1: \documentstyle[11pt,aasms4]{article}
2:
3: % Definitions of subexpressions
4: \newcommand{\gsim}{\mbox{\hspace{.2em}\raisebox{.5ex}{$>$}\hspace{-.8em}\raisebox{-.5ex}{$\sim$}\hspace{.2em}}}
5: % \newcommand{\gsim}{\stackrel{>}{\sim}}
6: \newcommand{\lsim}{\mbox{\hspace{.2em}\raisebox{.5ex}{$<$}\hspace{-.8em}\raisebox{-.5ex}{$\sim$}\hspace{.2em}}}
7: % \newcommand{\lsim}{\stackrel{<}{\sim}}
8: \newcommand{\ssst}{\scriptscriptstyle}
9: \newcommand{\E}[1]{\times 10^{#1}}
10: \newcommand{\etal}{et al.}
11: \newcommand{\lt}{\left} \newcommand{\rt}{\right}
12: \newcommand{\RA}[3]{\mbox{R.A.}={#1}^{{\rm h}}{#2}^{{\rm m}}{#3}^{{\rm s}}}
13: \newcommand{\decl}[3]{\mbox{decl.}={#1}^{\circ}{#2}'{#3}''}
14:
15: % Definitions of units
16: \newcommand{\s}{\,{\rm s}} \newcommand{\ps}{\,{\rm s}^{-1}}
17: \newcommand{\yr}{\,{\rm yr}} \newcommand{\Msun}{M_{\odot}}
18: \newcommand{\cm}{\,{\rm cm}} \newcommand{\km}{\,{\rm km}}
19: \newcommand{\parsec}{\,{\rm pc}}\newcommand{\kpc}{\,{\rm kpc}}
20: \newcommand{\ergs}{\,{\rm ergs}} \newcommand{\K}{\,{\rm K}}
21: \newcommand{\eV}{\,{\rm eV}} \newcommand{\keV}{\,{\rm keV}}
22:
23: % Definitions of symbols
24: \newcommand{\nel}{n_{e}} \newcommand{\NH}{N_{\ssst\rm H}}
25: \newcommand{\no}{n_{\ssst 0}} \newcommand{\rPDS}{r_{\ssst\rm PDS}}
26: \newcommand{\vPDS}{v_{\ssst\rm PDS}} \newcommand{\tPDS}{t_{\ssst\rm PDS}}
27: \newcommand{\Ts}{T_{s}}
28: \newcommand{\rs}{r_{s}} \newcommand{\vs}{v_{s}}
29: \newcommand{\nH}{n_{\ssst\rm H}} \newcommand{\mH}{m_{\ssst\rm H}}
30: \newcommand{\nHH}{n({\rm H}_{2})} \newcommand{\NHH}{N({\rm H}_{2})}
31: \newcommand{\xray}{X-ray} \newcommand{\Einstein}{{\em Einstein}}
32: \newcommand{\ROSAT}{{\sl ROSAT}} \newcommand{\ASCA}{{\sl ASCA}}
33: \newcommand{\du}{d_{8}} \newcommand{\Eu}{E_{51}}
34: \newcommand{\lpc}{l_{\rm pc}} \newcommand{\ru}{r_{3}}
35: \newcommand{\mE}{\langle E_{\ssst B-V}\rangle}
36:
37: \begin{document}
38:
39: \title{\ASCA\ Observations of the Thermal Composite Supernova Remnant 3C~391}
40:
41: \author{
42: Yang Chen\altaffilmark{1} and
43: Patrick O.\ Slane\altaffilmark{2}
44: }
45: \altaffiltext{1}{
46: Department of Astronomy, Nanjing University, Nanjing 210093, P.R.China;
47: email: ygchen@nju.edu.cn
48: }
49: \altaffiltext{2}{
50: Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
51: Cambridge, MA 02138;
52: email: slane@cfa.harvard.edu
53: }
54:
55: \vfil
56: %Text
57: \begin{abstract}
58: We present the results from ASCA observations of the centrally enhanced
59: supernova remnant 3C~391 (G31.9$+$0.0).
60: We use the \ASCA\ SIS data to carry out an investigation of the spatial
61: and spectral properties of the X-ray emission from this remnant.
62: The collisional equilibrium ionization and non-equilibrium ionization
63: spectral fits indicate that the hot gas within the
64: remnant has basically reached ionization equilibrium.
65: The variation of the hydrogen column density across the remnant
66: is in agreement with the presence of a molecular cloud to the northwest.
67: The comparisons of
68: hydrogen column and \xray\ hardness between the NW and SE portions
69: of the remnant support a scenario in which the SNR has broken out of
70: a dense region into an adjacent region of lower density.
71: The mean density within the SNR is observed to be much lower
72: than the immediate ambient cloud density. This and the centrally brightened
73: \xray\ morphology can be explained either by the evaporation of
74: engulfed cloudlets or by a radiative stage of evolution for the remnant.
75:
76: \keywords{
77: radiation mechanisms: thermal ---
78: supernova remnants: individual: 3C~391 (G31.9+0.0) ---
79: X-rays: ISM
80: }
81:
82: \end{abstract}
83:
84: \section{Introduction}
85:
86: Apart from the well-known classification of
87: shell-like, Crab-like, and composite supernova remnants (SNRs),
88: a collection of SNRs with shell-like radio emission and centrally brightened
89: thermal X-rays has attracted substantial attention
90: (e.g., White \& Long 1991 [=WL91];
91: Rho \& Petre 1998; Wilner, Reynolds, \& Moffett 1998).
92: Many of these so-called ``thermal composite'' or ``mixed morphology''
93: remnants (such as W28, W44, IC~443, G349.7+0.2, and 3C~391) are found to produce
94: the hydroxyl radical 1720 MHz maser emission that is characteristic of a shock
95: interaction with dense molecular gas (Green \etal\ 1997).
96:
97: 3C~391 (G31.9+0.0) is similar to other remnants in this
98: collection. Its X-ray brightness peaks well inside the radio
99: shell, and the spectrum is that of a thermal plasma. The remnant
100: shows an elongation from northwest (NW) to southeast (SE) in the
101: radio band, and the centroid of the soft X-ray emission lies in
102: the SE region. Reynolds \& Moffett (1993) (= RM93) first pointed
103: out that the radio and \xray\ morphologies can be explained with a
104: gas breakout from a molecular cloud into a lower density region.
105: In a \ROSAT\ \xray\ study, Rho \& Petre (1996) (= RP96) could not
106: actually distinguish between an increase in hydrogen column
107: density or a decrease in temperature across the remnant from SE to
108: NW from the spectra alone, and favored the hydrogen column
109: variation based on the morphology. Frail et al. (1996) suggest
110: that OH masers in the direction of 3C~391 are associated with the
111: remnant, and Reach \& Rho (1996) find strong enhancement in [OI]
112: emission near the northwestern edge of the remnant, which is
113: indicative of shock interaction. CO and other molecular line
114: observations confirm the location of the remnant at the
115: southwestern edge of a molecular cloud (Wilner \etal\ 1998; Reach
116: \& Rho 1999). The far-infrared H$_{2}$O and OH emission lines are
117: consistent with the passage of shock wave through dense clumps
118: (Reach \& Rho 1998), and broad CS and CO line sources are found
119: to coincide with one of the OH maser regions (Reach \& Rho 1999).
120:
121: Because of the limited spectral resolution of the \ROSAT\ PSPC
122: (0.5-2.2 keV) and \Einstein\ IPC ($<4\keV$), the previous \xray\
123: instruments did not resolve line features in the SNR spectrum.
124: \ASCA, with its broader energy range (0.5-10 keV) and better
125: energy response, allows us to analyze the line emission, element
126: abundances, and the narrow band images to investigate the physics
127: of the hot gas inside 3C~391 and inspect the breakout model more
128: fully.
129:
130: \section{Data Analysis}
131: The \ASCA\ observation of 3C~391 was made on 21-22 April 1994 with both
132: the solid-state imaging spectrometer (SIS) and the gas imaging
133: spectrometer (GIS).
134: In the SIS observation, both 1-CCD and 4-CCD clocking modes were used.
135: The observational data were screened with the standard ``rev2'' criteria
136: (see \ASCA\ ABC guide).
137: The SIS0 and SIS1 bright mode data were exploited in this analysis due to
138: the better spectral and spatial resolution of the SIS data over the GIS data.
139: The relevant observation parameters are listed in Table 1a.
140:
141: \subsection{Spectral Analysis}\label{sec:spec}
142: The spectra for the whole remnant (shown in Figure~1) were extracted from
143: a circular region of a radius 3.5~arcmin,
144: centered at $\RA{18}{49}{28}.3$, $\decl{-00}{56}{15}$ (J2000)
145: (see the diagram in Figure~3{\em a} below).
146: Background spectra were extracted from a nearby source-free field.
147: The net count rates of the two CCD mode SIS0 and SIS1 spectra
148: are tabulated in Table~1b.
149: Since data of different clocking modes have different spectral responses,
150: we would not merge the spectra but fit them simultaneously to the same model.
151:
152: In order to search for spectral variations across the remnant,
153: we also extracted spectra from two circular regions of
154: radius 1.7~arcmin in the NW and SE halves (as diagramed in Figure~3{\em a}).
155: The net count rates of these spectra are also tabulated in Table~1b.
156: All the spectra mentioned above were regrouped to contain at least
157: 25 net counts per bin.
158:
159: There are three prominent line features in the spectra.
160: To determine accurately the energy of the lines,
161: following Bamba \etal\ (2000), we first fitted the spectra of the whole
162: remnant to a thermal bremsstrahlung and three Gaussian lines with a
163: Morrison \& McCammon (1983) interstellar absorption.
164: The best-fit line centers are $1.35\pm0.01$, $1.85\pm0.01$,
165: and $2.46\pm0.02\keV$ (the errors are 90\% confidence).
166: They correspond to Mg He$\alpha$, Si~He$\alpha$, and S~He$\alpha$
167: emission lines.
168: The Fe L complex at 1-$1.5\keV$ could be present in the spectra.
169: The emission diminishes rapidly above the photon energy $\sim 4\keV$
170: and no Fe~K$\alpha$ emission is observed,
171: indicating that the gas temperature is not very high assuming
172: normal abundances (below $1\keV$).
173: We used an absorbed collisional equilibrium ionization (CEI) model,
174: VMEKAL (Mewe, Kaastra, \& Liedahl 1995), in the XSPEC code to
175: fit the two clocking mode SIS0 and SIS1 spectra of the whole remnant
176: simultaneously.
177: In the spectral fitting, the abundances of the elements (Mg, Si, \& S)
178: showing evident emission lines, are treated as free parameters,
179: while other element abundances are fixed at the default solar values.
180: The spectral fit for the whole remnant are shown in Figure~1.
181: Similar fit procedures are also applied to the spectra of
182: the NW and SE portions individually.
183: We find no significant contribution of a high energy tail that might be
184: associated with a strong nonthermal component to the spectrum.
185: The CEI fit results are summarized in Table 2.
186:
187: On the other hand, we also fit these spectra to VNEI, a non-equilibrium
188: ionization (NEI) model, in the XSPEC11.0 code.
189: The NEI fit results are presented in Table 3.
190: The goodness-of-fit and most of the fit parameters are
191: similar for the CEI and NEI cases, and
192: the ionization parameter $\nel t$ obtained from the NEI
193: models is $>10^{12}\cm^{-3}\s$,
194: so the hot gas in the remnant is basically in
195: ionization equilibrium.
196:
197: The metal abundances yielded with both the CEI and NEI models (see Tables
198: 2 \& 3) are essentially consistent with solar values and thus seem to be
199: consistent with an interstellar composition.
200: From the two tables we can not see significant difference in temperature
201: between the SE and NW regions of 3C~391 in either model,
202: but find different error ranges of $\NH$ for the two regions
203: (with higher $\NH$ in the NW than in the SE).
204: To check this, we compute the two-dimensional error contours for
205: $\NH$ and $kT$ for the SE and NW regions (Figure~2).
206: In both the CEI and NEI cases, the confidence contours for the two regions
207: are clearly disjoint, so $kT$ and $\NH$ highly correlated for each region.
208: The contours show similar range of temperature but different range of
209: column density for the two regions.
210: We also fit the NW and SE spectra simultaneously to the same temperature
211: and the same column density, respectively, using a CEI model;
212: here the metal abundances for the two regions given in Table~2 are used.
213: The best-fit results are summerized in Table~4.
214: For the case of the same temperature, one would again see that the
215: one dimensional error ranges of $\NH$ for the two regions do not overlap.
216: Assuming the same column density, on the other hand, leads to another
217: possible case in which the NW region is higher in temperature but lower
218: in emission measure than the SE.
219: In the latter case, considering the immediate adjacency of the molecular cloud
220: to the northwest of the remnant (Wilner \etal\ 1998),
221: it is quite unlikely that the gas density the NW region is lower than
222: that in the SE region,
223: so the hot gas in the defined circular NW region should actually occupy
224: larger volume than in the SE region.
225: Though this possibility can not be eliminated, a variation in column density
226: between the SE and NW region, ie., an increase in column density from
227: the SE to the NW is consistent with the proximity of the molecular cloud.
228:
229: \subsection{Spatial Analysis}\label{sec:img}
230: The SIS0 observation covers the entire remnant,
231: but the SIS1 observation misses a small part of the eastern rim.
232: We have thus used only SIS0 data for the image production.
233: The corrections for exposure and vignetting were made
234: for the two clocking modes (1-CCD and 4-CCD) separately, and
235: an adjustment was made to the reconstructed pointing direction
236: to correct for documented erroneous star-tracker readings (Gotthelf
237: et al.\ 2000).
238: The hard (2.6-10$\keV$) emission gray-scale image is overlaid with
239: the soft (0.5-2.6$\keV$) emission contours in Figure~3{\em a}.
240: The ``hard'' emission was extracted above the $2.6\keV$
241: so as to not include the lines of
242: Mg He$\alpha$, Si He$\alpha$, and S He$\alpha$
243: and can basically be regarded as thermal continuum.
244: In Figure~3{\em b} and Figure~3{\em c}, the 1.5 GHz radio contours
245: (from Moffett \& Reynolds 1994) are overlaid with the
246: soft and hard emission images, respectively.
247: We also present the
248: narrow band Mg He$\alpha$ (1.2-1.5$\keV$), Si He$\alpha$
249: (1.7-2.0$\keV$), and S He$\alpha$ (2.3-2.6$\keV$) images
250: in Figure~3{\em d}, Figure~3{\em e}, and Figure~3{\em f}.
251: The three energy ranges (ER1, ER2, and ER3), in which the narrow band
252: emissions are extracted, are labeled in Figure~1.
253: The three narrow band images contain 2308, 2587, and 833 counts,
254: individually.
255: Note that the three images are not corrected for the
256: underlying continuum contribution
257: and the Mg band (ER1) also includes emission from the Fe~L blend.
258: The Mg, Si, and S emissions have 56\%, 55\%, and 36\% energy
259: contributions, respectively, in each energy range.
260: All these image maps have been adaptively smoothed so as to contain a
261: minimum of 50 counts in the top hat filter.
262: The two OH maser points (Frail et al.\ 1996) are labeled in
263: the maps.
264:
265: \section{Discussion}
266: \subsection{Absorbing hydrogen}
267: Both the CEI and NEI models give a hydrogen column density $\NH$ around
268: $3.0\E{22}\cm^{-2}$ (Tables 2 \& 3).
269: In the spectral fits above, we show disjoint error ranges of $\NH$
270: for the NW and SE regions of 3C~391,
271: and favor a variation in $\NH$ between the two regions.
272: This is consistent with the northwestward enhancement of CO emission
273: (Wilner \etal\ 1998) and with the southeastward breakout
274: scenario (RM93).
275: In the CEI model fit, the best-fit value of $\NH$ for the NW portion is
276: $\sim4\E{21}\cm^{-2}$ higher than that for the SE portion;
277: and in the NEI model fit, this difference is $\sim6\E{21}\cm^{-2}$.
278: If the difference of $\NH$ for the two portions reflects the
279: density contrast between the inside and outside of the
280: molecular cloud,
281: the molecular column inside the cloud would be
282: $\NHH\sim3$-$4\E{20}\cm^{-2}$
283: (following RM93, considering that the photoionization cross section
284: for ${\rm H}_{2}$ at $1\keV$, per H atom,
285: is about about 8 times that for atomic hydrogen [Brown \& Gould 1970]).
286: In this way the mean molecular density inside the cloud would be
287: $\langle\nHH\rangle\sim100\lpc^{-1}\cm^{-3}$,
288: where $\lpc$ is the depth in pc of the molecular layer.
289: This would imply $\langle\nHH\rangle$ of order $\sim10$-$20\cm^{-3}$
290: if the line-of-sight depth of the SNR in the cloud is
291: similar to the remnant radius (6-9 pc).
292:
293: \subsection{Images}
294: The soft band map (Fig.3a,b) is very similar to the \ROSAT\ PSPC image (RP96).
295: The hard \xray\ image (Fig.3a,c) looks relatively bright in the NW
296: compared with that in the SE,
297: in contrast to the soft \xray\ image which looks faint in the NW.
298: This fact can be explained by increased extinction from the NW cloud,
299: and is consistent with a model in which the gas in the SE region
300: has broken out into a lower density environment.
301: There seems to be a narrow \xray\ bridge-like structure
302: at the center of the \xray\ images connecting the SE and NW portion,
303: which might possibly represents the ``nozzle" or ``tunnel" of
304: the NW-to-SE breakout.
305:
306: The emission dominated by Mg and Si (Fig.3d,e) is centrally brightened
307: in the SE half but faint in the NW (similar to the soft band map),
308: implying obscuration in the NW half.
309: An arc-like feature emerges in the NW of the S He$\alpha$ emission map
310: (Fig.3f), while it is absent in the Mg emission map (Fig.3d)
311: and marginally evident in the Si emission map (Fig.3e).
312: This is consistent with an absorption effect in the softer
313: Mg and Si lines.
314: This arc-like shaped feature is close to the NW radio shell
315: and reveals a hot gas structure behind the blastwave and
316: embedded in the dense cloud.
317:
318: In each \xray\ map the SE half is centrally brightened.
319: The hard emission (continuum) in the NW is also centrally brightened,
320: and is conspicuously faintest along the western limb where the radio
321: brightness is a maximum.
322: The mechanism
323: responsible for the enhanced central emission
324: may be interior cloudlet evaporation (WL91)
325: or, perhaps, the cooling of the rim gas (e.g.\ Harrus et al. 1997,
326: Rho \& Petre 1998, Cox et al. 1999, Shelton et al. 1999).
327: We discuss these two mechanisms in \S\ref{sec:dyn}).
328:
329: The peak of the X-ray brightness distribution is located at
330: about $\RA{18}{49}{33}.5$, $\decl{-00}{56}{37}$ (J2000)
331: for both the soft map and the narrow band maps,
332: and the hard map peaks at about
333: $\RA{18}{49}{25}$, $\decl{-00}{54}{32}$ (J2000).
334: The association of 3C~391 with a dense molecular cloud
335: is suggestive of a massive progenitor star.
336: The gravitational core collapse of the massive progenitor
337: should have left behind a compact star,
338: such as that observed for IC~443 (Olbert et al.\ 2001),
339: another ``thermal composite''
340: which is similar to 3C~391 in many aspects.
341: While the broad spatial response of ASCA prohibits a sensitive
342: search for an embedded compact star, future {\sl Chandra} and
343: {\sl XMM} observations near
344: the peaks in the soft and hard maps mentioned above are clearly
345: of interest.
346:
347: \subsection{Distance}
348: The presence of the HI absorption against 3C~391 out to the tangent point
349: velocity $\sim+105\km\ps$ puts the remnant beyond $7.2\kpc$ (for a
350: Galactocentric radius $8.5\kpc$),
351: and the absence of the absorption at negative velocities sets an
352: upper limit of $11.4\kpc$
353: (Caswell \etal\ 1971; Radhakrishnan \etal\ 1972; RM93).
354: This range is supported by the discovery of two OH 1720 MHz maser
355: features (at +105 and $+110\km\ps$) in the direction of 3C~391 (Wilner,
356: Reynolds, \& Moffett 1998).
357: With the hydrogen column density ($\NH\sim3\E{22}\cm^{-2}$)
358: obtained from the \xray\ spectral fits,
359: one can estimate another upper limit to the distance.
360: The extinction per unit distance within $2\kpc$ in the direction of 3C~391
361: is $\mE/d\sim0.60\,{\rm mag}\kpc^{-1}$ (Lucke 1978).
362: The extinction beyond $2\kpc$, in the direction of the Galactic center,
363: should be higher than this value.
364: The correlation $\NH=5.9\E{21}\mE\cm^{-2}$
365: (Spitzer 1978; Predehl \& Schmitt 1995) then
366: gives $d\lsim8.5\kpc$,
367: which is in agreement with the above range of distance.
368: In this paper $d\sim8\du\kpc$ will be used hereafter.
369:
370: \subsection{\xray\ emitting gas and its dynamics}\label{sec:dyn}
371: If the unabsorbed flux in the VMEKAL model (case CEI) is adopted,
372: the \xray\ (0.5-10 keV) luminosity is $L_{x}\sim2.7\E{36}\du^{2}\ergs\ps$.
373: The best-fit volume emission measure (EM) of the remnant
374: in the CEI case is $f\nel\nH V \sim 1.0\E{59}\du^{2}\cm^{-3}$,
375: where $f$ is the filling factor of the \xray\ emitting plasma;
376: the EM value in the NEI case is a bit smaller.
377: Based on the VLA observation (RM93),
378: we approximate the remnant volume as a cylinder of a diameter
379: $5'$ and a height $7'$.
380: With $\nel\approx1.2\nH$ assumed, the emission measure yields
381: $\nH\sim1.5f^{-1/2}\du^{-1/2}\cm^{-3}$.
382: The \xray\ emitting mass is
383: $M_{x}=1.4\nH\mH fV\sim92f^{1/2}\du^{5/2}\Msun$,
384: which indicates that the emission is
385: dominated by swept-up ambient matter.
386: (The NEI case would correspond to 0.86 times the above
387: gas density and mass.)
388: The spectral fits show a trend
389: of larger EM of the NW region than that of the SE region.
390: This may imply that the NW part of the remnant gas
391: is denser than the SE part.
392:
393: The mean interior hot gas density $\nH$ is much smaller than the environment
394: H-atom density $2\langle\nHH\rangle$ inside the molecular cloud,
395: indicating that a large amount of ambient matter does not act as
396: \xray\ emitting gas after being swept up or
397: engulfed by the supernova blastwave.
398: The explanation could be either that we are observing the hot, tenuous
399: internal gas while the dense material near the rim has cooled down,
400: or that the medium inside the cloud is clumpy and
401: most of the \xray\ emitting interclump medium (ICM) was
402: evaporated from the clumps.
403: The two mechanisms can both lead to the observed centrally brightened
404: \xray\ morphology, as we discuss below.
405:
406: \subsubsection{Cloud evaporation case}
407: Here we use the self-similar solution incorporating cloud evaporation
408: (WL91) to discuss the dynamics of SNR 3C~391.
409: Because cloud evaporation slowly increases the interior density, the mean
410: interior hot gas density could be a few times the postshock density.
411: This gas can provide the centrally emitting thermal X-ray component
412: observed ``mixed morphology'' remnants such as 3C~391 and others
413: (Harrus et al. 1997, Harrus et al. 2001).
414: The postshock temperature $\Ts$ can be derived from
415: the observed \xray\ emitting gas temperature $kT_{x}$($\sim0.52\keV$
416: in the CEI case)
417: using a scaling factor
418: \begin{equation}
419: q=KT_{x}/1.27\Ts,
420: \end{equation}
421: where $q$ and the energy ratio constant $K$ (scaled by Sedov value)
422: are dependent on $C/\tau$ (WL91).
423: Here $C$ is the ratio of the mass in the clumps to the mass of ICM,
424: and $\tau$ is the ratio of the cloud evaporation time to the SNR's age.
425: We follow RP96 and take $C/\tau\sim3$-5.
426: The velocity of the blastwave can be obtained from
427: $\vs=(16k\Ts/3\mu\mH)^{1/2}$
428: where the mean atomic weight $\mu=0.61$,
429: which then gives the dynamical age of the remnant
430: $t=2r_{s}/5\vs$.
431: Since the SNR is complicated in morphology, we scale the
432: radius with a mean value ($3'$) of the whole radio volume: $\rs=3'\ru$.
433: The hot gas density distribution is dependent upon the model parameters;
434: for simplicity, however, we assume the mean interior density is
435: twice the postshock density, which is broadly consistent with a range
436: of profiles for the cloud evaporation model.
437: Thus the undisturbed preshock ICM density $\no$ is about
438: $\nH/8\sim0.2f^{-1/2}\du^{-1/2}\cm^{-3}$.
439: This density can also be estimated from the \xray\ luminosity
440: $L_{x}$ ($\sim2.7\E{36}\du^{2}\ergs\ps$) using WL91's
441: eq.(21) and it is thus in a consistent range
442: $\sim0.1$-$0.2\ru^{-3/2}\du^{-1/2}$.
443: The explosion energy is given by
444: \begin{equation}
445: E=\frac{16\pi(1.4\no\mH)}{25(\gamma+1)K}\frac{r_{s}^{5}}{t^{2}},
446: \end{equation}
447: where the adiabatic index $\gamma=5/3$. The results of the
448: dynamical parameters obtained from the above relations are
449: tabulated in Table~5. The explosion energy (1.3-3.4)$\E{50}\ergs$
450: is somewhat lower than the canonical value of $10^{51}\ergs$. The
451: age estimate in RP96's evaporation model ($\sim6$-$8\E{3}\yr$) is
452: higher than that obtained here ($\sim4$-$5\E{3}\ru\du\yr$),
453: majorly because they used a larger-than-average radius of remnant.
454: RM93 used a Sedov model with a high preshock density, so their
455: estimate of age is rather large ($\sim1.7\E{4}\yr$).
456:
457: \subsubsection{Radiative rim case}
458: In this case we assume the clumpy mass is not important and
459: the mean molecular density $\langle\nHH\rangle$ would correspond to
460: a uniform undisturbed hydrogen density $\no\sim30\cm^{-3}$.
461: The radius of the SNR at the beginning of the radiative
462: pressure-driven snowplow (PDS) stage is given by
463: (Cioffi, McKee, \& Bertschinger 1988)
464: \begin{equation}
465: \rPDS=14.0\Eu^{2/7}\no^{-3/7}\zeta_{m}^{-1/7}\parsec,
466: \end{equation}
467: where $\zeta_{m}$ is the metallicity factor and is close to unity
468: for normal abundances. Adopting an explosion energy $\Eu\equiv
469: E/(10^{51}\ergs)\sim1$, we have $\rPDS\sim3.3\parsec$, which is
470: smaller than the remnant's radius $\rs\gsim6\du\parsec$ (here a
471: radius $2.5'$ of the NW radio shell is adopted). Thus, we expect
472: at least the NW part of the SNR, which appears to be in contact
473: with the dense cloud, to have already approached the PDS phase. In
474: fact, the \xray\ emission in the NW does not extend out to the
475: radio shell as it does in the SE. This is similar to what is
476: observed in W44, in which the remnant has entered the radiative
477: phase, shutting down the X-ray at the radio shell (Harrus et al.\
478: 1997). The newly detected near-infrared [Fe II] and the
479: mid-infrared 12-18 $\mu$m emission reveals the radiative shell of
480: 3C~391, particularly in the NW rim (Reach, Rho, \& Jarrett 2001).
481: We thus assume that the shell of 3C~391 has cooled sufficiently to
482: have reached the radiative phase. In this case the hot interior of
483: the remnant drives the cooled shell to expand and is responsible
484: for the centrally enhanced cooling (e.g.\ Cioffi et al.\ 1988).
485: Although ignored here, the effects of thermal conduction enhance
486: this process through smoothing of the interior temperature profile
487: (Cox et al. 1999, Shelton et al. 1999).
488:
489: The postshock temperature should be lower than that for
490: the Sedov case: $k\Ts<0.77kT_{x}\sim0.4\keV$;
491: the shock velocity $\vs$ is then slower than $\sim580\km\ps$.
492: In the PDS stage, $\rs$ and $\vs$ follow the formulae
493: (Cioffi et al.\ 1988)
494: \begin{eqnarray}
495: \rs &=& \rPDS\lt(\frac{4t}{3\tPDS}-\frac{1}{3}\rt)^{3/10}, \\
496: \vs &=& \vPDS\lt(\frac{4t}{3\tPDS}-\frac{1}{3}\rt)^{-7/10},
497: \end{eqnarray}
498: where
499: $\vPDS = 413\no^{1/7}\zeta_{m}^{3/14}\Eu^{1/14}\km\ps$ and
500: $\tPDS = 1.33\E{4}\Eu^{3/14}\no^{-4/7}\zeta_{m}^{-5/14}\yr$.
501: From the above, one has
502: \begin{equation}
503: \Eu = \lt(\frac{\rs}{14\parsec}\rt)^{98/31}
504: \lt(\frac{\vs}{413\km\ps}\rt)^{42/31}
505: \no^{36/31}\zeta_{m}^{5/31}, \\
506: \end{equation}
507: which yields $\Eu<9.2(\ru\du)^{98/31}$,
508: which is reasonable although it offers a weak overall constraint.
509: From Cioffi et al.\ (1988), the age of the remnant can readily be
510: obtained:
511: \begin{equation}
512: t=3.3\E{3}\lt(\frac{\rs}{14\parsec}\rt)
513: \lt(\frac{\vs}{413\km\ps}\rt)^{-1}
514: \lt[3 + \no^{-10/31} \zeta_{m}^{-10/31}
515: \lt(\frac{\rs}{14\parsec}\rt)^{-10/31}
516: \lt(\frac{\vs}{413\km\ps}\rt)^{40/31} \rt]\yr
517: \end{equation}
518: which yields $t>4.3\E{3}\ru\du\yr$.
519: If we use $\Eu$ as a parameter, then the present shock velocity is
520: $\vs\sim110\Eu^{31/42}(\ru\du)^{-7/3}\km\ps$
521: and the remnant age is
522: $t\sim1.9\E{4}\Eu^{-31/42}(\ru\du)^{10/3}\yr$.
523:
524: The cloud evaporation and the radiative rim mechanisms
525: both produce acceptable dynamical parameters.
526: A final judgment between them may depend on higher
527: resolution observations with {\sl Chandra} and {\sl XMM-Newton} which
528: may, for example, yield fine radial brightness and temperature profile or
529: reveal smaller scale features that might support
530: the clumpy ISM scenario.
531:
532: \section{Conclusion}
533: We have investigated the spatial and spectral properties of the SNR 3C~391
534: using \ASCA\ SIS data.
535: The CEI and NEI spectral fits indicate that the hot gas within the
536: SNR has basically reached ionization equilibrium.
537: The hydrogen column density in the direction of the NW portion of
538: the SNR is higher than that of the SE portion,
539: in agreement with the location of the molecular cloud to the NW.
540: The comparison of the hydrogen column
541: and \xray\ hardness between the NW and SE portions
542: supports the NW-to-SE breakout scenario, as suggested by earlier
543: observations of this SNR.
544: The much lower mean density within the SNR than the immediate
545: ambient cloud density and the centrally brightened \xray\ morphology
546: can be explained either by an SNR evolving in a clumpy cloud
547: inside of which gas is evaporated from the engulfed cloudlets,
548: or by an SNR which has entered the radiative stage with the interior
549: gas still hot, but with the rim material cooled down.
550:
551: \acknowledgements{
552: The authors would like to
553: thank Steve Reynolds and David Moffett for providing the radio
554: image of 3C~391.
555: We also thank Randall Smith for helpful discussions
556: related to this paper.
557: A special gratitude should be ascribed to an anonymous referee
558: whose comments help to improve the manuscript appreciably.
559: Part of YC's work was carried out in the CfA.
560: YC acknowledges support from CNSF grant 1007003
561: and grant NKBRSF-G19990754 of China Ministry of Science and Technology.
562: POS acknowledges support from NASA contract NAS8-39073 and grant
563: NAG5-9281.
564: }
565: \clearpage
566: \begin{thebibliography}{}
567: \bibitem{} Bamba, A., Yokogawa, J., Sakano, M., \& Koyama, K. 2000,
568: PASJ, 52, 259
569: \bibitem{} Brown, R.L., \& Gould, R.J. 1970, Phys.\ Rev.\ D, 1, 2252
570: \bibitem{} Caswell, J.L., Dulk, G.A., Goss, W.M., Radhakrishnan, V.,\&
571: Green, A.J. 1971, A\&A, 12, 271
572: \bibitem{} Cioffi, D.F., McKee, C.F., \& Bertschinger, E. 1988, ApJ,
573: 334, 252
574: \bibitem{} Cox, D. P., Shelton, R. L., Maciejewski, W., Smith, R. K.,
575: Plewa, T., Pawl, A., \& R\'{o}zyczka, M. 1999, ApJ, 524, 179
576: \bibitem{} Frail, D.A., Goss, W.M., Reynoso, E.M., Giacani, E.B.,
577: Green, A.J., \& Otrupcek, R. 1996, AJ, 111, 1651
578: \bibitem{} Gotthelf, E. V., Ueda, Y., Fujimoto, R., Kii, T., \& Yamaoka, K.
579: 2000, ApJ, 543, 417
580: \bibitem{} Green, A. J., Frail, D. A., Goss, W. M., Otrupcek, R.
581: 1997, AJ, 114, 2058
582: \bibitem{} Harrus, I. M., Hughes, J. P, Singh, K. P., Koyama, K.,
583: \& Asaoka, I. 1997, ApJ, 488, 781
584: \bibitem{} Harrus, I. M., Slane, P. O., Smith, R. K., \& Hughes, J. P.
585: 2001, ApJ, in press
586: \bibitem{} Mewe, R., Kaastra, J.S., \& Liedahl, D.A. 1995, Legacy, 6, 16
587: \bibitem{} Moffett, D. A., \& Reynolds, S. P., 1994, ApJ, 425, 668
588: \bibitem{} Morrison, R., \& McCammon, D., 1983, ApJ, 270, 119
589: \bibitem{} Olbert, C.M., Clearfield, C.R., Williams, N.E., Keohane, J.W.,
590: \& Frail, D.A. 2001, ApJ Letter, in press (astro-ph/0103268)
591: \bibitem{} Predehl, P., \& Schmitt, J.H.M.M. 1995, A\&A, 293, 889
592: \bibitem{} Radhakrishnan, V., Goss, W.M., Murray, J.D., \& Brooks, J.W.
593: 1972, ApJS, 24, 49
594: \bibitem{} Reach, W.T., \& Rho, J.H., 1996 A\&A, 315, L277
595: \bibitem{} ------------------, 1998, ApJ, 507, L93
596: \bibitem{} ------------------, 1999, ApJ, 511, 836
597: \bibitem{} Reach, W.T., Rho, J.H., \& Jarrett, T.H., 2001, astro-ph/0108173
598: \bibitem{} Reynolds, S.P., \& Moffett, D.A. 1993, AJ, 105, 2226 (RM93)
599: \bibitem{} Rho, J.H., \& Petre, R. 1996, ApJ, 467, 698 (RP96)
600: \bibitem{} Rho, J.H., \& Petre, R. 1998, ApJ, 503, L167
601: \bibitem{} Shelton, R. L., Cox, D. P., Maciejewski, W., Smith, R. K.,
602: Plewa, T.; Pawl, A., \& R\'{o}zyczka, M. 1999, ApJ, 524, 192
603: \bibitem{} Spitzer, L. Jr. 1978, Physical Processes in the Interstellar
604: Medium (Wiley, New York)
605: \bibitem{} Wilner, D.J., Reynolds, S.P., \& Moffett, D.A. 1998, ApJ, 115, 247
606: \bibitem{} White, R.L., \& Long, K.S. 1991, ApJ, 373, 543 (WL91)
607: \end{thebibliography}
608: \clearpage
609:
610: \centerline{\begin{tabular}{rccr}
611: \multicolumn{4}{c}{Table 1a: Summary of observations (used)} \\ \hline\hline
612: Seq.\ No. & Instrument & clocking mode & exposure \\ \hline
613: ad51017000s000102h & SIS0 & 4-CCD & 36 s \\
614: s000202h & SIS0 & 4-CCD & 10,479 s \\
615: s000302m & SIS0 & 1-CCD & 23,927 s \\
616: s000502m & SIS0 & 4-CCD & 26 s \\
617: s000602m & SIS0 & 1-CCD & 94 s \\
618: ad51017000s100102h & SIS1 & 4-CCD & 10,594 s \\
619: s100202m & SIS1 & 1-CCD & 24,056 s \\
620: s100402m & SIS1 & 4-CCD & 26 s \\ \hline
621: \end{tabular}}
622: \bigskip
623:
624: \centerline{\begin{tabular}{cccc}
625: \multicolumn{4}{c}{Table 1b: Summary of net count rates (in ${\rm counts}\ps$)
626: } \\ \hline\hline
627: CCD chips & whole & NW & SE \\ \hline
628: SIS0 1-CCD & $(18.7\pm0.3)\E{-2}$ & $(5.3\pm0.2)\E{-2}$ & $(7.2\pm0.2)\E{-2}$ \\
629: SIS1 1-CCD & $(15.2\pm0.3)\E{-2}$ & $(4.9\pm0.1)\E{-2}$ & $(5.8\pm0.2)\E{-2}$ \\
630: SIS0 4-CCD & $(17.3\pm0.4)\E{-2}$ & $(5.4\pm0.2)\E{-2}$ & $(6.3\pm0.3)\E{-2}$ \\
631: SIS1 4-CCD & $(13.0\pm0.4)\E{-2}$ & $(4.3\pm0.2)\E{-2}$ & $(4.5\pm0.2)\E{-2}$ \\ \hline
632: \end{tabular}}
633: \clearpage
634:
635: \centerline{\begin{tabular}{c|ccc}
636: \multicolumn{4}{c}{Table 2: VMEKAL (CEI) fitting results with the 90\%
637: confidence ranges}\\ \hline\hline
638: & whole & NW & SE \\ \hline
639: %\multicolumn{1}{c}{\bf With Fe abundance frozen (case CEI1)}\\
640: $f\nel\nH V/\du^{2}$ ($10^{58}\cm^{-3}$) & $10.39^{+1.95}_{-1.75}$ &
641: $4.16^{+1.46}_{-1.20}$ & $3.71^{+1.11}_{-0.93}$\\
642: $kT_{x}$ (keV) & $0.52\pm0.03$ & $0.52^{+0.05}_{-0.04}$ & $0.51\pm0.03$\\
643: {[Mg/H]} & $1.36^{+0.18}_{-0.17}$ & $1.16^{+0.35}_{-0.30}$ &
644: $1.50^{+0.31}_{-0.27}$ \\
645: {[Si/H]} & $1.00\pm0.10$ & $0.81^{+0.17}_{-0.14}$ &
646: $1.07^{+0.17}_{-0.19}$ \\
647: {[S/H]} & $0.67\pm0.14$ & $0.53^{+0.23}_{-0.21}$ &
648: $0.63^{+0.26}_{-0.22}$ \\
649: $\NH$ ($10^{22}\cm^{-2}$) & $3.03^{+0.10}_{-0.11}$ &
650: $3.34^{+0.17}_{-0.18}$ & $2.96\pm0.14$ \\
651: $F(0.5$-$10\keV)$ ($\ergs\cm^{-2}\ps$) & $4.7\E{-12}$ & $1.5\E{-12}$ &
652: $1.7\E{-12}$\\
653: $F^{(0)}(0.5$-$10\keV)$ ($\ergs\cm^{-2}\ps$) & $3.5\E{-10}$ &
654: $1.4\E{-10}$ & $1.2\E{-10}$\\
655: $\chi^{2}/{\rm d.o.f.}$ & $349/230$ & $125/108$ & $201/118$ \\ \hline
656: \multicolumn{4}{l}{Here $f$ denotes the filling factor of the hot gas,
657: $F$ the absorbed flux,}\\
658: \multicolumn{4}{l}{and $F^{(0)}$ the unabsorbed flux.} \\
659: \end{tabular}}
660: \bigskip
661:
662: \centerline{\begin{tabular}{c|ccc}
663: \multicolumn{4}{c}{Table 3: VNEI fitting results with the 90\%
664: confidence ranges}\\ \hline\hline
665: & whole & NW & SE \\ \hline
666: %\multicolumn{1}{c}{\bf Using XSPEC (NEI1)}\\
667: $f\nel\nH V/\du^{2}$ ($10^{58}\cm^{-3}$) & $7.38^{+1.46}_{-1.09}$ &
668: $3.50^{+1.08}_{-0.88}$ & $1.99^{+0.90}_{-0.44}$\\
669: $kT_{x}$ (keV) & $0.59^{+0.03}_{-0.01}$ & $0.57^{+0.04}_{-0.05}$ &
670: $0.63^{+0.07}_{-0.03}$\\
671: $n_{e}t$ ($10^{13}\cm^{-3}\s$) & $0.85^{+\infty}_{-0.10}$ &
672: $3.86^{+\infty}_{-0.52}$ & $3.76^{+\infty}_{-0.41}$ \\
673: {[Mg/H]} & $1.37^{+0.23}_{-0.19}$ & $1.25^{+0.46}_{-0.33}$ &
674: $1.64^{+0.21}_{-0.40}$ \\
675: {[Si/H]} & $1.22^{+0.15}_{-0.12}$ & $1.02^{+0.23}_{-0.19}$ &
676: $1.38^{+0.27}_{-0.24}$ \\
677: {[S/H]} & $0.75^{+0.17}_{-0.16}$ & $0.59^{+0.29}_{-0.26}$ &
678: $0.68^{+0.32}_{-0.29}$ \\
679: $\NH$ ($10^{22}\cm^{-2}$) & $2.90\pm0.07$ &
680: $3.33\pm0.15$ & $2.70^{+0.16}_{-0.06}$ \\
681: $F(0.5$-$10\keV)$ ($\ergs\cm^{-2}\ps$) & $4.8\E{-12}$ & $1.5\E{-12}$ &
682: $1.8\E{-12}$ \\
683: $F^{(0)}(0.5$-$10\keV)$ ($\ergs\cm^{-2}\ps$) & $2.3\E{-10}$ &
684: $1.1\E{-10}$ & $6.4\E{-11}$ \\
685: $\chi^{2}/{\rm d.o.f.}$ & $359/229$ & $133/107$ & $201/117$ \\ \hline
686: \multicolumn{4}{l}{Here $f$ denotes the filling factor of the hot gas,
687: $F$ the absorbed flux,}\\
688: \multicolumn{4}{l}{and $F^{(0)}$ the unabsorbed flux.} \\
689: \end{tabular}}
690: \bigskip
691: \clearpage
692:
693: \centerline{\begin{tabular}{c|ccc}
694: \multicolumn{3}{c}{Table~4: Results of simultaneous VMEKAL fitting}\\
695: \multicolumn{3}{c}
696: {for NW and SE, assuming the same $kT_{x}$}\\
697: \multicolumn{3}{c}{and the same $\NH$, respectively.}\\ \hline\hline
698: & NW & SE \\ \hline
699: %\multicolumn{1}{c}{\bf}\\
700: $kT_{x}$ (keV) & \multicolumn{2}{c}{$0.51^{+0.04}_{-0.02}$}\\
701: $\NH$ ($10^{22}\cm^{-2}$) & $3.36\pm0.12$ & $2.95^{+0.10}_{-0.11}$\\
702: $f\nel\nH V/\du^{2}$ ($10^{58}\cm^{-3}$)
703: & $4.35^{+0.84}_{-0.75}$ & $3.60^{+0.67}_{-0.61}$\\
704: $\chi^{2}/{\rm d.o.f.}$ & \multicolumn{2}{c}{$326/233$} \\ \hline
705: $\NH$ ($10^{22}\cm^{-2}$) & \multicolumn{2}{c}{$3.13^{+0.08}_{-0.09}$}\\
706: $kT_{x}$ (keV) & $0.57\pm0.03$ & $0.48\pm0.02$\\
707: $f\nel\nH V/\du^{2}$ ($10^{58}\cm^{-3}$)
708: & $3.00^{+0.49}_{-0.37}$ & $4.82^{+0.73}_{-0.68}$\\
709: $\chi^{2}/{\rm d.o.f.}$ & \multicolumn{2}{c}{$337/233$} \\ \hline
710: \end{tabular}}
711: \vspace{15mm}
712:
713: \centerline{\begin{tabular}{c|c}
714: \multicolumn{2}{c}{Table 5: Dynamical parameters for the}\\
715: \multicolumn{2}{c}{evaporation model of WL91}\\ \hline\hline
716: % & $r_{s}\sim2.5'$ & $r_{s}\sim4'$ \\ \hline
717: $C/\tau^{\dag}$ & 3-5\\
718: $K^{\ddag}$ & 0.385-0.189\\
719: $q^{\ddag}$ & 0.372-0.143\\
720: $k\Ts$ (keV) & 0.42-0.54\\
721: $\vs$ ($\km\ps$) & 590-670 \\
722: $r_{s}$ (pc) & $7.0\ru\du$ \\
723: $t$ ($10^{3}\yr$) & (4.6-4.0)$\ru\du$ \\
724: $E$ ($10^{50}\ergs$) & (1.3-3.4)$f^{-1/2}\ru^{3}\du^{5/2}$ \\ \hline
725: \multicolumn{2}{l}{$\dag$ following RP96} \\
726: \multicolumn{2}{l}{$\ddag$ adopted from WL91} \\
727: \end{tabular}}
728: \clearpage
729:
730: \begin{center}
731: \section{Figure captions}
732: \end{center}
733:
734: \figcaption[f1.eps]{
735: The SIS \xray\ spectra of 3C391 fitted with the VMEKAL model.
736: The three energy ranges (ER) are labelled, in which the narrow band
737: images are produced.
738: }
739:
740: \figcaption[f2.ps]{
741: Two dimensional confidence contours for the NW and SE regions of 3C~391.
742: The contours from inner to outer correspond to $\Delta\chi^{2}=$2.30 (68\%),
743: 4.61 (90\%), and 9.21 (99\%).
744: }
745:
746: \figcaption[f3.ps]{
747: The \xray\ images produced using SIS0 data.
748: Panel {\em a} is the hard (2.6-10$\keV$) emission gray-scale image
749: overlaid with the soft (0.5-2.6$\keV$) emission contours.
750: The three circles in panel {\em a} designate the areas
751: from which the SIS spectra are extracted.
752: Panels {\em b} and {\em c} display the gray-scale images of
753: soft and hard emission, respectively, overlaid with the dashed contours
754: of 1.5 GHz radio emission (from Moffett \& Reynolds [1994]).
755: Panels {\em d}, {\em e}, and {\em f} are the narrow band
756: gray-scale images and solid contours of
757: Mg He$\alpha$ (1.2-1.5$\keV$),
758: Si He$\alpha$ (1.7-2.0$\keV$),
759: and S He$\alpha$ (2.3-2.6$\keV$) emissions.
760: In panel {\em f} the image is also overlaid with the dashed radio contours.
761: The seven levels of solid contours are linear between the maximum
762: and 20\% maximum brightness.
763: The two cross labels in each panel denote the OH maser points
764: (Frail et al.\ 1996).
765: }
766: \end{document}
767: