 1996) show that the shock
354: outline remains roughly circular, while the center of the best-fit
355: curvature can move away from the true explosion location by as much as
356: 10 -- 15\% of the remnant radius.  The offsets between geometric
357: centers estimated here and the radio center (Matsui et al.\ 1984) are
358: roughly 5\% of the radius.  Therefore, from theoretical point of view,
359: our newly determined geometric centers can be possible true explosion
360: locations of Kepler's SNR.  We believe that the true explosion point
361: is around the geometric centers estimated here.
362: 
363: We consider the expansion center of Kepler's SNR at $\alpha =
364: $17$^{\mathrm h}$30$^{\mathrm m}$41$^{\mathrm s}$.6, $\delta = 
365: -21^{\circ}29^{\prime}17^{\prime\prime}$ (J2000) that is the center
366: point between the two center points of the circles representing the
367: northern-half and southern-half X-ray boundaries.  Then, the expansion
368: rate can be calculated by dividing the
369: proper-motion value in each region by the distance between the shock
370: front and the expansion center.  Combined with the age of the remnant
371: of 400 yr, the expansion index, $m$, is also calculated, where the
372: expansion of the forward shock of SNRs can be expressed as $R
373: \varpropto t^m$ (i.e., the remnant's radius is assumed to evolve as a
374: power law with age; see, e.g., Woltjer 1972).  Table.~\ref{tab:param}
375: summarizes these parameters and Fig.~\ref{fig:azimuth2} shows the
376: expansion indices as a function of azimuthal angle.  SNRs cease
377: being in pure free expansion ($m=1$) after a few years, as they interact with
378: circumstellar or interstellar gas.  Then, a reverse shock into the
379: ejecta is formed, and remnants evolve in what is usually called an
380: ejecta-driven phase.  In spherical symmetry, if both the ejecta and the CSM
381: have power-law density profiles characterized by $\rho \varpropto
382: r^{-n}$ and $\rho \varpropto r^{-s}$, respectively, the evolution becomes
383: self-similar and is given by $R \varpropto t^{(n-3)/(n-s)}$, i.e., the
384: expansion index, $m$, can be written as $(n-3)/(n-s)$ (e.g., Chevalier
385: 1982).  Here, the value of $s$ is 0 for a uniform CSM, and is 2 for a
386: constant wind velocity from the progenitor star.   
387: In general, the $s$-value is expected to be 2 for core-collapse
388: SNe, since massive stars which results in core-collapse SNe produce
389: stellar wind before SN explosions, while both $s = 0$ and $s = 2$ can 
390: happen for Type-{\scshape I}a SNe as mentioned in Section 1.  As for
391: $n$-values, $9 \lesssim n \lesssim 12$ is expected for a core-collapse
392: origin (e.g., Chevalier 1992), while $5 \lesssim n \lesssim 7$ for a
393: Type-Ia origin (e.g., Chevalier 1982).  Therefore, for core-collapse
394: SNe, the value of $m$ is expected to be greater than 6/7 (=0.86).  On
395: the other hand, for Type-{\scshape I}a SNe, the value of $m$ ranges
396: from 0.4 to 0.8.  Our measured $m$-values at all regions except for
397: regions 6 and 13 range from 0.47 to 0.82, which is consistent with
398: that expected from Type-{\scshape I}a SNe.  We find significant
399: variations of $m$-values in this 
400: remnant, which suggests complicated CSM structures around the
401: remnant.  It should be remarked that the northern rim shows
402: slower expansions than the southern rim does.  The north-south density 
403: asymmetry of the CSM suggested from variations of surface brightness
404: results in the variation of $m$-values between the northern rim and
405: the southern rim.
406: 
407: It is worth noting how the velocity difference of the forward shock
408: affects spectral features.  As already noticed by {\it XMM-Newton}
409: (Cassam-Chena$\ddot{\mathrm i}$ et al.\ 2004) and {\it Chandra}
410: (Reynolds et al.\ 2007) observations, a spectrum extracted around
411: Reg-6 on the southeastern rim is non-thermal (synchrotron) in origin,
412: whereas a spectrum around Reg-13 on the northwestern rim is thermal in
413: origin.  We can clearly see the spectral difference in
414: Fig.~\ref{fig:spec}.  This is considered as a result that the faster
415: shock (4700\,km\,sec$^{-1}$ at a distance of 3.3\,kpc) as well as the
416: lower ambient density suggested by lower surface brightness in Reg-6
417: produces the synchrotron emission more efficiently than the slower
418: (1200\,km\,sec$^{-1}$ at a distance of 3.3\,kpc) shock as well as
419: higher ambient density in Reg-13 does. 
420: 
421: We should note previous X-ray expansion measurements based on {\it
422: Einstein} and {\it ROSAT} observations performed by Hughes (1999).
423: The mean expansion rates in the entire remnant was derived to be
424: $\sim$0.24\% yr$^{-1}$.  This value is larger than that
425: estimated in our analysis; only one region (Reg-6) shows such a rapid
426: expansion.  Furthermore, he measured expansion rates as a function of
427: azimuthal angle, by comparing long radial profiles from the geometric
428: center to the X-ray boundaries between two {\it ROSAT} HRI
429: observations.  The expansion rates measured in the northern portion of
430: the remnant did not show lower values than those in the other
431: portions.  This also conflicts to our results.  These discrepancies
432: might come from the different method between Hughes (1999) and this
433: work.  Hughes (1999) measured radially averaged proper-motions of the
434: remnant, on the other hand, we measure proper-motions for the very
435: edge of the remnant, i.e., the forward shocks themselves.  Therefore,
436: the faster expansion rates derived in Hughes (1999) than those in this
437: work suggests that plasmas in the inner remnant show larger expansion
438: rates than those of the forward shocks.  Such a situation might
439: indicate recent rapid deceleration of the forward shocks.  However, as
440: mentioned above, we have not yet observed (at least strong) such 
441: indications so far.  Further detailed proper-motion measurements in
442: the inner remnant are strongly required to reveal the reason of the
443: discrepancies between the expansion rate by Hughes (1999) and that in
444: this work.  
445: 
446: \section{Conclusions}
447: 
448: We have measured proper-motions of the forward shocks on the overall 
449: rims of Kepler's SNR, using the archival {\it Chandra} data.
450: 
451: The expansion indices measured at various parts of the rim supports
452: the Type-{\scshape I}a SN which has the situation that the ejecta and
453: the CSM, respectively, have power-law density profiles with indices
454: of 5--7 and 0--2 rather than the core-collapse SN. 
455: 
456: We find that the shock velocities are
457: asymmetric: the shock velocities on the northern rim are 1.5--3 times
458: slower than those on the rest of the remnant.  We attribute this
459: asymmetry to the density inhomogeneities of the CSM surrounding the
460: remnant.  The shape of the X-ray boundary of the remnant as well as
461: the inhomogeneous CSM structures lead us to consider that the
462: expansion center is located at $\sim$5$^{\prime\prime}$ offset in the
463: north of the radio center.  
464: 
465: \acknowledgments
466: 
467: This work is partly supported by a Grant-in-Aid for Scientific Research
468: by the Ministry of Education, Culture, Sports, Science and Technology
469: (16002004).  S.K.\ is supported by JSPS Research Fellowship for Young
470: Scientists.  
471: 
472: \clearpage
473: 
474: \begin{thebibliography}{}
475: \bibitem[Baade 1943]{Baade1943}
476:   Baade, W. 1943, ApJ, 97, 119
477: \bibitem[Bandiera1987]{Bandiera1987}
478:   Bandiera, R. 1987, ApJ, 319, 885
479: \bibitem[Blair et al.\ (1991)]{Blair1991}
480:   Blair, W. P., Long, K. S., \& Vancura, O. 1991, ApJ, 366, 484
481: \bibitem[Blair et al.\ 2007]{Blair2007}
482:   Blair, W. P., et al.\ 2007, ApJ, 662, 998 
483: \bibitem[Chevalier 1982]{Chevalier1982}
484:   Chevalier, R. A. 1982, ApJ, 259, L85
485: \bibitem[Chevalier 1992]{Chevalier1992}
486:   Chevalier, R. A. 1992, ApJ, 394, 599
487: \bibitem[Dohm-Palmer \& Jones 1996]{Dohm-Palmer1982}
488:   Dohm-Palmer, R. C., \& Jones, T. W. 1996, ApJ, 471, 279
489: \bibitem[Cassam-Chenaa$\ddot{\mathrm i}$ et al.\ 2004]{Cassam-Chenai2004}
490:   Cassam-Chena$\ddot{\mathrm i}$, G., Decourchelle, A., Ballet, J.,
491:   Hwang, U., Hughes, J. P., \& Petre, R. 2004, A\&A, 414, 545
492: \bibitem[Dennefled 1982]{Dennefeld1982}
493:   Dennefeld, M. 1982, A\&A, 112, 215
494: \bibitem[Dickel et al.\ 1988]{Dickel1988}
495:   Dickel, J. R., Sault, R., Arendt, R. G., Korista, K. T., \& Matsui,
496:   Y. 1988, ApJ, 330, 254
497: \bibitem[Fesen et al.\ 1989]{Fesen1989}
498:   Fesen, R. A., Becker, R. H., Blair, W. P., \& Long, K. S. 1989, ApJ,
499:   338, L13
500: \bibitem[Green 1984]{Green1984}
501:   Green, D. A., 1984, MNRAS, 209, 449
502: \bibitem[Hatsukade et al.\ 1990]{Hatsukade1990}
503:   Hatsukade, I., Tsunemi, H., Yamashita, K., Koyama, K., Asaoka, Y.,
504:   \& Asaoka, I. 1990, PASJ, 42, 279
505: \bibitem[Hughes \& Helfand]{Hughes1985}
506:   Hughes, J. P., \& Helfand, D. J., 1985, ApJ, 291, 544
507: \bibitem[Hughes 1999]{Hughes1999}
508:   Hughes, J. P. 1999, ApJ, 527, 298
509: \bibitem[Katsuda et al.\ 2008]{Katsuda2008}
510:   Katsuda, S., Tsunemi, H., Mori, K. ApJ, 678, L35
511: \bibitem[Kinugasa \& Tsunemi]{Kinugasa1999}
512:   Kinugasa, K., \& Tsunemi, H. 1999, PASJ, 51, 239
513: \bibitem[Kotak et al.\ 2004]{Kotak2004}
514:   Kotak, R., Meikle, W. P. S., Adamson, A., \& Leggett, S. K. 2004,
515:   MNRAS, 354, L13
516: \bibitem[Matsui et al.\ 1984]{Matsui1984}
517:   Matsui, Y., Long, K. S., Dickel, J. R., \& Greisen, E. W. 1984, ApJ,
518:   287, 295
519: \bibitem[Reynolds et al.\ 2007]{Reynolds2007}
520:   Reynolds, S. P., et al. 2007, ApJ, 668, L135
521: \bibitem[Reynoso \& Goss 1999]{Reynoso1999}
522:   Reynoso, E. M., \& Goss, W. M. 1999, AJ, 118, 926
523: \bibitem[Sankrit et al.\ 2005]{Sankrit2005}
524:   Sankrit, R., Blair, W. P., Delaney, T., Rudnick, L., Harrus, I. M.,
525:   \& Ennis, J. A. 2005, Adv. Space Res. 35, 1027
526: \bibitem[Schaefer 1994]{Schaefer1994}
527:   Schaefer, B. E. 1994, ApJ, 426, 493
528: \bibitem[van den Bergh \& Kamper 1977]{vandenBergh1977}
529:   van den Bergh, S., \& Kamper, K. W. 1977, ApJ, 218, 617
530: \bibitem[Woltjer 1972]{Woltjer1972}
531:   Woltjer, L. 1972, ARA\&A, 10, 129
532: %\bibitem[]{}
533: 
534: \end{thebibliography}
535: 
536: %\clearpage
537: 
538: \begin{deluxetable}{ccccc}
539: \tabletypesize{\scriptsize}
540: 
541: \tablecaption{Summary of Proper-Motion Measurements}
542: \tablewidth{0pt}
543: \tablehead{
544: \colhead{Region} &\colhead{Azimuth (deg)} &\colhead{$\chi^2_\mathrm{
545:     min}$/d.o.f.} &\colhead{Shift between the two epochs (arcsec)}  &
546: \colhead{Proper motion (arcsec yr$^{-1}$)}  
547: }
548: \startdata
549: Reg-1 & 16 & 1.68 & 0.63 ($\pm$0.09 $\pm$0.10) & 0.109 ($\pm$0.013 $\pm$0.016)\\
550: Reg-2 & 32 & 0.54 & 0.66 ($\pm$0.08 $\pm$0.10) & 0.104 ($\pm$0.015 $\pm$0.016)\\
551: Reg-3 & 75 & 0.85 & 0.91 ($\pm$0.11 $\pm$0.10) & 0.149 ($\pm$0.017 $\pm$0.016)\\
552: Reg-4 & 96 & 1.47 & 1.16 ($\pm$0.09 $\pm$0.10) & 0.191 ($\pm$0.015 $\pm$0.016)\\
553: Reg-5 & 102 & 2.27 & 1.26 ($\pm$0.10 $\pm$0.10)& 0.206 ($\pm$0.016 $\pm$0.016)\\
554: Reg-6 & 135 & 1.56 & 1.84 ($\pm$0.29 $\pm$0.10)& 0.302 ($\pm$0.048 $\pm$0.016)\\
555: Reg-7 & 155 & 1.71 & 1.08 ($\pm$0.18 $\pm$0.10)& 0.178 ($\pm$0.030 $\pm$0.016)\\
556: Reg-8 & 172 & 1.39 & 1.48 ($\pm$0.16 $\pm$0.10)& 0.242 ($\pm$0.025 $\pm$0.016)\\
557: Reg-9 & 230 & 1.80 & 1.25 ($\pm$0.13 $\pm$0.10)& 0.206 ($\pm$0.021 $\pm$0.016)\\
558: Reg-10 & 242 & 1.00 & 0.99 ($\pm$0.13 $\pm$0.10)& 0.162 ($\pm$0.022 $\pm$0.016)\\
559: Reg-11 & 250 & 1.25 & 1.29 ($\pm$0.24 $\pm$0.10)& 0.212 ($\pm$0.040 $\pm$0.016)\\
560: Reg-12 & 258 & 1.06 & 1.10 ($\pm$0.17 $\pm$0.10)& 0.180 ($\pm$0.027 $\pm$0.016)\\
561: Reg-13 & 319 & 0.89 & 0.46 ($\pm$0.19 $\pm$0.10)& 0.076 ($\pm$0.032 $\pm$0.016)\\
562: Reg-14 & 345 & 1.06& 0.70 ($\pm$0.08 $\pm$0.10)& 0.114 ($\pm$0.014 $\pm$0.016)\\
563: 
564: %\hline
565: \enddata
566: 
567: \tablecomments{First-term errors represent 90\% statistical
568:   uncertainties and second-terms errors represent systematic
569:   uncertainties. } 
570: \label{tab:summary}
571: \end{deluxetable}
572: 
573: \begin{deluxetable}{ccccc}
574: \tabletypesize{\scriptsize}
575: 
576: \tablecaption{Summary of Expansion Rates and Expansion Indices}
577: \tablewidth{0pt}
578: \tablehead{
579: \colhead{Region} &\colhead{Azimuth (deg)} &\colhead{Exapnsion Rate (\%)}
580: &\colhead{X-Ray Expansion Index}  & \colhead{Radio
581:   Expansion Index$^{\mathrm a}$}  
582: }
583: \startdata
584: Reg-1 & 16 & 0.12 ($\pm$0.01 $\pm$0.02)& 0.49 ($\pm$0.06 $\pm$0.07)& 0.45\\
585: Reg-2 & 32 & 0.12 ($\pm$0.02 $\pm$0.02)& 0.47 ($\pm$0.07 $\pm$0.07)& 0.45\\
586: Reg-3 & 75 & 0.16 ($\pm$0.02 $\pm$0.02)& 0.63 ($\pm$0.07 $\pm$0.07)& 0.65\\
587: Reg-4 & 96 & 0.17 ($\pm$0.01 $\pm$0.01)& 0.68 ($\pm$0.05 $\pm$0.06)& ---\\
588: Reg-5 & 102 & 0.17 ($\pm$0.01 $\pm$0.01)& 0.70 ($\pm$0.05 $\pm$0.05)& ---\\
589: Reg-6 & 135 & 0.25 ($\pm$0.04 $\pm$0.01)& 0.98 ($\pm$0.16 $\pm$0.05)& ---\\
590: Reg-7 & 155 & 0.16 ($\pm$0.03 $\pm$0.01)& 0.63 ($\pm$0.11 $\pm$0.06)& ---\\
591: Reg-8 & 172 & 0.20 ($\pm$0.02 $\pm$0.01)& 0.82 ($\pm$0.08 $\pm$0.05)& ---\\
592: Reg-9 & 230 & 0.18 ($\pm$0.02 $\pm$0.01)& 0.71 ($\pm$0.07 $\pm$0.06)& ---\\
593: Reg-10 & 242 & 0.14 ($\pm$0.02 $\pm$0.01)& 0.55 ($\pm$0.08 $\pm$0.06)& 0.55\\
594: Reg-11 & 250 & 0.18 ($\pm$0.03 $\pm$0.01)& 0.72 ($\pm$0.14 $\pm$0.05)& 0.55\\
595: Reg-12 & 258 & 0.16 ($\pm$0.02 $\pm$0.01)& 0.62 ($\pm$0.09 $\pm$0.06)& 0.55\\
596: Reg-13 & 319 & 0.08 ($\pm$0.04 $\pm$0.02)& 0.34 ($\pm$0.14 $\pm$0.07)& 0.35\\
597: Reg-14 & 345 & 0.12 ($\pm$0.02 $\pm$0.02)& 0.50 ($\pm$0.06 $\pm$0.07)& 0.35\\
598: 
599: %\hline
600: \enddata
601: \tablecomments{First-term errors represent 90\% statistical uncertainties and
602:   second-term errors represent systematic uncertainties.  
603:   $^{\mathrm a}$Dickel et al.\ (1988) without particulary
604:   uncertain data.} 
605: \label{tab:param}
606: \end{deluxetable}
607: 
608: 
609: \begin{figure}
610: \includegraphics[angle=0,scale=0.45]{f1.eps}\hspace{1cm}
611: \caption{{\it Left}: {\it Chandra} 1.0--8.0\,keV band image obtained
612:   in the second-epoch observation.  The image is binned by
613:   0$^{\prime\prime}$.492 and has been smoothed by Gaussian kernel of
614:   $\sigma = 0^{\prime\prime}$.984.  The intensity scale is logarithmic.  
615:   We measure proper-motions of the forward shock in 14 regions
616:   indicated as rectangules (from Reg-1 to Reg-14).  {\it Right}:
617:   Linearly scaled difference (2006$-$2000) image between the two epochs.  
618: } 
619: \label{fig:image}
620: \end{figure}
621: 
622: \begin{figure}
623: \includegraphics[angle=0,scale=0.65]{f2.eps}
624: \caption{{\it Left}: Radial profiles binned with a
625:   0$^{\prime\prime}$.25 scale derived for Reg-4.  Data points with circles and
626:   triangles represent the first- and the second-epoch observations,
627:   respectively.   The shock motion is in the left direction.  {\it
628:   Right}: Same as {\it left} but for Reg-14.}    
629: \label{fig:profiles}
630: \end{figure}
631: 
632: \begin{figure}
633: \includegraphics[angle=0,scale=0.65]{f3.eps}
634: \caption{Proper motions as a function of azimuthal angle (lower
635:   x-axis) and region number (upper x-axis).  Note that
636:   quated errors represent only statistical uncertainties.} 
637: \label{fig:azimuth}
638: \end{figure}
639: 
640: \begin{figure}
641: \includegraphics[angle=0,scale=0.65]{f4.eps}
642: \caption{Same as Fig.~\ref{fig:azimuth} but for expansion indices.} 
643: \label{fig:azimuth2}
644: \end{figure}
645: 
646: \begin{figure}
647: \includegraphics[angle=0,scale=0.65]{f5.eps}
648: \caption{Best-fit circles to represent the X-ray boundaries of the
649:   northern and southern halves of the remnant are indicated as white
650:   dashed pie superposed on the second-epoch {\it Chandra}
651:   0.3--8.0\,keV band image.  
652:   The radio center (Matsui et al.\ 1984) is also indicated as a small
653:   black circle.  White contours (5 counts per pixel) are the X-ray
654:   boundaries used to estimate the best-fit circles shown in this image. 
655: }    
656: \label{fig:center}
657: \end{figure}
658: 
659: \begin{figure}
660: \includegraphics[angle=0,scale=0.65]{f6.eps}
661: \caption{Spectra extracted from the southeastern rim around Reg-6
662:   and the northwestern rim around Reg-13.}
663: \label{fig:spec}
664: \end{figure}
665: 
666: 
667: \end{document}
668: