astro-ph0406045/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[apjl]{emulateapj}
3: %\usepackage{apjfonts}
4: 
5: \begin{document}
6: 
7: \title{Kinematics of X-ray Emitting Components in Cassiopeia A}
8: 
9: \author{Tracey DeLaney\altaffilmark{1}, Lawrence Rudnick\altaffilmark{1}, 
10: Robert A. Fesen\altaffilmark{2}, T. W. Jones\altaffilmark{1}, Robert 
11: Petre\altaffilmark{3}, and Jon A. Morse\altaffilmark{4}}
12: 
13: \altaffiltext{1}{Department of Astronomy, University of Minnesota, 116 Church 
14: Street SE, Minneapolis, MN  55455; tdelaney@astro.umn.edu, 
15: larry@astro.umn.edu, twj@astro.umn.edu}
16: \altaffiltext{2}{Department of Physics \& Astronomy, Dartmouth College, 
17: Hanover, NH  03755; fesen@snr.dartmouth.edu}
18: \altaffiltext{3}{NASA Goddard Space Flight Center, Greenbelt, MD  20771; 
19: petre@lheavx.gsfc.nasa.gov}
20: \altaffiltext{4}{Department of Physics \& Astronomy, Arizona State 
21: University, Box 871504, Tempe, AZ  85287-1504; Jon.Morse@asu.edu}
22: 
23: \begin{abstract}
24: 
25: We present high-resolution X-ray proper motion measurements of Cassiopeia A 
26: using \emph{Chandra X-ray Observatory} observations from 2000 and 2002.  We 
27: separate the emission into four spectrally distinct classes: 
28: Si-dominated, Fe-dominated, low-energy-enhanced, and continuum-dominated.  
29: These classes also represent distinct spatial and kinematic components.  The 
30: Si- and Fe-dominated classes are ejecta and have a mean expansion rate of 
31: 0.2\% yr$^{-1}$.  This is the same as for the forward shock filaments but 
32: less than the 0.3\% yr$^{-1}$ characteristic of optical ejecta.  The 
33: low-energy-enhanced spectral class possibly illuminates a clumpy circumstellar 
34: component and has a mean expansion rate of 0.05\% yr$^{-1}$.  The 
35: continuum-dominated emission likely represents the forward shock and consists 
36: of diffuse circumstellar material which is seen as a circular ring around the 
37: periphery of the remnant as well as projected across the center.
38: 
39: \end{abstract}
40: 
41: \keywords{ISM: supernova remnants --- ISM: individual (Cassiopeia A)}
42: 
43: \section{Introduction}
44: 
45: Cassiopeia A (Cas A) is the youngest known ($\sim$ 330 yr, Thorstensen, 
46: Fesen, \& van den Bergh 2001) Galactic supernova remnant (SNR).  It is about 
47: 3.4 kpc \citep{rhf95} away and was most likely the result of either a type 
48: Ib or IIn supernova explosion \citep{co03}.  As one of the brightest X-ray 
49: and radio sources in the sky, Cas A is one of the best targets for studying 
50: SNR evolution.  At radio, X-ray, and optical wavelengths, Cas A is dominated 
51: by a 3$\farcm$5 diameter bright ring of emission with a fainter 5$\arcmin$ 
52: diameter plateau of radio and X-ray emission.  The X-ray emission is 
53: dominated by a thermal spectrum rich in emission lines from highly ionized 
54: atoms.  The radio emission is consistent with synchrotron radiation from 
55: relativistic electrons accelerated at shocks.  The optical emission 
56: originates from a complex system of chemically enriched knots.  
57: 
58: The kinematics of small and large scale radio features of Cas A are well 
59: studied.  The small-scale radio knots are significantly decelerated relative 
60: to the X-ray and optical emission.  They are not in homologous expansion and 
61: there are even inward moving knots.  The expansion rates of the small-scale 
62: features of the bright ring vary by over a factor of two as a function of 
63: azimuth with a mean rate of 0.11\% yr$^{-1}$ \citep{ar95,krg98,tuf86}.  
64: The bulk expansion of Cas A has been measured in several ways by 
65: \citet{ar95}, \citet{ag99}, and \citet{dr03} with reported expansion rates 
66: of $\approx$0.11\% yr$^{-1}$, $\approx$0.22\% yr$^{-1}$, and 
67: $\approx$0.07\% yr$^{-1}$, respectively.  For this discussion, we adopt the 
68: \citet{dr03} value of 0.07\% yr$^{-1}$ because it is the least contaminated 
69: by brightness changes and azimuthal asymmetries.
70: 
71: The optical components of Cas A consist of slow-moving ($\lesssim$ 500 
72: km s$^{-1}$) quasi-stationary flocculi (QSFs) and fast moving knots (FMKs, 
73: 4000$-$6000 km s$^{-1}$, $\approx$0.3\% yr$^{-1}$) \citep{kv76,vk85}.  The 
74: QSFs are thought to be shocked circumstellar material (CSM) from the 
75: presupernova wind of the progenitor while most of the FMKs are located on the 
76: bright ring and represent emission from ejecta that have interacted with the 
77: reverse shock.  There are also outlying knots of enriched composition with 
78: velocities from 8000$-$15,000 km s$^{-1}$ ($\approx$0.3\% yr$^{-1}$) 
79: \citep{fes01}.  
80: 
81: Previous X-ray proper motion measurements, conducted with \emph{Einstein} 
82: and \emph{ROSAT} at a resolution of 5\arcsec, showed that the mean 
83: expansion rate was 0.2\% yr$^{-1}$ \citep{krg98,vbk98}.  This established a 
84: paradoxical difference in expansion rates for the cospatial bright ring 
85: measured in optical, X-ray, and radio of approximately 3:2:1, respectively 
86: (Thorstensen, Fesen, \& van den Bergh 2001 and references therein).  The 
87: factor of two discrepancy between the X-ray and radio expansion is a common 
88: theme in young SNRs and has been found in both Kepler's SNR 
89: \citep{hug99,dsa88} and Tycho's SNR \citep{hug00}.  
90: 
91: With the advent of the \emph{Chandra X-ray Observatory}, we are now able to 
92: measure proper motions with unprecedented spatial resolution.  The results 
93: for the forward shock filaments around the outer edge of the remnant showed a 
94: mean expansion rate of 0.21\% yr$^{-1}$ \citep{dr03}, the same as the bright 
95: ring.  In this paper we extend our proper motion analysis to the rest of the 
96: SNR, for four spectrally distinct components.  One of these components 
97: represents a new class of X-ray emission while three components have been 
98: previously identified \citep{hrb00}.  We interpret the proper motions based 
99: on these spectral components and propose a solution to the expansion rate 
100: paradox.
101: 
102: \section{Observations, Analysis, and Results}
103: 
104: Cas A was observed for 50 ks with the ACIS-S3 chip onboard
105: \emph{Chandra} on 2000 January 30-31 and 2002 February 6.  For details
106: of the observing parameters and data reduction see \citet{hhp00},
107: \citet{gkr01}, and \citet{dr03}.  To measure proper motions, we
108: registered the 2000 and 2002 broadband (0.3$-$10 keV) X-ray images by
109: aligning them on the point source as described in \citet{dr03}.  The
110: uncertainty of the registration to the point source is 0$\farcs$015 based on 
111: Monte Carlo simulations which translates to a proper motion error of 0.005\%
112: yr$^{-1}$ at the mean radius of the exterior forward shock filaments
113: (150\arcsec).  The possible 0$\farcs$02 yr$^{-1}$ \citep{tfv01} motion
114: of the point source would result in a maximum proper motion error of
115: 0.013\% yr$^{-1}$ at 150$\arcsec$.  We constrained rotation between the two 
116: images by requiring the bright knots in the northeast jet to have no 
117: average rotation.  The jet knots are the least decelerated features in the 
118: remnant and so should be the least affected by CSM interactions.  The 
119: uncertainty in rotation alignment is $<0\fdg05$.  The proper motion 
120: measurements for 261 knots and filaments were then made by minimizing 
121: $\chi^2$ for a $\approx$7$\arcsec$ region around each feature, using the 
122: difference maps between the 2000 image and a set of 2002 images, shifted over 
123: a range of $\pm 2.5\arcsec$ in both RA and DEC. \cite{dr03} contains a
124: detailed description of the one-dimensional analog of this procedure.
125: We assigned an error of 0\farcs07 yr$^{-1}$ to all 261 features, which 
126: corresponds to the top 20\% of the Monte Carlo errors derived in \citet{dr03} 
127: where the mean error was 0\farcs04 yr$^{-1}$.  These Monte Carlo simulations 
128: include problems that would arise from features of varying sizes and the 
129: different point spread functions across the remnant, such as affects the 261 
130: features measured here.  The positions and proper motions for the 261 
131: features are shown in Table \ref{tbl1}.  We determined the radial component 
132: of the motion for each knot or filament and calculated the radial expansion 
133: rate ($v_r$) using the distance from the optical expansion center at 
134: $23^{\mathrm{h}}23^{\mathrm{m}}27\fs77$, 
135: $+58\degr48\arcmin49\farcs4$ (J2000) \citep{tfv01}.  
136: 
137: The 261 knots and filaments were spectrally classified using both extracted 
138: spectra and spectral tomography, as discussed below.  Based on the extracted 
139: spectra, four major classes were identified:  Si-dominated, 
140: Fe-dominated, low energy (LowE) enhanced, and continuum-dominated.  
141: The Si-, Fe-, and continuum-dominated classes were previously identified by 
142: \citet{hrb00}.  Representative spectra are shown in Figure \ref{spectra}.  
143: The Si-dominated emission is marked by very strong line emission from Si and 
144: S with contributions from O, Ca, Ar, and Mg.  The Fe-dominated spectrum shows 
145: strong FeL and FeK emission and can in some cases be modeled with nearly pure 
146: Fe emission \citep{hl03,lh03}.  The LowE-enhanced emission shows a relative 
147: increase in emission at low energies ($\approx$ 0.8$-$1.6 keV) compared to 
148: the global spectrum.  The continuum-dominated emission is marked by very 
149: little or no line emission and an increase in 4$-$6 keV emission.  Only 50\% 
150: of the knots and filaments can be cleanly classified using their spectra 
151: alone.  The remaining features combine the characteristics of two or more of 
152: the four basic classes identified above.
153: 
154: To separate overlapping spectral structures and to spatially visualize the 
155: spectral components, we used spectral tomography.  The spectral tomography 
156: technique has been used to analyze the spectra of radio galaxies \citep{kr97} 
157: and supernova remnants \citep{dkr02}.  It involves taking differences between 
158: images from two different energies with a scale factor chosen to accentuate 
159: features of interest.  In Figure \ref{tomos} we show four tomography images 
160: each representing the dominant spectral type indicated by the spectra in 
161: Figure \ref{spectra}.  To make the tomography images, we first constructed 
162: emission line images for Si (1.72$-$1.96 keV which includes Al and possibly 
163: weak Mg), FeL+Ne (0.9$-$1.15 keV which includes Ni), and FeL+Mg (1.35$-$1.55 
164: keV) and an image of the high energy continuum (HEC) region from 4$-$6 keV.  
165: These energy ranges were chosen to exploit the differences between the 
166: spectral classes.  The images were convolved to 3$\arcsec$ resolution and 
167: there was no background subtraction or correction applied for differential 
168: absorption across the remnant.  The Si-dominated image was constructed by 
169: scaling and subtracting the HEC image from the Si image.  Regions that are 
170: bright on the Si-dominated image represent ejecta whose emission is dominated 
171: by strong emission lines of the Si-group (Si, S, Ar, Ca) which all scale with 
172: the Si line \citep{hhp00}.  Regions that are dark on the Si-dominated image 
173: represent features where the line emission is not dominant with respect to 
174: the continuum.  The scale factor was chosen to separate the bimodal 
175: distribution of ratios in the log(Si/HEC) image.  The appearance of the image 
176: in Figure \ref{tomos} is only weakly dependent on the value of the scale 
177: factor.  Because only the HEC image was subtracted to create the Si-dominated 
178: image, there are low levels of contamination from the Fe-dominated and 
179: LowE-enhanced classes.  Subtracting out these two components does not 
180: significantly alter the Si-dominated image because the relative strength of 
181: the Si emission is very strong compared to the other classes.  
182: 
183: The continuum-dominated image was made in the same way as the Si-dominated 
184: image, but with a slightly different scaling factor to better show the 
185: continuum-dominated component.  The brightness on the image was reversed so 
186: the continuum-dominated emission appears positive while the Si-dominated 
187: emission appears negative.  The Fe-dominated image was constructed from 
188: scaling and subtracting the FeL+Mg image from the FeL+Ne image.  The bright 
189: areas represent regions where the ratio of FeL+Ne to FeL+Mg is greater than 
190: one, which is unique to the Fe-dominated ejecta class.  The LowE-enhanced 
191: image was made by first scaling and subtracting the Si image from the FeL+Ne 
192: image.  Because the resulting image also contained significant Fe-dominated 
193: ejecta emission, a second scaling and subtraction was applied using the 
194: positive brightness areas on the Fe-dominated image to eliminate regions 
195: with strong Fe ejecta emission.  The resulting image only shows those regions 
196: with spectra similar to the LowE-enhanced spectrum in Figure \ref{spectra}.  
197: 
198: The Si-dominated and Fe-dominated tomography images basically reproduce the 
199: spatial distributions of Si and Fe as shown by the equivalent width images 
200: in \citet{hhp00}, with the exception that the western region does not show as 
201: much Si and Fe emission as is actually present due to the increased 
202: absorption of low energy X-ray emission in that region \citep{kra96,wbv02}.  
203: The continuum-dominated image shows the same forward shock filaments around 
204: the outer edge of the remnant as described by \citet{gkr01}.  There is also 
205: continuum-dominated filamentary emission interior to the bright ring of the 
206: remnant which \citet{hrb00} speculated results from X-ray synchrotron emission 
207: associated with the forward shock.  The continuum-dominated image matches the 
208: 8-15 keV \emph{XMM} continuum image \citep{bwv01}, in the west particularly, 
209: indicating that this emission may indeed have a strong nonthermal component.  
210: 
211: The LowE-enhanced image represents emission that has not previously been 
212: explored.  Some of the LowE-enhanced emission is associated with optical QSF 
213: emission such as the features in the interior of the remnant, the bright 
214: northeast region at the outer edge of the bright ring, and the southern arc 
215: \citep{lmu95}.  Some of the LowE-enhanced emission has no QSF counterparts 
216: such as the emission to the southeast.  This may be because the low energy 
217: spectral region contains emission lines of O, Fe, Ne, and Mg, making it 
218: possible that some of the emission in the LowE-enhanced image is due to more 
219: than one spectral component.
220: 
221: The power of the tomography technique is in its ability to isolate spatially 
222: overlapping spectral components.  Using spectra alone, one may misidentify 
223: features as intermediate spectral types when they are simply a combination of 
224: unrelated overlapping structures.  The fact that the tomography images 
225: isolate X-ray structures that match emission features in other wavebands 
226: proves the efficacy of the technique.  This will be explored in more detail 
227: in a future paper (DeLaney et al. in preparation).
228: 
229: Using both extracted spectra and spectral tomography, 80\% of the 261 knots 
230: were classified into one of the four spectral classes.  Histograms showing 
231: the radial expansion rates broken down by spectral class are shown in Figure 
232: \ref{hist}.  Also included for completeness are the expansion rates from the 
233: 29 forward shock filaments presented in \citet{dr03}.  The expansion rate 
234: distribution of the knots and filaments is very broad; however, when broken 
235: down by spectral class, it is clear that each species has its own 
236: characteristic motion.  The expansion rate distribution of each class holds 
237: for both the tomography and spectrum identified sets independently -- proving 
238: the robustness of the tomography identifications.  The Si- and Fe-dominated 
239: ejecta knots have the fastest expansion rates with a mean value of 
240: 0.2$\pm$0.01\% yr$^{-1}$.  The expansion rate, expressed in \% yr$^{-1}$, is 
241: independent of projection effects.  
242: 
243: To compare to Doppler measurements, we can project the expansion rates onto 
244: the plane of the sky.  The Si-dominated knots are at an average radius of 
245: 95$\arcsec$ which converts to a transverse velocity of 3100 km s$^{-1}$ for a 
246: distance of 3.4 kpc to Cas A.  The Fe-dominated knots lie further out on 
247: average at a radius of 120$\arcsec$ yielding an average transverse velocity 
248: of 3900 km s$^{-1}$ for the same expansion rate.  The Doppler measurements 
249: from \emph{XMM} show that the Fe emission is at a larger radius on average 
250: than the Si emission and that the average line-of-sight velocities are 
251: smaller than the average transverse velocities at approximately 1500 
252: km s$^{-1}$ for the Fe and 1000 km s$^{-1}$ for the Si \citep{wbv02}.  The 
253: Doppler velocities measured for Si with \emph{Chandra} are slightly faster 
254: than those measured with \emph{XMM} at 2000-3000 km s$^{-1}$ \citep{hsp01}.  
255: The difference between the Doppler measurements for Si could be due to the 
256: way in which the zero point in velocity was determined or possibly to the 
257: different spectral and spatial resolutions of \emph{XMM} and \emph{Chandra}.  
258: The mismatch of Doppler velocities and transverse velocities is not 
259: surprising.  The average transverse and line-of-sight velocities would only 
260: match if the ejecta were uniformly distributed and thus observed at all 
261: projected radii such that, on average, projection effects would contribute 
262: equally in each case.  For Cas A, the ejecta do not appear to be uniformly 
263: distributed \citep{wbv02}.  
264: 
265: The LowE-enhanced knots are moving quite slowly with a mean expansion rate of 
266: 0.05$\pm$0.01\% yr$^{-1}$.  The Si- and Fe-dominated knot distribution and 
267: the LowE-enhanced knot distribution are narrow with respect to the separation 
268: between them which is consistent with virtually non-overlapping populations.  
269: The continuum-dominated knots and filaments are split into ``exterior'' and 
270: ``projected interior'' where the exterior features are the forward shock 
271: filaments around the periphery of the remnant.  The exterior features have a 
272: mean expansion rate of 0.2$\pm$0.01\% yr$^{-1}$ while the expansion rates of 
273: the projected interior population varies significantly and includes apparent 
274: inward motions.  Most of the inward motions are concentrated to the south and 
275: west between 170$\degr$ and 300$\degr$ azimuth.  Radio proper motion 
276: measurements also show inward moving features in this region \citep{ar95}.
277: 
278: \section{Discussion}
279: 
280: As was found previously with the \emph{Einstein} and \emph{ROSAT} proper 
281: motion measurements \citep{krg98,vbk98}, the X-ray ejecta are moving slower, 
282: as a class, relative to the cospatial \emph{population} of optical ejecta and 
283: moving faster, as a class, than the cospatial \emph{population} of radio 
284: knots (On a case-by-case basis, the small number of matched X-ray/optical and 
285: X-ray/radio knots show the same motions (DeLaney et al. in preparation)).  
286: The X-ray/optical population difference is thought to be due to the density 
287: difference between the two populations \citep{ajr94,hsp01,fes01,dr03}.  High 
288: density ejecta experience very little deceleration, and thus very little 
289: disruption, allowing the ejecta to cool rather quickly and emit at optical 
290: wavelengths.  Low density ejecta experience more deceleration, and are heated 
291: to higher temperatures.  These ejecta emit at X-ray wavelengths and may even 
292: be completely disrupted before they can cool enough to emit optically.  The 
293: radio emission arises from synchrotron radiation, which requires 
294: amplification of magnetic fields to be strong enough for the emission to be 
295: seen. This amplification results from the turbulence associated with ejecta 
296: deceleration. If significant deceleration is needed for amplification, radio 
297: emitting knots would be decelerated the most. 
298: 
299: Of particular interest is the fact that the Si- and Fe-dominated ejecta have 
300: experienced the same fractional deceleration in their expansion rate from the 
301: free expansion rate of 0.3\% yr$^{-1}$ to 0.2\% yr$^{-1}$.  The Fe-rich 
302: emission (which results from explosive silicon burning near the core of the 
303: progenitor) is on average exterior to the Si-group emission (which results 
304: from explosive oxygen burning farther out from the star's center).  This 
305: spatial overturning of material could be due to several scenarios such as:  
306: 1.  The Si and Fe ejecta may have had different initial velocities imparted 
307: from the SN explosion such that the Fe layer overtook the Si layer.  2.  The 
308: Fe ejecta may have experienced less deceleration at the reverse shock than 
309: the Si ejecta (perhaps due to density differences between the two types of 
310: ejecta).  3.  The overturning may have occurred in the star prior to or during 
311: the SN explosion resulting in a spatial separation between the Fe and Si even 
312: in the event of the same initial explosion velocities.  These scenarios would 
313: not necessarily predict that the Si and Fe would have the same expansion 
314: rate, as observed.  Furthermore, the X-ray ejecta have the same expansion 
315: rate as the forward shock filaments (i.e. homologous expansion on average), 
316: perhaps implying that the two populations are part of a coupled dynamical 
317: system in which the pressure in the X-ray ejecta is determined by the forward 
318: shock speed; although this does not explain why the optical ejecta would not 
319: be affected.  Even if initially the different types of ejecta were not 
320: coupled to eachother or to the forward shock speed, after interaction with 
321: the reverse shock and secondary shocks, the ejecta may now be coupled and 
322: expanding with the forward shock \citep{lh03}.
323: 
324: At the other end of the expansion rate spectrum are the LowE-enhanced knots 
325: and filaments.  Spatially, many of these features match very well with 
326: optical QSF emission, particularly the arc to the south, the emission on the 
327: bright ring to the northeast, and much of the interior emission 
328: (DeLaney et al., in preparation).  On average, QSFs show little or no motion 
329: \citep{vk85} while the mean expansion rate of the LowE-enhanced knots is 
330: 0.05$\pm$0.01\% yr$^{-1}$.  Based on the spatial and dynamical similarities, 
331: we propose that the LowE-enhanced X-ray emission, like the optical QSF 
332: emission, represents the clumpy component of the CSM that was deposited by 
333: the wind from the progenitor.  The slight difference in expansion rates 
334: between the optical QSF population and the X-ray LowE-enhanced population 
335: might be due to a pre-existing density distribution in the progenitor wind 
336: \citep{co03}.  On average, less dense X-ray clumps may be accelerated more by 
337: the forward shock than more dense optical clumps.  The low-energy enhancement 
338: in the X-ray spectrum relative to the global X-ray spectrum (i.e. a 
339: ``softer'' X-ray spectrum) is expected as dense clumps are overrun by the 
340: forward shock resulting in slow moving transmitted shocks through the clumps 
341: \citep{co03} which will produce an X-ray spectrum with a ``cooler'' 
342: characteristic temperature compared to other X-ray components 
343: \citep{hc86,pbg03}.
344: 
345: Another likely CSM component, a diffuse component, was discovered when the 
346: ``first light'' observation of \emph{Chandra} revealed a thin, 
347: continuum-dominated ring of emission around the outer edge of the remnant 
348: which has been identified as the forward shock \citep{gkr01}.  We find 
349: similar continuum-dominated knots and filaments across the face of the 
350: remnant and propose that these are also forward shock filaments seen in 
351: projection.  These ``projected interior'' filaments are associated with 
352: edges and filaments of radio emission (DeLaney et al., in preparation) just 
353: as the ``exterior'' forward shock filaments are associated with the edge of 
354: the radio plateau \citep{gkr01}.  The projected interior and exterior 
355: filaments also have the same X-ray spectral shape (bottom panel of Figure 
356: \ref{spectra}) and they are both associated with steep radio spectral index 
357: features \citep{ar96}.  The projected interior filaments are not evenly 
358: distributed across the face of the remnant, but rather are concentrated to 
359: the west where there is significant X-ray absorption \citep{kra96,wbv02}.  
360: The ``inward'' motions and the apparent motions faster than free expansion 
361: for the projected filaments are consistent with the forward shock following 
362: the path of least resistance through the clumpy medium and showing apparent 
363: chaotic behavior when seen face on.  We may also be observing a more complex 
364: pattern involving both motions and changes in synchrotron brightness as 
365: magnetic fields are amplified in shear layers and in turbulent wakes due to 
366: the forward shock wrapping around CSM clumps \citep{jk93}.  The interior 
367: continuum-dominated filaments have also been interpreted as nonthermal 
368: bremsstrahlung emission associated with the ejecta/reverse shock 
369: \citep{bwv01}.  The \emph{XMM} hard X-ray continuum images show that the 
370: continuum-dominated emission in the west is significantly brighter than the 
371: exterior forward shock filaments possibly indicating that the two 
372: continuum-dominated populations arise through different emission processes.  
373: If the interior filaments are instead internal to the SNR, they are unlikely 
374: to arise through X-ray synchrotron emission because the high magnetic field 
375: strength will cause rapid aging of electrons at these energies \citep{vl03}.
376: 
377: In summary, the high-resolution proper motion measurements, along with the 
378: spectral classes, reveal that the X-ray emission can be separated into three 
379: components each with a distinct kinematic signature.  The ejecta component 
380: consists of both Si- and Fe-dominated emission populations, is associated 
381: with optical FMK emission, and has been significantly decelerated relative to 
382: the FMKs.  This greater deceleration of the X-ray ejecta is likely due to the 
383: lower densities in this material relative to the optical FMKs.  The 
384: slow-moving LowE-enhanced emission may be analogous to optical QSF emission 
385: thus representing a clumpy CSM component.  A diffuse CSM component may be 
386: associated with continuum-dominated emission from material swept up by the 
387: forward shock.  In this scenario, the chaotic motions of the interior 
388: continuum-dominated filaments would be explained as the projected behavior of 
389: the forward shock interacting with an inhomogeneous CSM.   The exterior 
390: continuum-dominated filaments have the same expansion rate as the X-ray 
391: ejecta suggesting a dynamic coupling between the forward shock and the ejecta. 
392: 
393: \acknowledgments
394: 
395: We thank Una Hwang for valuable conversations and manuscript comments.  This 
396: work was supported at the University of Minnesota under Smithsonian grant 
397: SMITHSONIAN/GO1-2051A/NASA, NASA grants HST-AR-09537.01 and NAG5-10774 and by 
398: the Minnesota Supercomputer Institute.
399: 
400: \begin{thebibliography}{}
401: 
402: \bibitem[Ag\"{u}eros \& Green(1999)]{ag99} Ag\"{u}eros, M. A. \& Green, 
403: D. A. 1999, \mnras, 305, 957 
404: 
405: \bibitem[Anderson et al.(1994)]{ajr94} Anderson, M., Jones, T. W., Rudnick, 
406: L., Tregillis, I. L., \& Kang, Hyesung 1994, \apj, 421, L31
407: 
408: \bibitem[Anderson \& Rudnick(1995)]{ar95} Anderson, M. \& Rudnick, L. 1995, 
409: \apj, 441, 307
410: 
411: \bibitem[Anderson \& Rudnick(1996)]{ar96} Anderson, M. \& Rudnick, L. 1996, 
412: \apj, 456, 234
413: 
414: \bibitem[Bleeker et al.(2001)]{bwv01} Bleeker, J. A. M., Willingale, R., 
415: van der Heyden, K., Dennerl, K., Kaastra, J. S., Aschenbach, B. \& Vink, J. 
416: 2001, \mnras, 365, L225
417: 
418: \bibitem[Chevalier \& Oishi(2003)]{co03} Chevalier, R. \& Oishi, J. 2003, 
419: \apj, 593, L23
420: 
421: \bibitem[DeLaney et al.(2002)]{dkr02} DeLaney, T., Koralesky, B., \& Rudnick, 
422: L. 2002, \apj, 580, 914
423: 
424: \bibitem[DeLaney \& Rudnick(2003)]{dr03} DeLaney, T. \& Rudnick, L. 2003, 
425: \apj, 589, 818
426: 
427: \bibitem[Dickel et al.(1988)]{dsa88} Dickel, J. R., Sault, R., 
428: Arendt, R. G., Matsui, Y., \& Korista, K. T. 1988, \apj, 330, 254
429: 
430: \bibitem[Fesen(2001)]{fes01} Fesen, R. A. 2001, \apjs, 133, 161
431: 
432: \bibitem[Fesen et al.(2001)]{fmc01} Fesen, R. A., Morse, J. A., Chevalier, 
433: R. A., Borkowski, K. J., Gerardy, C. L., Lawrence, S. S., \& van den Bergh, 
434: S. 2001, \aj, 122, 2644
435: 
436: \bibitem[Gotthelf et al.(2001)]{gkr01} Gotthelf, E., Koralesky, B., 
437: Rudnick, L., Jones, T., Hwang, U., \& Petre, R. 2001, \apj, 552, L39
438: 
439: \bibitem[Hester \& Cox(1986)]{hc86} Hester, J. J. \& Cox, D. P. 1986, \apj, 
440: 300, 675
441: 
442: \bibitem[Hughes(1999)]{hug99} Hughes, J. P. 1999, \apj, 527, 298
443: 
444: \bibitem[Hughes(2000)]{hug00} Hughes, J. P. 2000, \apjl, 545, L53
445: 
446: \bibitem[Hughes et al.(2000)]{hrb00} Hughes, J. P., Rakowski, C. E., 
447: Burrows, D. N., \& Slane, P. O. 2000, \apj, 528, L109
448: 
449: \bibitem[Hwang, Holt, \& Petre(2000)]{hhp00} Hwang, U., Holt, S., \& Petre, 
450: R. 2000, \apj, 537, L119
451: 
452: \bibitem[Hwang \& Laming(2003)]{hl03} Hwang, U. \& Laming, J. M. 2003, \apj, 
453: 597, 362
454: 
455: \bibitem[Hwang et al.(2001)]{hsp01} Hwang, U., Szymkowiak, A. E., Petre, R., 
456: \& Holt, S. S. 2001, \apj, 560, L175
457: 
458: \bibitem[Jones \& Kang(1993)]{jk93} Jones, T. W. \& Kang, Hyesung 1993, 
459: \apj, 402, 560 
460: 
461: \bibitem[Kamper \& van den Bergh(1976)]{kv76} Kamper, K. \& van den Bergh, S. 
462: 1976, \apjs, 32, 351
463: 
464: \bibitem[Katz-Stone \& Rudnick(1997)]{kr97} Katz-Stone, D. M. \& Rudnick, 
465: L. 1997, \apj, 488, 146
466: 
467: \bibitem[Koehane, Rudnick \& Anderson(1996)]{kra96} Koehane, J. W., Rudnick, 
468: L., \& Anderson, M. C. 1996, \apj, 466, 309
469: 
470: \bibitem[Koralesky et al.(1998)]{krg98} Koralesky, B., Rudnick, L., 
471: Gotthelf, E., \& Keohane, J. 1998, \apj, 505, L27
472: 
473: \bibitem[Lawrence et al.(1995)]{lmu95} Lawrence, J. L., MacAlpine, G. M., 
474: Uomoto, A., Woodgate, B. E., Brown, L. W., Oliversen, R. J., Lowenthal, 
475: J. D., \& Liu, C. 1995, \aj, 109, 2635
476: 
477: \bibitem[Laming \& Hwang(2003)]{lh03} Laming, J. M. \& Hwang, U. 2003, \apj, 
478: 597, 347
479: 
480: \bibitem[Park et al.(2003)]{pbg03} Park, S., Burrows, D. N., Garmire, G. P., 
481: Nousek, J. A., Hughes, J. P., \& Williams, R. M. 2003, \apj, 586, 210
482: 
483: \bibitem[Reed et al.(1995)]{rhf95} Reed, J. E., Hester, J. J., Fabian, A. C., 
484: \& Winkler, P. F. 1995, \apj, 440, 706
485: 
486: \bibitem[Thorstensen, Fesen, \& van den Bergh(2001)]{tfv01} Thorstensen, J., 
487: Fesen, R., \& van den Bergh, S. 2001, \aj, 122, 297
488: 
489: \bibitem[Tuffs(1986)]{tuf86} Tuffs, R. 1986, \mnras, 219, 13
490: 
491: \bibitem[van den Bergh \& Kamper(1985)]{vk85} van den Bergh, S. \& Kamper, 
492: K. 1985, \apj, 293, 537
493: 
494: \bibitem[Vink et al.(1998)]{vbk98} Vink, J., Bloemen, H., Kaastra, 
495: J. S., \& Bleeker, J. A. M. 1998, \aap, 339, 201
496: 
497: \bibitem[Vink \& Laming(2003)]{vl03} Vink, J. \& Laming, J. M. 2003, 
498: \apj, 584, 758
499: 
500: \bibitem[Willingale et al.(2002)]{wbv02} Willingale, R., Bleeker, 
501: J. A. M., van der Heyden, K. J., Kaastra, J. S., \& Vink, J. 2002, \aap, 
502: 381, 1039
503: 
504: \end{thebibliography}
505: 
506: \clearpage
507: 
508: \begin{deluxetable}{cccccccccc}
509: \tabletypesize{\scriptsize}
510: \tablecaption{Knot and Filament Proper Motions \label{tbl1}}
511: \tablewidth{0pt}
512: \tablehead{
513: \colhead{\#} & 
514: \colhead{blc$_x$} & 
515: \colhead{blc$_y$} &
516: \colhead{trc$_x$} &
517: \colhead{trc$_y$} & 
518: \colhead{$\mu_x$} & 
519: \colhead{$\mu_y$} &
520: \colhead{$v_r$} &
521: \colhead{$\delta v_r$} & 
522: \colhead{class}
523: \\
524: \colhead{ } &
525: \colhead{$\arcsec$} &
526: \colhead{$\arcsec$} &
527: \colhead{$\arcsec$} &
528: \colhead{$\arcsec$} &
529: \colhead{$\arcsec$ yr$^{-1}$} &
530: \colhead{$\arcsec$ yr$^{-1}$} &
531: \colhead{\% yr$^{-1}$} &
532: \colhead{\% yr$^{-1}$} &
533: \colhead{ }
534: }
535: \startdata
536: 1 & \phs\phn2.0 & $-$105.3 & \phs\phn8.4 & \phn$-$99.9 & \phs0.01 & $-$0.20 &  \phs0.19 & 0.07 & L \\
537: 2 & \phs60.5 & $-$122.5 & \phs67.4 & $-$115.6 & \phs0.05 & $-$0.01 & \phs0.03 & 0.05 &  M \\
538: 3 & \phs25.6 & \phn$-$83.6 & \phs32.0 & \phn$-$77.2 & \phs0.01 & $-$0.03 & \phs0.03 & 0.09 & M \\
539: 4 & $-$32.0 & \phn$-$65.4 & $-$25.1 & \phn$-$58.1 & $-$0.01 & $-$0.10 & \phs0.15 & 0.11 & S \\
540: 5 & \phn$-$7.4 & \phn$-$33.0 & \phn$-$1.0 & \phn$-$27.1 & $-$0.02 & \phs0.05 & $-$0.15 & 0.24 & L \\
541: \enddata
542: \tablecomments{The complete version of this table is in the electronic 
543: edition of the Journal.  The printed edition contains only a sample.  All 
544: knots have been assigned a constant error of 0\farcs07 yr$^{-1}$.  Col. 
545: 1. Knot or filament designation number.  Cols. 2-5.  Location of bottom left 
546: corner and top right corner of box defining the region used to measure each 
547: knot or filament and extract spectra (arcseconds of offset on epoch 2002 
548: image with respect to the optical expansion center at 
549: $23^{\mathrm{h}}23^{\mathrm{m}}27\fs77$, 
550: $+58\degr48\arcmin49\farcs4$ (J2000) \citep{tfv01}).  Cols. 6-7. Proper 
551: motion in right ascension and declination ($\arcsec$ yr$^{-1}$).  Positive 
552: right ascension offset or motion is defined to be westward.  Col. 8. Radial 
553: component of motion (\% yr$^{-1}$).  Col. 9. Error in radial expansion 
554: (\% yr$^{-1}$).  Col. 10.  Spectral classification:  S=Si-dominated, 
555: F=Fe-dominated, L=LowE-enhanced, C=continuum-dominated, M=mixture of two or 
556: more classes.}
557: \end{deluxetable}
558: 
559: \clearpage
560: 
561: \begin{figure}
562: \epsscale{0.5}
563: \plotone{f1.eps}
564: \caption{\emph{Chandra} ACIS-S spectra of Cas A showing the global spectrum 
565: (in red) and typical spectra for the four spectral classes identified in the 
566: text.  The Si-dominated emission is marked by very strong Si group emission.  
567: The Fe-dominated spectrum shows strong FeL and FeK emission.  The 
568: LowE-enhanced emission shows a relative increase in emission at low energies 
569: compared to the global spectrum.  The continuum-dominated emission is marked 
570: by very little or no line emission and an increase in 4$-$6 keV emission 
571: relative to the global spectrum.
572: \label{spectra}}
573: \end{figure}
574: 
575: \begin{figure}
576: \epsscale{1}
577: \plotone{f2.eps}
578: \caption{Representative tomography images used to spatially identify 
579: emission from the four classes whose spectra are shown in Figure 
580: \ref{spectra}.  The Si- and Fe- dominated images represent ejecta emission.  
581: Note that, because of absorption, the ejecta emission on the western side of 
582: the remnant is underrepresented.  The LowE-enhanced image likely represents 
583: clumpy CSM emission analogous to optical QSF emission.  The 
584: continuum-dominated image possibly represents emission from diffuse CSM 
585: swept up by the forward shock.  
586: \label{tomos}}
587: \end{figure}
588: 
589: \begin{figure}
590: \epsscale{0.5}
591: \plotone{f3.eps}
592: \caption{Histogram of expansion rates in Cas A.  Arrows represent the average 
593: measurement error for each population.  Dashed lines represent the mean 
594: expansion values for the Si-dominated and LowE-enhanced populations.  There 
595: are also two continuum-dominated knots with expansion rates less than 
596: $-$0.4\% yr$^{-1}$.  The ``exterior'' continuum-dominated filaments are from 
597: \citet{dr03} and have been included here for completeness.  Free expansion 
598: corresponds to 0.3\% yr$^{-1}$.
599: \label{hist}}
600: \end{figure}
601: 
602: \end{document}
603: 
604: