1: %
2: %\documentstyle[12pt,aasms4]{article}
3: \documentstyle[11pt,aaspp4]{article}
4:
5: %\received{4 August 1988}
6: %\accepted{23 September 1988}
7: %\journalid{337}{15 January 1989}
8: %\articleid{11}{14}
9:
10: \lefthead{}
11: \righthead{}
12:
13: \def\einstein{{\it Einstein}}
14: \def\rosat{{\it ROSAT}}
15: \def\asca{{\it ASCA}}
16: \def\chandra{{\it Chandra}}
17: \def\CXO{{\it Chandra X-Ray Observatory}}
18: \def\eohone{E0102.2$-$7219}
19: \def\kms{km s$^{-1}$}
20: \def\mydegree{^\circ\mskip-5mu}
21: \def\myarcmin{^\prime\mskip-5mu}
22: \def\myarcsec{\mskip1mu^{\prime\prime}\mskip-7mu.\mskip2mu}
23: \def\lsim{\hbox{\raise.35ex\rlap{$<$}\lower.6ex\hbox{$\sim$}\ }}
24: \def\gsim{\hbox{\raise.35ex\rlap{$>$}\lower.6ex\hbox{$\sim$}\ }}
25:
26: \begin{document}
27:
28: \title{Electron Heating and Cosmic Rays at a Supernova
29: Shock from \chandra\ X-ray Observations of \eohone}
30:
31: \author{
32: John P.~Hughes\altaffilmark{1,2},
33: Cara E.~Rakowski\altaffilmark{1,2}, and
34: Anne Decourchelle\altaffilmark{2}
35: }
36: %\affil{
37: %Service d'Astrophysique, L'Orme des Merisiers,
38: %CEA-Saclay, 91191 Gif-sur-Yvette Cedex France
39: %}
40:
41: \altaffiltext{1}{Department of Physics and Astronomy, Rutgers The State
42: University of New Jersey, 136 Frelinghuysen Road, Piscataway NJ 08854-8019;
43: E-mail: jph@physics.rutgers.edu and rakowski@physics.rutgers.edu
44: }
45:
46: \altaffiltext{2}{
47: Service d'Astrophysique, L'Orme des Merisiers,
48: CEA-Saclay, 91191 Gif-sur-Yvette Cedex France
49: }
50:
51:
52: \begin{abstract}
53:
54: In this {\it Letter} we use the unprecedented spatial resolution of
55: the \CXO\ to carry out, for the first time, a measurement of the
56: post-shock electron temperature and proper motion of a young SNR,
57: specifically to address questions about the post-shock partition of
58: energy among electrons, ions, and cosmic rays. The expansion rate,
59: $0.100\% \pm 0.025\% \rm\, yr^{-1}$, and inferred age, $\sim$1000 yr,
60: of \eohone, from a comparison of X-ray observations spanning 20 years,
61: are fully consistent with previous estimates based on studies of high
62: velocity oxygen-rich optical filaments in the remnant. With a radius
63: of 6.4 pc for the blast wave estimated from the \chandra\ image, our
64: expansion rate implies a blast wave velocity of $\sim$6000 km s$^{-1}$
65: and a range of electron temperatures 2.5 - 45 keV, dependent on the
66: degree of collisionless electron heating. Analysis of the
67: \chandra\ ACIS spectrum of the immediate post-shock region reveals a
68: thermal plasma with abundances and column density typical of the Small
69: Magellanic Cloud and an electron temperature of 0.4--1 keV.
70: The measured electron temperature is significantly lower than the
71: plausible range above, which can only be reconciled if we assume
72: that a significant fraction of the shock energy, rather than
73: contributing to the heating of the post-shock electrons and ions, has
74: gone into generating cosmic rays.
75:
76: \end{abstract}
77:
78: \keywords{
79: cosmic rays ---
80: ISM: individual (\eohone) ---
81: shock waves --
82: supernova remnants ---
83: X-rays: ISM
84: }
85:
86: \section{INTRODUCTION}
87:
88: \par
89:
90: Supernova remnants (SNRs) are among that rare class of astronomical
91: objects whose dynamical evolution can be probed directly through both
92: radial velocity and proper motion measurements. Together such studies
93: provide important constraints on the evolution of young SNRs and the
94: structure of the ejecta and ambient medium. Furthermore, when its
95: distance is known, measurement of a remnant's proper motion is
96: equivalent to a measurement of the velocity of the supernova blast
97: wave. This fundamental quantity determines the total energy available
98: to the remnant for partition into bulk kinetic energy of the ambient
99: medium, shock-heated electrons and ions, and relativistic particles,
100: i.e., cosmic rays (see Chevalier 1983; Decourchelle, Ellison, \&
101: Ballet 2000). Conservation laws provide some constraints on how the
102: shock energy is divided up into the several forms, but some issues,
103: such as the fraction of shock energy that goes into cosmic rays
104: and the extent of collisionless heating of the electrons, are still
105: theoretically uncertain and need to be determined observationally.
106:
107: The brightest X-ray supernova remnant (SNR) in the Small Magellanic
108: Cloud (SMC), \eohone, is dominated by line emission from highly
109: ionized atoms of O, Ne, and Mg (Hayashi et al.~1994; Gaetz et
110: al.~2000) and its X-ray morphology is that of a strongly
111: limb-brightened shell, roughly 40$^{\prime\prime}$ in diameter (Hughes
112: 1988). Optical filaments in the remnant are also rich in O and Ne,
113: extend over a diameter of 24$^{\prime\prime}$ (Dopita et al.~1981),
114: and display radial velocities that span a range of 6500 \kms\ (Tuohy
115: \& Dopita 1983). At the distance to the SMC (60 kpc, e.g., van den
116: Bergh 2000), the proper motion of the optical filaments should be
117: $0\myarcsec011$ yr$^{-1}$ or an expansion rate of 0.092\% yr$^{-1}$.
118: Herein we determine the X-ray expansion rate, equivalent to the
119: velocity of the shock, to obtain an estimate of the total shock
120: energy. We then compare this to the electron temperature in the
121: immediate post-shock region, through fits to spectra isolated by
122: \chandra's superb spatial resolution, to determine the fraction of
123: energy in shock-heated electrons.
124:
125: \section{EXPANSION RATE}
126:
127: Prior to the launch of \chandra\ the highest spatial resolution images
128: of \eohone\ were those taken by the \einstein\ and \rosat\ High
129: Resolution Imagers, hereafter EHRI and RHRI. During Orbital Activation
130: and Check-out of \chandra, \eohone\ was observed by the back-side
131: illuminated chip (S3) of the ACIS-S instrument (Garmire 1997). A
132: processed event file (id no.~1231 created on 1999-09-07T18:39:11) was
133: obtained from the \chandra\ X-ray Center and analysis was carried out
134: using standard astronomical software. For spectral analysis (\S 3) the
135: events were gain-corrected using the appropriate gain map for a focal
136: plane temperature of $-$100 C. Events were time filtered to remove
137: intervals of high background as well as frames with bad or no
138: aspect. Dead and flickering pixels were also removed. An ephemeris of
139: all three observations is given in Table 1.
140:
141: The EHRI and RHRI each made broad-band X-ray images, but in slightly
142: different spectral bands. The possibility of spectral variations with
143: position has been the main factor limiting the widespread use of EHRI
144: and RHRI images for carrying out precise proper motion studies of
145: X-ray SNRs. Now, however, the \chandra\ ACIS-S data provide a
146: spectral dimension that can be used to take full advantage of the long
147: time baseline for proper motion studies set by the earlier
148: observations.
149:
150: In our approach we slice the ACIS-S data into a large number (50--100)
151: of images according to observed photon energy using narrow X-ray
152: spectral bands. These separate images are then multiplied by the
153: ratio of effective area between the EHRI (for example) and the
154: \chandra\ ACIS-S at the photon energy appropriate to the spectral band
155: of each individual image slice. The final ACIS-S image, corrected to
156: the EHRI band in this case, is then the sum of the image slices. This
157: is done for both of the HRIs. Clearly the accuracy of this procedure
158: depends on how well the various effective area functions are known, an
159: effect we explore below.
160:
161: The on-orbit 50\% encircled energy radius for the \chandra\ ACIS-S,
162: $R_{50\%}=0\myarcsec4$ (Dewey 1999), is about an order of magnitude
163: smaller than that of the RHRI or EHRI, both of which had $R_{50\%}
164: \approx 4^{\prime\prime}$ (Giacconi et al.~1979; David et al.~1998).
165: We can therefore ignore the small \chandra\ point-response-function
166: (PRF) and convolve the band-corrected ACIS-S image with the
167: appropriate EHRI or RHRI PRF models. In our preliminary investigations
168: we found that the parameters of the analytical PRF models (EHRI: Henry
169: \& Henriksen 1986; RHRI David et al.~1998) needed to be modified
170: slightly from their published values in order to obtain good fits to
171: the \eohone\ data. For both datasets it was the largest spatial scale
172: component of the PRF model that needed adjustment. For the RHRI the
173: best-fit normalization of the $\sim$30$^{\prime\prime}$-scale
174: exponential component was 33\% higher than nominal and for the EHRI
175: the normalization of the $\sim$13$^{\prime\prime}$-scale power-law
176: component was less than nominal by 22\%. \eohone\ is small enough that
177: spatial variations in the background, vignetting, or PRF can all be
178: safely neglected.
179:
180: The final piece of information required for the expansion study is the
181: plate scales of the three instruments. The plate scales of the EHRI
182: and RHRI are $0\myarcsec4965 \pm 0\myarcsec0012$ per pixel and
183: $0\myarcsec499 \pm 0\myarcsec001$ per pixel, respectively (David et
184: al.~1998). For the ACIS-S the plate scale near the center of the field
185: of view is $0\myarcsec49115 \pm 0\myarcsec00010$ per pixel based on
186: \chandra\ in-flight observations of the star cluster NGC 2516
187: (M.~Markevitch 1999, private communication). This value is within
188: 0.2\% of the pre-flight estimate of $0\myarcsec4920$ per pixel.
189:
190: The approach we follow for determining the expansion rate is that
191: described in Hughes (1999). In brief each HRI image (the ``data'') is
192: compared to the appropriate band-corrected ACIS-S image (the
193: ``model'') in a fit for parameters using a maximum-likelihood
194: estimator as the figure-of-merit function. The fitted parameters are
195: the relative intensity scale between the two images, the relative
196: pixel position, the fractional amount of expansion or contraction of
197: the spatial scale, and the normalization of the largest-scale PRF
198: component. The difference in background levels (e.g., ACIS-S vs.\
199: EHRI or HRI) is fixed at the value determined from source-free
200: portions of the image and is included as a spatially uniform term. In
201: the fitting software the expansion or contraction of the ``model''
202: image is done first using a rebinning scheme that conserves flux, then
203: the scaled image is convolved with the appropriate instrumental PRF,
204: and finally the background term is included. The only interesting
205: parameter that we obtain from this comparison is the expansion (or
206: contraction) factor, which tells us the global mean expansion rate of
207: the remnant. This procedure explicitly assumes that the fractional
208: expansion rate is uniform over the entire remnant, both radially and
209: azimuthally.
210:
211: Figure 1 shows the fitted results. \eohone\ has expanded by
212: $0.74\pm0.31$\% based on comparing the RHRI and ACIS-S data and it has
213: expanded by $1.98\pm0.64$\% based on comparing the EHRI and ACIS-S
214: data. The quoted errors are purely statistical at 1 sigma and include
215: both Poisson error (dominated by the EHRI and RHRI data) as well as
216: the uncertainty in the plate scales. The two points are described best
217: by a uniform expansion rate of $0.100\pm0.025$\% yr$^{-1}$. The null
218: hypothesis that the remnant has remained constant in size can be
219: rejected at the $4\sigma$ confidence level.
220:
221: It is extremely unlikely that systematic effects could account for
222: these results. For example, the \chandra\ plate scale would have to
223: be incorrect by more than 1\%, which is 50 times the statistical
224: uncertainty on the calibration measurement. We investigated
225: systematic errors due to uncertainty in the analytical PRF models and
226: in the effective area functions and the boxes plotted in fig.~1 show
227: an estimate of their likely effect on the expansion measurements.
228: Neither is expected to be a dominant source of error.
229:
230: \section {POST-SHOCK ELECTRON TEMPERATURE}
231:
232: The eastern limb of \eohone\ in the \chandra\ image (Blair et
233: al.~2000, Gaetz et al.~2000) displays a smooth, nearly circular, faint
234: rim of X-ray emission. Beyond the edge of the rim the surface
235: brightness drops by more than two orders of magnitude to the average
236: background level over a distance of order 1$^{\prime\prime}$. Herein
237: we identify this jump as the SN shock wave plowing into the
238: ambient medium (hereafter ``blast wave'').
239:
240: We extracted the spectrum of the blast wave region from within a
241: partial annulus of thickness $1\myarcsec5$ and outer radius
242: 21$^{\prime\prime}$ that covered the angular range from $-$30$^\circ$
243: to 155$^\circ$ (angles measured counterclockwise from north). We
244: ignored the background since it was negligibly small (less than 1\%).
245: The spectral data were contained within a single readout node of chip
246: S3, nevertheless the spectrum included data from several detector
247: regions with potentially different intrinsic responses to incident
248: X-rays. Individual response functions (both effective area functions
249: and spectral redistribution matrices) were generated for the several
250: regions covered by the extraction region and then combined, weighting
251: by the number of X-ray events, into a single global response function.
252: The results we obtain are insensitive to this procedure:
253: our fitted parameters are nearly unchanged if we use the response
254: functions from any individual extraction region. Because
255: of the limited statistical precision of our spectrum (which contained
256: only $\sim$1500 detected events) the uncertainty on derived parameters
257: is dominated by Poisson counting errors rather than calibration
258: uncertainties.
259:
260: Figure 2 plots the blast wave spectrum and best-fit nonequilibrium
261: ionization model (NEI) (Hughes \& Singh 1994). The spectrum shows a
262: well-resolved \ion{O}{8} Ly$\alpha$ line at 0.653 keV, a blend of
263: \ion{Ne}{9} and \ion{Ne}{10} K-shell lines in addition to Fe L-shell
264: lines around 1 keV, and the \ion{Mg}{11} He-like 2$\rightarrow$1 line
265: complex at $\sim$1.34 keV. The mere presence of these lines suggests
266: a moderately low electron temperature, $kT \sim 1$ keV, as our
267: detailed fits confirm. Table 2 gives best-fit parameters and 1-sigma
268: statistical errors determined from two classes of NEI models: (1) a
269: ``single ionization timescale'' model which assumes a single, constant
270: value for both the temperature and ionization timescale in the blast
271: wave region and (2) a ``planar shock'' model which also assumes a
272: constant temperature, but integrates the emission from spectral
273: components with ionization timescales varying linearly from 0 to
274: $n_et$ (the fit parameter) and weighted by equal fractional intervals
275: in $n_et$. The first model provides a marginally acceptable fit to
276: the \chandra\ data (the $\chi^2$ of 57 for 38 degrees of freedom can
277: be rejected at the 97.5\% confidence level). The second model, with
278: no additional parameters, yields a much improved, and statistically
279: acceptable, fit.
280:
281: Elemental abundances were determined only for the species with obvious
282: line emission in the observed spectrum, viz.\ O, Ne, Mg, and Fe. The
283: abundances of the other astrophysically common species were fixed at
284: values appropriate to interstellar gas in the SMC (Russell \& Dopita
285: 1992), an assumption that resulted in a significantly better fit than
286: one where the abundances of the other species were fixed to standard
287: solar abundances. The line-of-sight absorbing column density, $N_{\rm
288: H}$, was separated into a component with normal solar composition
289: arising from the Galaxy and a component from the SMC with lower
290: abundances. The Galactic component was fixed at an $N_{\rm H}$ value
291: consistent with reddening estimates and the Galactic \ion{H}{1} column
292: to \eohone, while the SMC component was allowed to vary freely.
293:
294: The abundances from both model fits are nicely consistent with the
295: well-known low metallicity of the gas in the SMC ($\sim$25\% solar),
296: strongly suggesting that the emission we see from the outer rim comes
297: from the shock wave propagating through interstellar
298: gas.\footnotemark\footnotetext{One should be wary of interpreting the
299: fitted abundances much beyond this. In particular, the Fe abundance
300: comes from a complex of L-shell lines that are unresolved at the
301: modest spectral resolution of our data. Current models (including
302: ours) are known to be incomplete as regards the Fe L-shell emission
303: (Brickhouse et al.~2000), which introduces additional uncertainty
304: beyond the statistical errors in the measured value of the
305: Fe-abundance.} Further in from the blast wave where the X-ray
306: emission of \eohone\ has become quite bright, the spectral character
307: has changed as well. In fig.~2 the spectrum from the bright region in
308: the southeast quadrant is plotted (labeled ``ejecta'').
309: The superposed model spectrum, shown for comparison, is the blast wave
310: NEI model scaled in intensity to fit the continuum level of the ejecta
311: spectrum. There are clear differences between the ejecta and blast
312: wave spectra in terms of both thermodynamic state (note, for example,
313: the strength of the \ion{O}{7} He-like 2$\rightarrow$1 line complex at
314: $\sim$0.57 keV in the ejecta which is virtually absent in the blast
315: wave) and the elemental abundances (especially of O and Ne). The
316: enhanced abundances and less advanced ionization state of the ejecta
317: are consistent with expectations for emission from the reverse shock
318: propagating into higher density, metal-rich gas.
319:
320: \section{DISCUSSION}
321:
322: The X-ray expansion rate of \eohone\ is fully consistent with the
323: remnant's proper motion estimated from the location and radial
324: velocities of the oxygen-rich optical filaments. Specifically our
325: estimate of the remnant's age assuming free expansion,
326: $1000^{+340}_{-200}$ yr, agrees with that of Tuohy \& Dopita (1983).
327: At the outermost edge of the X-ray emission from \eohone\ ($R\approx
328: 22^{\prime\prime} = 6.4 \rm \, pc$), the inferred velocity is
329: $6200^{+1500}_{-1600}$ km s$^{-1}$, which we consider in the following
330: to be the speed of the blast wave. According to the Rankine-Huguniot
331: jump conditions in the absence of cosmic ray acceleration (Landau \&
332: Lifshitz 1976), this velocity corresponds to a mean post-shock
333: temperature of $kT_S = {3\over 16}\, \mu m_p v_S^2 = 45^{+25}_{-20}$
334: keV, using a mean mass per particle of $\mu =0.61$. Our measurement
335: of the post-shock electron temperature from the \chandra\ X-ray
336: spectrum, $kT\ \lsim 1$ keV, is at least 25 times smaller than this
337: estimate.
338:
339: There are two principle assumptions that go into our estimate for the
340: mean temperature: (1) that the post-shock thermal energy is
341: partitioned equally between electrons and ions, and (2) that only a
342: negligible fraction of the total energy of the shock goes into cosmic
343: rays.
344:
345: It has been debated whether electron and ion temperatures are quickly
346: equilibrated by plasma processes behind collisionless high Mach number
347: SN shocks (McKee 1974; Cargill \& Papadopoulos 1988) or if the
348: electron and ion temperatures initially differ by their mass ratio
349: ($m_p/m_e = 1836$) and then equilibrate slowly through Coulomb
350: collisions (Shklovskii 1968). For \eohone\ we find, ignoring for the
351: moment any Coulombic heating of the electrons downstream, that the
352: ratio of electron thermal energy to upstream kinetic energy, $E_{\rm
353: e,th}/E_{\rm kin} = kT_e/({1\over 2}\,m_pv_S^2)\ \lsim1\%$, is much
354: smaller than the value of 12\% expected by Cargill \& Papadopoulos
355: (1988).
356:
357: Downstream of the blast wave, electrons and ions (mainly protons) will
358: exchange energy through Coulomb interactions, which provide the
359: minimum level of heating expected in the case of nonequipartition.
360: The electron temperature varies as
361:
362: $${dT_e \over dn_et} = 0.13 {T_p - T_e \over T_e^{3/2}}$$
363:
364: \noindent
365: (in cgs units) (Spitzer 1978, p.~22), where we express the temperature
366: variation in terms of $n_et$, since we have an observational estimate
367: of this quantity from the X-ray spectral fits (Table 2). This
368: equation predicts that the mean post-shock electron temperature
369: (averaging over $n_et$ from 0 to $4\times 10^{11}\rm\, cm^{-3}\, s$)
370: would be between 4.5 keV and 8 keV for proton temperatures
371: corresponding to the measured shock velocity (i.e., $T_p = 40$--120
372: keV). Note that the dependence of electron temperature on the actual
373: value of $n_et$ up to which the averaging is done is weak. If the
374: lower limit on the measured $n_et$ value, $2\times 10^{11}\rm\,
375: cm^{-3}\, s$, from our spectral fits is used, then the average $kT_e$
376: range becomes 3.5--6 keV. This estimate assumes that the total
377: pressure ($n_pT_p + n_eT_e$) remains constant in the post-shock
378: region, when, in fact, some adiabatic decompression should occur as
379: the shocked gas elements expand outward. Estimates of this effect,
380: assuming the post-shock gas acquires a speed of $3/4~v_S$, indicate
381: that the predicted electron temperatures may be reduced to no lower
382: than $\sim$2.5 keV, closer to, but still considerably higher than, the
383: measured blast wave electron temperature in \eohone. These simple
384: models suggest that post-shock proton temperatures of 40--120 keV are
385: just too high and that they need to be considerably reduced in order
386: to match the \chandra\ measurements. This in turn argues that cosmic
387: ray production is {\it not} negligible in \eohone\ and that the
388: relativistic particles may be absorbing a significant fraction of the
389: SN shock energy.
390:
391: Recently Decourchelle et al.~(2000) have shown that the fraction of
392: energy going into cosmic rays is likely to be high in young SNRs and
393: that this effect has direct consequences for the thermal X-ray
394: emission. In such a situation, the shock jump conditions are modified
395: (e.g., Blandford \& Eichler 1987) so that for a given shock velocity
396: the compression ratio increases and the post-shock temperature
397: decreases. Nonlinear models of shock acceleration (Ellison 2000)
398: appear fully consistent with the observed shock parameters of \eohone.
399: In particular, these models predict a mean post-shock temperature of 1
400: keV (see figure 1 in Ellison 2000) for standard cosmic ray injection
401: efficiencies and a high Mach number shock, i.e., values of 100--300,
402: as appropriate for \eohone. Thus it appears that efficient cosmic ray
403: acceleration alone is sufficient to explain our observations of
404: \eohone, even without the additional influence of nonequipartition of
405: electron and ion temperatures.
406:
407: Unfortunately, a direct measurement of the post-shock proton
408: temperature in \eohone\ is not possible, due to the absence of any H
409: line emission from the remnant. Therefore, constraints on the
410: efficiency of cosmic ray production in \eohone\ will need to come from
411: future detailed studies of the compression ratio, the structure of the
412: forward and reverse shocks, and the X-ray line spectrum using NEI
413: hydrodynamical models of SNRs that include cosmic-ray acceleration
414: (e.g., Decourchelle et al.~2000). In addition, studies of other types
415: of remnants, in particular Balmer-dominated ones for which estimates
416: of the proton temperature are available from H$\alpha$ line profiles,
417: should prove helpful in further elucidating the physics of
418: collisionless shock fronts.
419:
420: \acknowledgements
421:
422: Helpful discussions with David Burrows, Don Ellison, Maxim
423: Markevitch, Jon Morse, and Wallace Tucker on the scientific content of
424: the article are gratefully acknowledged. We appreciate Monique
425: Arnaud's support and hospitality during the course of this project.
426: This work was partially supported by NASA Grant NAG5-6420.
427:
428: \vfill\eject
429:
430: \begin{references}
431:
432: \reference{}
433: Blair, W.~P., et al.~2000, ApJ, in press
434:
435: \reference{}
436: Blandford, R.~D. \& Eichler, D.~1987, Phys. Reports, 154, 1
437:
438: \reference{}
439: Brickhouse, N.~S., Dupree, A.~K., Edgar, R.~J., Liedahl, D.~A., Drake,
440: S.~A., White, N.~E., \& Singh, K.~P.~2000, ApJ, 530, 387
441:
442: \reference{}
443: Cargill, P.~J., \& Papadopoulos, K.~1988, ApJ, 329, L29
444:
445: \reference{}
446: Chevalier, R.~A.~1983, ApJ, 272, 765
447:
448: \reference{}
449: David, L.~P., et al.~1998, The ROSAT High Resolution Imager (HRI)
450: Calibration Report
451: (http://hea-www.harvard.edu/rosat/rsdc-www/hricalrep.html)
452:
453: \reference{}
454: Decourchelle A., Ellison, D.~C., \& Ballet, J.~2000, ApJL, submitted.
455:
456: \reference{}
457: Dewey, D.~1999, PSF Studies...Encircled Energy Curves: Expected and Measured,
458: AXAF Science Center Calibration memo from 8/19/99
459:
460: \reference{}
461: Dopita, M.~A., Tuohy, I.~R., \& Mathewson, D.~S.~1981, ApJ, 248, L105
462:
463: \reference{}
464: Ellison, D.~C. 2000, in AIP Conf. Proc. 528, Acceleration and
465: Transport of Energetic Particles Observed in the Heliosphere,
466: ed.\ R.~A. Mewaldt, J.~R. Jokipii, M.~A. Lee, E.~Moebius,
467: \&~T.~H. Zurbuchen, in press (astro-ph/0003214)
468:
469: \reference{}
470: Gaetz, T.~J., Butt, Y.~M., Edgar, R.~J., Eriksen, K.~A., Plucinsky,
471: P.~P., Schlegel, E.~M., \& Smith R.~K.~2000, ApJ, 534, L47
472:
473: \reference{}
474: Garmire, G.~P.~1997, AAS Meeting, 190, 34.04
475:
476: \reference{}
477: Giacconi, R., et al.~1979, \apj, 230, 540
478:
479: \reference{}
480: Hayashi, I., Koyama, K., Ozaki, M., Miyata, E., Tsunemi, H.,
481: Hughes, J.~P., \& Petre, R.~1994, \pasj, 46, L121
482:
483: \reference{}
484: Henry, J.~P., \& Henriksen, M.~J.~1986, \apj, 301, 689
485:
486: \reference{}
487: Hughes, J.~P.~1988, in IAU Colloq.~101, Supernova Remnants and the
488: Interstellar Medium, ed.~R.~S.~Rogers \& T.~L.~Landecker (Cambridge:
489: Cambridge University Press), 125.
490:
491: \reference{}
492: Hughes, J.~P.~1999, \apj, 527, 298
493:
494: \reference{}
495: Hughes, J.~P., \& Singh, K.~P.~1994, ApJ, 422, 126
496:
497: \reference{}
498: Landau, L.~D., \& Lifshitz, E.~M.~1976, Fluid Mechanics, (New York:
499: Pergamon Press)
500:
501: \reference{}
502: McKee, C.~E.~1974, ApJ, 188, 335
503:
504: \reference{}
505: Russell, S.~C., \& Dopita, M.~A.~1992, ApJ, 384, 508
506:
507: \reference{}
508: Shklovskii, I.~S.~1968, Supernovae (New York: Wiley)
509:
510: \reference{}
511: Spitzer, L.~1978, Physical Processes in the Interstellar Medium (New
512: York: Wiley-Interscience)
513:
514: \reference{}
515: Tuohy, I.~R., \& Dopita, M.~A.~1983, \apj, 268, L11
516:
517: \reference{}
518: van den Bergh, S.~2000, PASP, 112, 529
519:
520: \reference{}
521: Weisskopf, M.~C., O'Dell, S.~L., \& Van Speybroeck, L.~P.~1996,
522: Proc.~SPIE, 2805, 2
523:
524: \end{references}
525:
526: \newpage
527:
528: \begin{deluxetable}{cccc}
529: \tablecaption{Observations of \eohone}
530: \tablewidth{4.5truein}
531: \tablehead{
532: \colhead{Observatory} & \colhead{Start Date} &
533: \colhead{Average MJD} & \colhead{Duration (s)}
534: }
535: \startdata
536: {\it Einstein} (EHRI) & 1980 Apr 18 & 44347.5 & $22500$ \nl
537: {\it ROSAT} (RHRI) & 1991 Nov 9 & 48598.2 & $21900$ \nl
538: {\it ACIS-S} & 1999 Aug 23 & 51413.9 & $\phantom{0}8883$ \nl
539: \enddata
540: \end{deluxetable}
541:
542: \newpage
543:
544: \begin{deluxetable}{ccc}
545: \tablewidth{0pt}
546: \tablecaption{NEI Models of E0102.2-0.5\label{neimodels}}
547: \tablehead{
548: \colhead{} & \colhead{Single Ionization} &
549: \colhead{}\\
550: \colhead{Parameters} & \colhead{Timescale} &
551: \colhead{Planar Shock}
552: }
553: \startdata
554: $kT$ (keV)
555: & $0.48^{+0.08}_{-0.05}$
556: & $0.78^{+0.16}_{-0.15}$ \\
557: $n_{e}t$ (cm$^{-3}$ s)
558: & $4^{+21}_{~-2}\times10^{11}$
559: & $4^{+4}_{-2}\times10^{11}$\\
560: $N_{\mathrm{H,SMC}}$\tablenotemark{a}~ (cm$^{-2}$)
561: & $(5\pm 3)\times10^{20}$
562: & $<2.5\times10^{20}$ \\
563: %$N_{\mathrm{H,Gal}}$\tablenotemark{b}~ (cm$^{-2}$)
564: % & $5\times10^{20}$ & $5\times10^{20}$ \\
565: % Emission Integral & & \\
566: $n_{\mathrm{e}}n_{\mathrm{H}}$V/4$\pi$D$^{2}$ (cm$^{-5}$)
567: & $1.1^{+0.4}_{-0.3}\times10^{11}$
568: & $4.1^{+2.0}_{-0.8}\times10^{10}$ \\
569: Oxygen
570: & $0.26^{+0.19}_{-0.14}$
571: & $0.34^{+0.15}_{-0.12}$ \\
572: Neon
573: & $0.31^{+0.11}_{-0.09}$
574: & $0.67^{+0.20}_{-0.19}$\\
575: Magnesium
576: & $0.28^{+0.12}_{-0.10}$
577: & $0.47\pm0.16$ \\
578: Iron
579: & $0.01\pm0.005$
580: & $0.06\pm0.03$ \\
581: $\chi ^{2}$ (d.o.f.)
582: & 57.13 (38) & 46.07 (38) \\
583: \enddata
584: \tablenotetext{a}{Uses mean SMC abundances; fits also include a fixed
585: $N_{\mathrm{H,Gal}}$ of $5\times10^{20}$ cm$^{-2}$ using solar abundances}
586: %\tablenotetext{b}{Fixed, uses solar abundances}
587: \end{deluxetable}
588:
589: \newpage
590:
591: \clearpage
592:
593: \figcaption[f1.ps]{Global mean percentage expansion of SNR \eohone\
594: versus the time difference between imaging measurements. The point
595: indicating the shortest time difference comes from comparing the {\it
596: ROSAT} and \chandra\ observations. The other point comes from
597: comparing the {\it Einstein} and \chandra\ observations. The error
598: bars show the statistical uncertainty, while the boxes that surround
599: each data point give an estimate of the systematic uncertainty. The
600: dotted curves show the best-fit percentage expansion and 1 sigma
601: errors. The inferred speed of the shock front is $\sim$6000 km
602: s$^{-1}$ and the remnant's age is 1000 yr.
603: \label{Figure 1}}
604:
605: \figcaption[f2.ps]{\chandra\ spectra of a portion of the outer blast
606: wave and bright ejecta of SNR \eohone\ as indicated. The spectra were
607: rebinned to a minimum of 25 events per channel. The best-fit
608: nonequilibrium ionization model for the blast wave is shown as the
609: solid histogram. The model appears compared against the ejecta
610: spectrum as well in order to demonstrate the gross spectral
611: differences between the two regions. The residuals are shown only for
612: the blast wave spectrum.
613: \label{Figure 2}}
614:
615: %\end{document}
616:
617: \clearpage
618:
619: \begin{figure}
620: \plotfiddle{./fig1.ps}{5in}{0}{100}{100}{-300}{-200}
621: \end{figure}
622:
623: \clearpage
624:
625: \begin{figure}
626: \plotfiddle{./fig2.ps}{5in}{0}{100}{100}{-300}{-200}
627: \end{figure}
628:
629: \end{document}
630: