astro-ph0109556/pp.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: \slugcomment{Accepted for publication in the Astrophysical Journal}
4: 
5: \shorttitle{Cygnus Loop Nonradiative Shock - \textit{FUSE}}
6: \shortauthors{Sankrit and Blair}
7: 
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: 
10: \begin{document}
11: 
12: \title{\textit{FUSE} Observations of the Cygnus Loop: O~VI Emission
13: from a Nonradiative Shock \altaffilmark{1}}
14: 
15: \author{Ravi Sankrit \& William P. Blair}
16: \affil{The Johns Hopkins University}
17: \affil{Department of Physics and Astronomy,
18: 3400 N. Charles Street,  Baltimore, MD 21218}
19: \email{ravi@pha.jhu.edu}
20: 
21: \altaffiltext{1}{Based on data obtained for the Guaranteed Time Team by the
22: NASA-CNES \textit{FUSE} mission operated by the Johns Hopkins University.  Financial
23: support to U.S. participants has been provided by NASA contract NAS5-32985.}
24: 
25: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
26: 
27: \begin{abstract}
28: 
29: We present \textit{Far Ultraviolet Spectroscopic Explorer (FUSE)}
30: observations of a Balmer filament in the northeast region of the Cygnus
31: Loop supernova remnant.  The data consist of one spectrum obtained
32: through the $30\arcsec\times30\arcsec$ (LWRS) aperture and three
33: spectra at adjacent positions obtained through the
34: $4\arcsec\times20\arcsec$ (MDRS) aperture.  The nonradiative shocks in
35: the region giving rise to these faint optical filaments produce strong
36: O~VI~$\lambda\lambda$1032,1038 emission, which is detected in all the
37: spectra.  The O~VI emission is resolved by \textit{FUSE} into a strong
38: component centered at 0~km~s$^{-1}$, and weaker components centered at
39: $\pm~140$~km~s$^{-1}$.  The MDRS spectra allow us to study the
40: variation of O~VI emission in the post-shock structure.  We find that
41: the zero velocity emission is associated directly with the Balmer
42: filament shock, while the high velocity emission comes from a more
43: uniformly distributed component elsewhere along the line of sight.  We
44: also find that the shocks producing the emission at
45: $\pm~140$~km~s$^{-1}$ have velocities between 180~km~s$^{-1}$ and
46: 220~km~s$^{-1}$, if we assume that the ram pressure driving them is the
47: same as for the zero velocity component shock.  In the context of the
48: cavity model for the Cygnus Loop, the interaction of the blast wave
49: with the spherical shell that forms most of the cavity wall can
50: naturally give rise to the similar red and blue-shifted components that
51: are observed.
52: 
53: \end{abstract}
54: 
55: \keywords{ISM: individual (Cygnus Loop) --- ISM: supernova remnants --- 
56: shock waves}
57: 
58: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
59: \section{Introduction}
60: 
61: Among the most prominent lines in the ultraviolet spectra of several
62: supernova remnant (SNR) shocks are O~VI~$\lambda\lambda$1032,1038, a
63: doublet with the ground state as their common lower level.  O~VI is
64: produced in shocks with velocities of at least 160~km~s$^{-1}$.  For
65: shocks with velocities up to about 400~km~s$^{-1}$, the strength of the
66: O~VI lines is very sensitive to the shock velocity.  In addition, the
67: O~VI line strength also depends on the column density swept up by the
68: shock.  The regions producing O~VI emission are typically at
69: temperatures near 300,000~K\@.  Therefore, these lines are the best
70: tracers of SNR gas that is cooler than the x-ray emitting regions, and
71: hotter than the regions of bright optical emission.
72: 
73: The optically faint filaments in the northeast region of the Cygnus
74: Loop SNR produce a large amount of O~VI emission.  In this region, the
75: optical emission is dominated by Balmer lines, which are collisionally
76: excited in a narrow zone behind the shock front \citep{che80}.  The
77: ultraviolet lines, including O~VI~$\lambda\lambda$1032,1038, come from
78: the hot post-shock gas further downstream where elements are moving to
79: higher stages of ionization.  The shocked gas has not had time to
80: recombine and cool radiatively and so these shocks are termed
81: ``nonradiative''.
82: 
83: Strong O~VI emission was detected in a \textit{Hopkins Ultraviolet
84: Telescope (HUT)} spectrum of a bright nonradiative filament in the
85: northeast Cygnus Loop by
86: \citet{lon92}.  They found that the ratio of line strengths
87: I$_{1038}$/I$_{1032}$ implied significant resonant scattering along the
88: line of sight, and inferred a pre-shock density between 5 and
89: 12~cm$^{-3}$.  However, they also found that the overall \textit{HUT}
90: spectrum was best fit by a 180~~km~s$^{-1}$ shock running into material
91: with a density of 2~cm$^{-3}$.  Other studies of the same filament
92: based on spectra of the Balmer lines and on \textit{International
93: Ultraviolet Explorer (IUE)} data also suggest that the preshock density
94: is $\sim 2$~cm$^{-3}$ \citep{ray83, hes94}.  More recently, analysis of
95: the distribution of C~IV~$\lambda$1549 and N~V~$\lambda$1240 emission
96: in a region behind the shock front (seen in a spatially resolved
97: \textit{Space Telescope Imaging Spectrograph} spectrum) showed that the
98: pre-shock density is at least 2~cm$^{-3}$ (but probably closer to
99: 4~cm$^{-3}$), and that the shock velocity is $\sim 180$~km~s$^{-1}$
100: \citep{san00}.
101: 
102: Here, we present \textit{Far Ultraviolet Spectroscopic Explorer (FUSE)}
103: observations of a portion of this well studied  Balmer filament.  The
104: only SNR lines detected in the \textit{FUSE} bandpass are
105: O~VI~$\lambda\lambda$1032,1038.  However, in contrast to the
106: \textit{HUT} spectrum, the two lines of the doublet are clearly
107: separated from each other and from the Ly$\beta$ airglow line, and they
108: are resolved into complex line profiles that vary with spatial position
109: behind the shock.
110: 
111: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
112: \section{Observations}
113: 
114: \textit{FUSE} observations were obtained on 2000 June 6, as part of the
115: Guaranteed Time Team project on SNRs.  The first observation was
116: obtained through the low resolution (LWRS) $30\arcsec~\times~30\arcsec$
117: aperture on the filament (centered at $\alpha_{2000}=20^{\rm{h}}
118: 56^{\rm{m}} 06\fs57$,
119: $\delta_{2000}=+31\arcdeg\ 56\arcmin\ 05\farcs6$).  Three other
120: observations were obtained through the medium resolution (MDRS)
121: 20\arcsec~$\times$~4\arcsec\ aperture with the long dimension parallel
122: to the filament.  The first MDRS position was centered on the same
123: location as the LWRS observation, and the next two were obtained with
124: the slit stepped back each time by $\sim$3\arcsec\ perpendicular to its
125: length.  (Henceforth, we will refer to these positions as P1, P2 and
126: P3.) Details of the observations are presented in Table \ref{tblobs}.
127: In Figure~\ref{SLITS}, the aperture locations are shown overlaid on a
128: WFPC2 H$\alpha$ image of the filament \citep{bla99}.
129: 
130: \textit{FUSE} has four independent channels, LiF1, LiF2, SiC1, and
131: SiC2, each with two segments, ``A'' and ``B'' (see \cite{moo00} for
132: details).  Four segments, LiF1A, LiF2B, SiC1A, and SiC2B cover the
133: wavelength region $\sim$1000 -- 1100\AA, which includes the
134: O~VI~$\lambda\lambda$1032,1038 doublet.  Of the four segments, LiF1A is
135: the one with the largest effective area at 1035\AA, and the O~VI lines
136: are well detected.  Furthermore, since the LiF1A channel is used for
137: guiding, only the LiF1 apertures can be held accurately in position for
138: the duration of an observation -- other channels drift with respect to
139: LiF1 due to thermal motions \citep{sah00}.  Therefore, in this paper we
140: present only data from the LiF1A segment.
141: 
142: For each observation, we combined the raw data from all individual
143: exposures and then used CALFUSE pipeline version 1.8.7 to produce
144: calibrated spectra.  A shift was applied to the flux vectors for each
145: observed LiF1A spectrum in order to line up geocoronal Ly$\beta$
146: emission at the appropriate wavelength in the heliocentric frame of
147: reference.  (The average geocentric to heliocentric velocity is
148: calculated and included in the headers of the calibrated FITS files.)
149: The pipeline slightly oversubtracts the background, so we added a
150: correction to the flux vectors -- this step was done to avoid negative
151: fluxes in the plots; the measured line strengths are not affected.
152: 
153: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
154: \section{Results}
155: 
156: The LWRS spectrum between 1024\AA\ and 1040\AA\ is shown in Figure
157: \ref{LWRS}.  Each O~VI line has a strong central component, and emission
158: on both red and blue wings.  The wings are prominent in the
159: 1032\AA\ line but much weaker in the 1038\AA\ line.  The spectral
160: resolution for emission lines from extended sources is the filled slit
161: width $\sim0.34$\AA, which corresponds to $\sim100$~km~s$^{-1}$ at
162: 1034\AA\@.  (This is the width of the airglow lines in the spectrum.)
163: The width of the central component of the O~VI emission is about the
164: filled slit width.  The integrated line fluxes of the 1032\AA\ and
165: 1038\AA\ lines (including the central component and the wings) are
166: $7.67\times10^{-13}$~erg~s$^{-1}$~cm$^{-2}$ and
167: $3.66\times10^{-14}$~erg~s$^{-1}$~cm$^{-2}$, respectively.
168: 
169: The MDRS spectra between 1024\AA\ and 1040\AA\ are shown in Figure
170: \ref{MDRS}.  The filled slit spectral resolution with this aperture is
171: $\sim0.045$\AA, which corresponds to $\sim13$~km~s$^{-1}$ at
172: 1034\AA\@.  The O~VI lines are broader than the filled slit width
173: (compare the O~VI lines with the airglow lines).  The two components on
174: the wings of the 1032\AA\ line are clearly seen in each of the
175: spectra.  The O~VI line strengths decrease from P1 to P3.  Also, in the
176: P3 spectrum, the lines are broader and the central component of
177: O~VI~$\lambda$1032 is double peaked.  The wing components are about
178: equally strong at all three positions.  The O~VI line fluxes in the
179: MDRS spectra, along with the fluxes in the LWRS spectrum are given in
180: Table \ref{tblflux}.  The fluxes were obtained by simple trapezoidal
181: integration over the line profiles.  The error in each measurement is
182: dominated by the absolute flux calibration, which is accurate to 10\%
183: \citep{sah00}.
184: 
185: The O~VI line fluxes at the three MDRS positions are plotted against
186: velocity in Figure \ref{VEL}.  The top panel shows O~VI~$\lambda$1032
187: and the bottom panel shows O~VI~$\lambda$1038.  The central components
188: of the 1032\AA\ line and the 1038\AA\ line are centered at
189: 0~km~s$^{-1}$, and their FWHMs are $\sim100$~km~s$^{-1}$.  The
190: components on the 1032\AA\ line wings are centered at approximately
191: $\pm140$~km~s$^{-1}$.
192: 
193: The intrinsic line profiles of O~VI~$\lambda$1032 and
194: O~VI~$\lambda$1038 are expected to be identical, except that the
195: 1032\AA\ line is twice as strong as the 1038\AA\ line in the optically
196: thin limit.  The features on the wings of the 1032\AA\ line in the \textit{FUSE}
197: spectra are not likely to be spurious since we see them in the LiF2B
198: spectra as well.  (The LiF2B spectra are not shown here.)  Their absence
199: in the 1038\AA\ emission can be attributed to absorption by molecular
200: hydrogen along the line of sight between us and the Cygnus Loop.  The
201: Lyman band transitions of H$_{2}$, R(1)$_{5-0}$ and P(1)$_{5-0}$, have
202: rest wavelengths of 1037.146\AA\ and 1038.156\AA, respectively
203: \citep{mor76}.  Relative to the central O~VI wavelength (1037.617\AA)
204: these are at $-136$~km~s$^{-1}$ and $+155$~km~s$^{-1}$, respectively.
205: If H$_2$ is present along the line of sight, then we expect these
206: strong transitions to affect the spectrum strongly around the O~VI
207: 1038\AA\ line.
208: 
209: To check this quantitatively, we used a \textit{FUSE} spectral simulation
210: routine, FSIM, to examine the effects of H$_{2}$ absorption.  O~VI
211: emission components at zero velocity and at $\pm140$~km~s$^{-1}$ were
212: included.  The peak fluxes on the 1032\AA\ wings were chosen to be
213: $4\times10^{-14}$~erg~s$^{-1}$~cm$^{-2}$~\AA$^{-1}$; the peak fluxes of
214: the 1038\AA\ line were chosen to be half this value (valid for
215: optically thin line emission).  We find that $10^{16}$~cm$^{-2}$ of
216: H$_{2}$ at a temperature of 250~K would be sufficient to absorb the
217: O~VI~$\lambda$1038 wings to the point that the emission would be
218: undetectable.  Although we do not attempt to set any limits, we
219: conclude that there is sufficient molecular hydrogen towards the Cygnus
220: Loop that results in the O~VI~$\lambda$1038 wings being absorbed.
221: (O~VI line profiles showing the presence of overlying H$_2$ absorption
222: are also seen in \textit{FUSE} spectra of other regions in the Cygnus Loop
223: (Blair et al.~in preparation).)
224: 
225: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
226: \section{Discussion}
227: 
228: The set of MDRS spectra were obtained in order to map the spatial
229: distribution of the O~VI emission perpendicular to the shock front.  We
230: find that the central component of O~VI emission is well correlated
231: with the H$\alpha$ emission from the filament.  The O~VI flux is
232: highest at P1, where the filament is most edge-on and the H$\alpha$ is
233: brightest (Figure \ref{SLITS}), and is lower at positions P2 and P3,
234: which lie behind the leading edge of the filament.  Furthermore, the
235: O~VI and H$\alpha$ emission are both centered at 0~km~s$^{-1}$ (Figure
236: \ref{VEL}, and \citet{hes94}).  The high velocity components, on the
237: other hand, do not vary much among the three MDRS positions (Table
238: \ref{tblflux}) and we infer that they come from regions that are more
239: uniformly distributed (in the plane of the sky) than the gas behind the
240: optical filament.
241: 
242: The O~VI surface brightnesses measured in the \textit{FUSE} spectra and in the
243: HUT spectrum are listed in Table \ref{tblsb}.  The HUT observation was
244: obtained with the $9\farcs4\times116\arcsec$ placed with its length
245: along the Balmer filament (see Figure 1 of \citet{lon92}).  At the
246: distance of the Cygnus Loop -- 440~pc \citep{bla99}, the aperture
247: length corresponds to $\sim0.25$~pc.  The O~VI surface brightness of
248: the region observed by HUT is as high as it is at P1.  Therefore, we
249: infer that strong O~VI emission is more or less uniformly distributed
250: along the length of the filament.  Spatial variations in the O~VI flux
251: are much larger in the direction perpendicular to the shock front.
252: 
253: Approximately 1/3 the area of the LWRS aperture covers a region that is
254: ahead of the H$\alpha$ filament (Figure \ref{SLITS}).  Therefore it is
255: not surprising that the O~VI surface brightness in the LWRS spectrum is
256: even lower than the surface brightness in the MDRS P3 spectrum.  The
257: average surface brightnesses of the three MDRS apertures are
258: $13.5\times10^{-16}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$ for
259: O~VI~$\lambda$1032 and
260: $6.9\times10^{-16}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$ for
261: O~VI~$\lambda$1038.  If we assume that 2/3 of the LWRS aperture is
262: filled with O~VI, we can calculate corrected LWRS surface brightnesses
263: by multiplying the values in Table \ref{tblsb} by 3/2.  The corrected
264: surface brightnesses are
265: $12.8\times10^{-16}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$ and
266: $6.2\times10^{-16}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$ for the
267: 1032\AA\ and 1038\AA\ lines, respectively.  These numbers are much
268: closer to the average MDRS brightnesses.  This analysis suggests
269: that there is very little O~VI emitted in the region just ahead
270: of the filament.
271: 
272: By comparing various spectra taken by \textit{FUSE} and HUT, we have inferred
273: some of the properties of the distribution of O~VI emission in the
274: vicinity of the Balmer line filament.  Now, we consider the properties
275: of the emission components in the MDRS spectra and what can be learned
276: from them about the shock and cloud properties.
277: 
278: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
279: 
280: \subsection{The Zero Velocity Component}
281: 
282: The distribution of O~VI emission seen in the three MDRS spectra is
283: affected by the history and the geometry of the shock
284: front.  According to our current understanding of the Cygnus Loop,
285: based on the the cavity model \citep{hes94,lev98}, the shock now giving
286: rise to the Balmer filaments encountered the cavity ``wall'' several
287: hundred years ago.  Although the radial density profile of this
288: ``wall'' is not known, it is likely that the pre-shock density was
289: lower and the shock velocity higher at earlier times.  The contribution
290: to the O~VI flux from these earlier times is uncertain, but probably
291: not very high: a fast shock will ionize oxygen past O$^{5+}$ rapidly,
292: and the lower pre-shock density implies a lower O~VI yield.  Hence it is
293: likely that the differences in the zero velocity component of O~VI
294: emission among the three MDRS spectra are due mainly to the shock
295: geometry.  In this paper, we will not consider the shock history
296: further.
297: 
298: The shock front is a rippled sheet with a peak to peak amplitude of
299: $\sim$10\arcsec, measured in the WFPC2 H$\alpha$ image of the filament
300: \citep{bla99}.  On larger scales, the shock front is curved with the
301: convex face outward \citep{lev98}.  It may be assumed that the filament
302: similarly curves along the line of sight.  The O~VI flux is highest at
303: P1 because it is closest to the edge -- i.e. it is
304: ``limb-brightened''.  The flux at position P3 is lower because the path
305: length through the emitting gas is shorter than at P1.  Because the
306: path length is shorter, the effects of resonance scattering are smaller
307: and the I$_{\lambda1038}$/I$_{\lambda1032}$ is closer to 0.5 at P3 than
308: at P1 (Table \ref{tblflux}).  The O~VI lines are slightly broader in
309: the P3 spectrum than in the P1 and P2 spectra, and the 1032\AA\ line
310: shows a double peak.  This is due to the curvature of the filament --
311: much of the emission at P3 comes from portions of the shock front that
312: have a small radial velocity component.  
313: 
314: We have accounted for the spatial variation of the O~VI emission seen
315: in the MDRS spectra in a qualitative way using geometric arguments.
316: However, the spatial resolution and coverage of the observations are
317: insufficient to distinguish between different values of the shock
318: velocity and pre-shock density.  Therefore, in the following
319: discussion, we use the average O~VI surface brightness of the three
320: MDRS positions.
321: 
322: The total flux of the O~VI$\lambda$1032 zero velocity component in the
323: MDRS spectra (P1,P2 and P3) is
324: $2.72\times10^{-13}$~erg~s$^{-1}$~cm$^{-2}$ (Table \ref{tblflux}),
325: which corresponds to a surface brightness of
326: $1.13\times10^{-15}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$.  In order to
327: compare the observed flux with shock model predictions, two corrections
328: have to be made.  First, the interstellar extinction has to be
329: accounted for.  We use the extinction curve presented by \citet{fit99}
330: and find, for color excess E$_{B-V}$=0.08 to the Cygnus Loop
331: \citep{fes82} and $R_V = 3.1$ (the standard ISM value), that the
332: correction factor is 2.8 at 1032\AA\@.  Second, since we are viewing
333: the shock close to edge-on, the resonance scattering of the O~VI lines
334: can significantly attenuate their intensities.  If resonant scattering
335: effects were very small (i.e.~the optical depth in the lines was
336: negligible) then the ratio I$_{\lambda1038}$/I$_{\lambda1032}$ would be
337: 0.5.  The observed ratio is about 0.60, which implies that the optical
338: depth in the 1032\AA\ line is $\sim0.8$.  Following the procedure
339: outlined in \citet{lon92}, we find that the attenuation factor for the
340: line is $\sim1.5$.  The corrected surface brightness of the central
341: component of O~VI$\lambda$1032 (obtained by multiplying the measured
342: surface brightness by these correction factors derived above) is
343: $4.7\times10^{-15}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$.
344: 
345: We choose parameters based on earlier studies of the filament
346: \citep{ray83,lon92,hes94,san00} and use these in shock models to see
347: how the predictions compare with the observation.  Three models were
348: calculated, with shock velocities of 170, 180 and 190~km~s$^{-1}$.
349: Diffuse ISM abundances, where O:H = 8.70:12.00 on a logarithmic scale
350: \citep{cow86}, and a pre-shock hydrogen number density of 3~cm$^{-3}$
351: were used in all the models.  The models were calculated using an
352: updated version of the code described by \citet{ray79}.  The models
353: predict the face-on O~VI intensity as a function of the swept up column
354: density.  
355: 
356: In order to compare the predicted intensity with the observed
357: intensity it is necessary to take into account the viewing aspect
358: ratio, i.e.~the ratio of actual shock surface area lying within the
359: slit to the projected area.  The observed region is
360: 10\arcsec$\times$20\arcsec (the MDRS positions overlap - see Figure
361: \ref{SLITS}).  If the extent of the shock front along the line of sight
362: is $l$ arcseconds, then since the MDRS apertures are placed parallel to
363: the filament, the aspect ratio is $\sim~l/10$.  Figure 1 of
364: \citet{hes94} shows that the filament is about 300\arcsec\ long.  If we
365: assume that the extent of the filament along the line of sight is
366: comparable to its extent on the sky (there is no reason to believe
367: otherwise), then the aspect ratio is about 30.  Now, the filament we
368: are studying is the brightest Balmer filament in the northeast Cygnus
369: Loop.  This may be a selection effect based on a longer path length
370: along the line of sight so in the following discussion we also
371: consider the possibility of a higher aspect ratio as well.
372: 
373: In Figure \ref{MODS} the O~VI intensities predicted by the shock models
374: are plotted as functions of swept up hydrogen column density.  The
375: horizontal lines show the observed intensities (corrected for
376: interstellar extinction and resonance scattering as described above)
377: for aspect ratios of 30 and 60, which may be considered as approximate
378: bounds on the true value.  Note that these models are all calculated
379: for pre-shock density, n$_{0}=3$~cm$^{-3}$; for a given shock velocity
380: and swept up column density, the O~VI intensity scales linearly with
381: n$_{0}$.  Analyses of other ultraviolet spectra \citep{lon92, san00}
382: have shown that the swept up column density lies in the range
383: $\sim0.8$--$2\times10^{17}$~cm$^{-2}$.  From the plot we see that the
384: models predict the O~VI emission correctly.  The important conclusion
385: of the preceding analysis is that the observed O~VI flux can be
386: accounted for by nonradiative shock emission alone -- it is not necessary to
387: invoke other mechanisms such as cloud evaporation, or ``coronal''
388: emission from cooling gas ionized at an earlier time.
389: 
390: The 180~km~s$^{-1}$ shock model predicts a total O~VI line intensity
391: $\sim16$ times the H$\alpha$ intensity for a swept up column of
392: $1\times10^{17}$~cm$^{-2}$.  Although only the O~VI doublet lines lie
393: in the \textit{FUSE} bandpass, other strong UV lines have been detected
394: in spectra of this filament.  Predicted intensities of these UV lines
395: from similar shock models are given by \citet{lon92}.  Several lines of
396: He~I, He~II, O~IV and O~V at wavelengths below 912\AA\ are predicted to
397: be comparable to O~VI in intensity.  The strongest of these are
398: He~II~$\lambda$304 (a few times as strong as O~VI), and
399: O~V~$\lambda$630 (about 1.5 times as strong as O~VI).  These lines
400: cannot be observed because of interstellar absorption, but they can be
401: important for photoionization of the pre-shock gas \citep{ham88}.
402: 
403: A few lines in the optical range are expected to be present at the
404: level of a few percent the brightness of H$\alpha$, in particular
405: He~I~$\lambda$6678, He~II~$\lambda$4686 \citep{har99} and
406: [O~III]~$\lambda$5007.  The brightest predicted line,
407: [Ne~V]~$\lambda$3425 at about 10\% the H$\alpha$ intensity, was
408: detected by \cite{ray83}.  Infrared emission lines are predicted to be
409: very faint, the brightest being [O~IV]~25.9$\mu$m and [Ne~VI]~7.6$\mu$m
410: lines at about 1\% the intensity of H$\alpha$.
411: 
412: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
413: 
414: \subsection{The Wing Components}
415: 
416: The total flux of the high velocity O~VI at the three MDRS positions is
417: $5.3\times10^{-14}$ erg s$^{-1}$ cm$^{-2}$ (Table \ref{tblflux}).  The
418: correction for interstellar extinction is exactly the same as for the
419: main component (a factor of 2.8).  As we discuss below, resonance
420: scattering effects may be neglected.  So, the corrected surface
421: brightness for the wing components (red and blue total) is
422: $6.2\times10^{-16}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$.  It is seen
423: from Figure \ref{VEL} that flux in the blue wing is approximately the
424: same as the flux in the red wing.  To check this quantitatively, we fit
425: gaussians to the wing profiles, and also evaluated the integrals under
426: the curve, between suitably chosen points.  Both methods showed that
427: the difference in flux between the two wings is $\lesssim$10\%.
428: For the following discussion, we assume that the two components are
429: equally bright and consider only the blue wing -- the arguments used
430: and results derived apply equally well to both components.
431: 
432: We start by making the assumption that the $-140$~km~s$^{-1}$ emission
433: is due to a steady shock traveling towards us, and that we are looking
434: through just a single shock structure.  (The line optical depth will be
435: small, and attenuation of the flux due to resonance scattering can be
436: neglected.) The angle between the shock normal and the line of sight
437: ($\theta$), the observed and face-on surface brightnesses, and the
438: observed velocity and shock velocity are related by
439: \begin{equation}
440: cos\theta = \frac{v_{obs}}{v_s} = \frac{I_0}{I_{obs}}
441: \end{equation}
442: and the latter relation may be written as
443: \begin{equation}
444: I_0 v_s = I_{obs} v_{obs}.
445: \end{equation}
446: Thus, there are a family of solutions for the shock velocity and
447: face-on surface brightness that will match the observations.
448: 
449: We next assume that the pressure driving this shock is equal to the
450: pressure driving the shock responsible for the H$\alpha$ filament and
451: the zero velocity O~VI emission.  This is a reasonable assumption since
452: ROSAT x-ray maps show no evidence for large pressure variations in the
453: region \citep{lev97}.  Then the isobaric relation, $n_0
454: v_s^2~=~constant$, holds.  The constant, based on the parameters used
455: in the model described above (\S4.1), is
456: $3~\times~180^2$=97200~(cm$^{-3}$)(km~s$^{-1}$).  The isobaric relation
457: can be used to specify the pre-shock density for every assumed value of
458: shock velocity.
459: 
460: As discussed in \S4.1, for a given shock velocity and pre-shock
461: density, the O~VI intensity
462: depends on the swept up column density.  A third constraint on
463: the shock parameters is obtained from the relation
464: \begin{equation}
465: n_0 v_s t_{shock} = N_{\rm{H}}.
466: \end{equation}
467: Here $t_{shock}$ is the time since the shock interaction started (the
468: age of the shock), and $N_{\rm{H}}$ is the swept up column density.
469: 
470: We ran a sequence of models with shock velocities 160 --
471: 300~km~s$^{-1}$.  The pre-shock densities were calculated using the
472: isobaric condition.  Equation 2 was used to find the required O~VI
473: surface brightness for each shock velocity.  For each model this was
474: achieved at some value of the swept up column density.  (The exception
475: is the 160~km~s$^{-1}$ shock, which does not produce enough O~VI
476: emission to match the observations.) The age of the shock is then found
477: from equation 3.  In Table \ref{tblmod} we list the parameters used in
478: the models, the required O~VI~1032\AA\ surface brightnesses, and the
479: resulting swept up column density and shock age.  The swept up column
480: density required to produce the observed O~VI increases with shock
481: velocity.  As a consequence of the isobaric condition and equation 3,
482: $t_{shock}\propto N_{\rm{H}}v_s$, so the shock age also increases with
483: shock velocity.  There is a sudden jump in the required $N_{\rm{H}}$,
484: and hence the shock age between shock velocities 200 and
485: 220~km~s$^{-1}$.  This is because in faster shocks, the oxygen rapidly
486: ionizes to O$^{6+}$ and beyond and the required O~VI intensity is
487: produced only after the gas recombines to O$^{5+}$.
488: 
489: The age of the Cygnus Loop is $\sim9,000\times(D_{pc}/440)$~yr, where
490: $D_{pc}$ is the distance to the remnant \citep{lev98}.  Since the shock
491: age cannot exceed the age of the remnant, shock velocities
492: $\gtrsim260$~km~s$^{-1}$ are ruled out (Table \ref{tblmod}).  However,
493: more stringent limits can be placed on the shock velocity because
494: detailed studies of the northeast Cygnus Loop \citep{hes94} show that
495: the interaction started $\sim1000$~yr ago.  Therefore, the allowed
496: range of shock velocities is 180--220~km~s$^{-1}$ (Table
497: \ref{tblmod}).  The angles between the line of sight and the shock
498: normal for the lower and upper velocity limits are 39\arcdeg\ and
499: 50\arcdeg, respectively (equation 1).
500: 
501: We note that the exact upper limit on the shock velocity derived above
502: depends on the value of the constant in the isobaric relation.  It also
503: depends upon the observed O~VI surface brightness, for which the
504: error is dominated by the uncertainty in our knowledge of the
505: reddening correction at 1032\AA\@.  However, the expected range of
506: values for these parameters is sufficiently narrow that though the
507: limit may be revised upward by some tens of km~s$^{-1}$, the basic
508: picture for the production of O~VI emission would not change.
509: 
510: Our analysis has shown that either high velocity component of O~VI can
511: be produced by a shock with properties similar to the one producing the
512: Balmer line emission.  The similarity between the blue and red wing
513: components (equal flux, symmetric about zero in velocity space) can be
514: understood in the context of the cavity model for the Cygnus Loop.  In
515: this model \citep{hes94,lev98}, the blast wave is interacting with the
516: walls of the cavity cleared by the progenitor star.  According to the
517: picture presented by \citet{lev98}, the cavity wall consists of a
518: neutral atomic shell covering most of the surface, and several large
519: cloud boundaries filling in the rest.  The shell was formed by
520: recombination at the edges of the pre-supernova H~II region.  In such a
521: model we expect to see symmetrical red- and blue-shifted shocks along
522: lines of sight that do not intersect the large clouds.  Thus, in the
523: high velocity wings observed with \textit{FUSE} we are seeing the shock
524: wave interacting with the neutral shell.  Because the shell is
525: spherically symmetric, our line of sight will pass through shock fronts
526: inclined at equal angles toward us and away from us.
527: 
528: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
529: \section{Concluding Remarks}
530: 
531: We have detected strong O~VI emission from a Balmer filament in the
532: Cygnus Loop.  The \textit{FUSE} data presented here have the spectral
533: resolution not only to separate the two O~VI lines from each other and
534: from the Ly$\beta$ airglow, but also to separate kinematically
535: different emission components along the line of sight.  The O~VI flux
536: distribution, sampled at a spatial resolution of $\sim$4\arcsec, is
537: correlated with the H$\alpha$ emission.  A nonradiative shock with
538: properties derived from other ultraviolet observations can produce the
539: observed O~VI\@.  We have also resolved the O~VI emission in the form
540: of wings on the main component.  These high velocity emission
541: components are centered symmetrically in velocity space, at
542: $\pm140$~km~s$^{-1}$, and have equal flux.  They probably arise
543: in shocks driven into the spherically symmetric neutral shell,
544: which, in some cavity models, has a large covering factor around
545: the surface of the remnant.
546: 
547: O~VI~$\lambda\lambda$1032,1038 emission is an important channel through
548: which the Cygnus Loop SNR loses energy.  Measurements of the global
549: O~VI emission have been obtained using \textit{Voyager 1} and
550: \textit{2} \citep{bla91, van93}, and a rocket-borne experiment
551: \citep{ras92}.  These studies show that the O~VI luminosity is about
552: the same as the luminosity in the 0.1--4.0~keV x-ray band, the latter
553: based on \textit{Einstein} observations \citep{ku84}.  The contribution
554: of nonradiative shocks to the total O~VI emission may be estimated
555: based on our results.  The face-on surface brightness of the shock
556: causing the observed wing emission is
557: $\sim3\times10^{-16}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$ (total of
558: 1032\AA\ and 1038\AA\ lines).  Assuming a covering factor of 50\% for
559: such shocks, and distance to the Cygnus Loop of 440~pc, the derived
560: O~VI luminosity is $\sim1.2\times10^{36}$~erg~s$^{-1}$.  While
561: radiative shocks are known to contribute to the total O~VI emission
562: \citep{bla91b,dan00}, nonradiative shocks could account for most of the
563: total O~VI emitted by the Cygnus Loop.  The study of O~VI emission is
564: clearly important for unraveling the details of the interaction between
565: the Cygnus Loop and the surrounding material, and by extension for
566: understanding the nature of other middle-aged cavity SNRs.
567: 
568: \acknowledgements
569: 
570: We thank the people who worked on the development of \textit{FUSE} and
571: those operating the satellite, and we thank Bryce Roberts for help with
572: tracking down the accurate slit position angles from the Mission
573: Planning files.  We thank John Raymond for useful discussions, and for
574: giving us a copy of his shock code to use for model calculations.  RS
575: thanks Nancy Levenson for useful discussions.  We acknowledge financial
576: support provided by NASA contract NAS5-32985 to the Johns Hopkins
577: University.
578: 
579: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
580: \clearpage
581: \begin{thebibliography}
582: 
583: \bibitem[Blair et al.~(1991a)]{bla91}
584: Blair, W. P., Long, K. S., Vancura, O. \& Holberg, J. B.
585: 1991, \apj, 372, 202
586: 
587: \bibitem[Blair et al.~(1991b)]{bla91b}
588: Blair, W. P. et al.
589: 1991, \apjl, 379, L33
590: 
591: \bibitem[Blair et al.~(1999)]{bla99}
592: Blair, W. P., Sankrit, R., Raymond, J. C., \& Long, K. S.
593: 1999, \aj, 118, 942
594: 
595: \bibitem[Chevalier, Kirshner \& Raymond (1980)]{che80}
596: Chevalier, R. A., Kirshner, R. P. \& Raymond, J. C.
597: 1980, \apj, 235, 186
598: 
599: \bibitem[Cowie \& Songaila (1986)]{cow86}
600: Cowie, L. L., \& Songaila, A.
601: 1986, \araa, 24, 499
602: 
603: \bibitem[Danforth, Blair \& Raymond (2001)]{dan00}
604: Danforth, C. W., Blair, W. P., \& Raymond, J. C.
605: 2001, \aj, in press.
606: 
607: \bibitem[Fesen, Blair, \& Kirshner (1982)]{fes82}
608: Fesen, R. A., Blair, W. P., \& Kirshner, R. P.
609: 1982, \apj, 262, 171
610: 
611: \bibitem[Fesen \& Itoh (1985)]{fes85}
612: Fesen, R. A., \& Itoh, H.
613: 1985, \apj, 295, 43
614: 
615: \bibitem[Fitzpatrick (1999)]{fit99}
616: Fitzpatrick, E. L.
617: 1999, \pasp, 111, 63
618: 
619: \bibitem[Hamilton \& Fesen (1988)]{ham88}
620: Hamilton, A. J. S., \& Fesen, R. A.
621: 1988, \apj, 327, 178
622: 
623: \bibitem[Hartigan (1999)]{har99}
624: Hartigan, P.
625: 1999, \apj, 526, 274
626: 
627: \bibitem[Hester, Raymond \& Blair (1994)]{hes94}
628: Hester, J. J., Raymond, J. C., \& Blair, W. P.
629: 1994, \apj, 420, 721
630: 
631: \bibitem[Ku et al.~(1984)]{ku84}
632: Ku, W. H-M., Kahn, S. M., Pisarski, R., \& Long, K. S.
633: 1984, \apj, 278, 615
634: 
635: \bibitem[Levenson et al.~(1997)]{lev97}
636: Levenson, N. A. et al.
637: 1997, \apj, 484, 304
638: 
639: \bibitem[Levenson et al.~(1998)]{lev98}
640: Levenson, N. A., Graham, J. R., Keller, L. D., \& Richter, M. J.
641: 1998, \apjs, 118, 541
642: 
643: \bibitem[Long et al.~(1992)]{lon92}
644: Long, K. S., Blair, W. P., Vancura, O., Bowers, C. W.,
645: Davidsen, A. F., \& Raymond, J. C.
646: 1992, \apj, 400, 214
647: 
648: \bibitem[Moos et al.~(2000)]{moo00}
649: Moos, H. W., et al.
650: 2000, \apjl, 538, L1
651: 
652: \bibitem[Morton \& Dinerstein (1976)]{mor76}
653: Morton, D. C., \& Dinerstein, H. L.
654: 1976, \apj, 204, 1
655: 
656: \bibitem[Rasmussen \& Martin (1992)]{ras92}
657: Rasmussen, A., \& Martin, C.
658: 1992, \apjl, 396, L103
659: 
660: \bibitem[Raymond (1979)]{ray79}
661: Raymond, J. C.
662: 1979, \apjs, 39, 1
663: 
664: \bibitem[Raymond (1991)]{ray91}
665: Raymond, J. C.
666: 1991, \pasp, 103, 781
667: 
668: \bibitem[Raymond et al.~(1983)]{ray83}
669: Raymond, J. C., Blair, W. P., Fesen, R. A., \& Gull, T. R.
670: 1983, \apj, 275, 636
671: 
672: \bibitem[Raymond et al.~(1988)]{ray88}
673: Raymond, J. C., Hester, J. J., Cox, D., Blair, W. P.,
674: Fesen, R. A., \& Gull, T. R.
675: 1988, \apj, 324, 869
676: 
677: \bibitem[Sahnow et al.~(2000)]{sah00}
678: Sahnow, D. J., et al.
679: 2000, \apjl, 538, L7
680: 
681: \bibitem[Sankrit et al.~(2000)]{san00}
682: Sankrit, R., Blair, W. P., Raymond, J. C., \& Long, K. S.
683: 2000, \aj, 120, 1925
684: 
685: \bibitem[Vancura et al.~(1993)]{van93}
686: Vancura, O., Blair, W. P., Long, K. S., Raymond, J. C., \& Holberg, J. B.
687: 1993, \apj, 417, 663
688: 
689: \end{thebibliography}
690: 
691: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
692: \clearpage
693: \figcaption{Aperture locations overlaid on a WFPC2 H$\alpha$ image of
694: the Balmer filament in the Cygnus Loop.  The large box (dashed line)
695: represents the 30\arcsec$\times$30\arcsec\ LWRS aperture and the
696: smaller boxes represent the 4\arcsec$\times$20\arcsec\ MDRS aperture.
697: The shock direction is shown on the image.
698:              \label{SLITS} }
699: \figcaption{LWRS spectrum of the filament showing the region
700: around the O~VI~$\lambda\lambda$1032,1038 lines.
701: The line at about 1025.8\AA\ is Ly$\beta$, and the other airglow
702: lines are from O~I\@.  The feature on the blue wing of the
703: Ly$\beta$ is a detector artifact.
704:              \label{LWRS} }
705: \figcaption{MDRS spectra showing the region around the O~VI lines.
706: Position 1 refers to the aperture location that is furthest ahead, on
707: the optical filament (Figure \protect\ref{SLITS}).  Positions 2 and 3
708: are stepped back perpendicular to the shock normal.  All spectra
709: have been binned by 4 pixels.  Airglow lines are marked on the
710: Position 1 spectrum.
711:              \label{MDRS} }
712: \figcaption{O~VI line fluxes plotted as a function of velocity for the
713: three spectra.  The top panel shows the 1032\AA\ line and the bottom
714: panel shows the 1038\AA\ line.  The scales are the same on both plots
715: so they can be compared easily.  The units of F$_{\lambda}$ are
716: erg~s$^{-1}$~cm$^{-2}$~\AA$^{-1}$.  In the bottom panel, the positions
717: of a pair of H$_2$ Lyman band transitions are shown.  These H$_2$
718: transitions are responsible for absorbing the O~VI~$\lambda$1038 wing
719: emission.
720:              \label{VEL} }
721: \figcaption{The cumulative face-on O~VI $\lambda$1032 intensities
722: predicted by shock models plotted versus swept up hydrogen column
723: density.  Predictions from three models, with shock velocities of 170,
724: 180 and 190~km~s$^{-1}$ are shown.  All models use a pre-shock hydrogen
725: number density of 3~cm$^{-3}$.  The intensity for a given shock
726: velocity and swept up column scales linearly with the pre-shock
727: density.  The horizontal lines show the observed intensity divided by
728: 30 and 60, to account for the viewing aspect ratio of the shock front.
729: Intensity units are: $10^{-16}$~erg~s$^{-1}$~cm$^{-2}$~arcsec$^{-2}$.
730:              \label{MODS} }
731: 
732: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
733: \clearpage
734: \begin{deluxetable}{ccccc}
735: 
736: \tablecaption{List of Observations
737:                               \label{tblobs}}
738: \tablewidth{0pt}
739: \tablehead{
740:   \colhead{Obs. ID} & \colhead{$\alpha_{J2000}$} &
741:   \colhead{$\delta_{J2000}$} & \colhead{t$_{exp}$ (s)} & \colhead{Aperture}
742: }
743: \startdata
744: P1140401  &  $20^{\rm{h}}56^{\rm{m}}06\fs57$ & +31\arcdeg\ 56\arcmin\ 05\farcs6
745:           &  10234  & LWRS  \\
746: P1140402  &  $20^{\rm{h}}56^{\rm{m}}06\fs57$ & +31\arcdeg\ 56\arcmin\ 05\farcs6
747:           &  9559 & MDRS (P1)  \\
748: P1140501  &  $20^{\rm{h}}56^{\rm{m}}06\fs40$ & +31\arcdeg\ 56\arcmin\ 03\farcs5
749:           &  9247 & MDRS (P2)  \\
750: P1140601  &  $20^{\rm{h}}56^{\rm{m}}06\fs23$ & +31\arcdeg\ 56\arcmin\ 01\farcs4
751:           &  9767 & MDRS (P3)  \\
752: \enddata
753: 
754: \tablecomments{The slit position angle was 308\fdg5 (E of N) for all
755: observations.}
756: \end{deluxetable}
757: 
758: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
759: \clearpage
760: \begin{deluxetable}{lcccc}
761: 
762: \tablecaption{Observed Fluxes and Flux Ratios 
763:                                    \label{tblflux}}
764: \tablewidth{0pt}
765: \tablehead{
766:   \colhead{Component} & \colhead{MDRS P1} & 
767:   \colhead{MDRS P2} & \colhead{MDRS P3} & \colhead{LWRS}
768: }
769: \startdata
770: 1032\AA\ total   &  1.37  &  1.12  &  0.76  &  7.67   \\
771: 1032\AA\ center  &  1.22  &  0.95  &  0.55  & \nodata \\
772: 1038\AA\ \tablenotemark{a}  &  0.79  &  0.54  &  0.31  &  3.66 \\
773: 1032\AA\ wings\tablenotemark{b} & 0.15 & 0.17 & 0.21 & \nodata  \\
774: \\
775: I$_{1038}$/I$_{1032(center)}$  & 0.65 & 0.57 & 0.56 & 0.48\tablenotemark{c} \\
776: \enddata
777: 
778: \tablecomments{Flux units: 10$^{-13}$ erg s$^{-1}$ cm$^{-2}$.}
779: \tablenotetext{a}{The central component dominates, as the
780: wings are absorbed by interstellar H$_{2}$ lines.}
781: \tablenotetext{b}{This is simply the difference between the
782: total and center fluxes given in the first two rows.}
783: \tablenotetext{c}{The components have not been deconvolved in the
784: LWRS spectrum so the ratio I$_{1038}$/I$_{1032(total)}$ is tabulated.}
785: 
786: \end{deluxetable}
787: 
788: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
789: \clearpage
790: \begin{deluxetable}{lccccc}
791: 
792: \tablecaption{Observed Surface Brightnesses
793:                                    \label{tblsb}}
794: \tablewidth{0pt}
795: \tablehead{
796:   \colhead{Component} & \colhead{MDRS P1} & \colhead{MDRS P2} &
797:   \colhead{MDRS P3} & \colhead{LWRS} & \colhead{HUT\tablenotemark{a}}
798: }
799: \startdata
800: 1032\AA\ total   & 17.1 & 14.0 &  9.5 &  8.5 & 16.8 \\
801: 1038\AA\         &  9.9 &  6.8 &  3.9 &  4.1 & 10.1\tablenotemark{b} \\
802: \enddata
803: 
804: \tablecomments{Surface Brightness 
805: units: 10$^{-16}$ erg s$^{-1}$ cm$^{-2}$ arcsec$^{-2}$.}
806: \tablenotetext{a}{Data from Long et al.~(1992).  The HUT aperture
807: was $9\farcs4 \times 116\arcsec$.}
808: \tablenotetext{b}{Long et al.~(1992) reported the total flux in both
809: lines of the doublet.  The best fit to their spectrum yielded
810: a ratio I$_{1038}$/I$_{1032}$=0.6, which we use here.}
811: 
812: \end{deluxetable}
813: 
814: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
815: \clearpage
816: \begin{deluxetable}{ccccr}
817: 
818: \tablecaption{Model Parameters and Results
819:                                    \label{tblmod}}
820: \tablewidth{0pt}
821: \tablehead{
822:   \colhead{$v_s$} & \colhead{$n_0$} & \colhead{$I_0$\tablenotemark{a}} &
823:   \colhead{$N_{\rm{H}}$} & \colhead{$t_{shock}$} \\
824:   \colhead{(km s$^{-1}$)} & \colhead{(cm$^{-3}$)} &
825:   \colhead{} &
826:   \colhead{(cm$^{-2}$)} & \colhead{(years)}
827: }
828: \startdata
829: 160 & 3.80 & 2.71 & \nodata             & \nodata  \\
830: 180 & 3.00 & 2.41 & $1.86\times10^{17}$ &   110    \\
831: 200 & 2.43 & 2.17 & $2.21\times10^{17}$ &   140    \\
832: 220 & 2.01 & 1.97 & $1.52\times10^{18}$ &  1090    \\
833: 240 & 1.69 & 1.81 & $5.65\times10^{18}$ &  4420    \\
834: 260 & 1.44 & 1.67 & $1.26\times10^{19}$ & 10650    \\
835: 280 & 1.24 & 1.55 & $2.18\times10^{19}$ & 19890    \\
836: 300 & 1.08 & 1.45 & $3.29\times10^{19}$ & 32140    \\
837: \enddata
838: 
839: \tablenotetext{a}{Units are $10^{-16}$ erg s$^{-1}$ cm$^{-2}$ arcsec$^{-2}$.}
840: 
841: \end{deluxetable}
842: 
843: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
844: \end{document}
845: