1: \documentclass[12pt,preprint]{aastex}
2:
3: %% manuscript produces a one-column, double-spaced document:
4:
5: %% \documentclass[manuscript]{aastex}
6:
7: %% preprint2 produces a double-column, single-spaced document:
8:
9: %% \documentclass[preprint2]{aastex}
10:
11: %% Sometimes a paper's abstract is too long to fit on the
12: %% title page in preprint2 mode. When that is the case,
13: %% use the longabstract style option.
14:
15: %% \documentclass[preprint2,longabstract]{aastex}
16:
17: %% If you want to create your own macros, you can do so
18: %% using \newcommand. Your macros should appear before
19: %% the \begin{document} command.
20: %%
21:
22: \shorttitle{X-Ray Absorption Halo}
23: \shortauthors{Bregman and Lloyd-Davies}
24:
25: \begin{document}
26:
27: \title{X-Ray Absorption from the Milky Way Halo and the Local Group}
28:
29: \author{Joel N. Bregman and Edward J. Lloyd-Davies}
30: \affil{University of Michigan, Department of Astronomy, Ann Arbor, MI 48109}
31: \email{jbregman@umich.edu, ejdavies@umich.edu}
32:
33: \begin{abstract}
34: Million degree gas is present at near-zero redshift and is due either to a
35: gaseous Galactic Halo or a more diffuse but very massive Local Group
36: medium. We can discriminate between these models because the column
37: densities should depend on location in the sky, either relative to the
38: Galaxy bulge or to the M31-Milky Way axis. To search for these
39: signatures, we measured the \ion{O}{7} K$\alpha$ absorption line strength toward 25
40: bright AGNs, plus LMC X-3, using {\it XMM-Newton RGS\/} archival data. The data
41: are in conflict with a purely Local Group model, but support the Galactic Halo
42: model. The strongest correlation is between the \ion{O}{7} equivalent widths
43: and the {\it ROSAT\/} background emission measurement in the R45 band (0.4-1 keV),
44: for which \ion{O}{7} emission makes the largest single contribution. This
45: suggests that much of the \ion{O}{7} emission and absorption are cospatial, from which
46: the radius of a uniform halo appears to lie the range 15-110 kpc.
47: The present data do not constrain the type of halo gas model and
48: an equally good fit is obtained in a model where the gas density decreases
49: as a power-law, such as r$^{-3/2}$.
50: For a uniform halo with a radius of 20 kpc, the electron density would be
51: 9$\times$10$^{{\rm -4}}$ cm$^{-3}$, and the gas mass is 4$\times$10$^{{\rm 8}}$ M$_{\odot}$.
52: The redshift of the four highest S/N \ion{O}{7} measurements is consistent with a Milky
53: Way origin rather than a Local Group origin.
54:
55: \end{abstract}
56:
57: \keywords{Galaxy: halo --- ISM: abundances --- Local Group ---
58: X-rays: diffuse background}
59:
60: \section{Introduction}
61:
62: One of the important recent results from {\it Chandra, XMM,\/} and {\it FUSE\/} is
63: the detection of hot gas at nearly zero redshift. The low redshift of the
64: gas implies that it is either in the Local Group or in a halo around the
65: Milky Way, and discriminating between these two possibilities is critical.
66: If the gas is extended and fills the Local Group on the Mpc scale, its mass
67: exceeds the baryonic mass in all other galaxies in the Local Group. This
68: would account for some of the ``missing baryons'' in the Local Universe,
69: of fundamental importance for cosmology. Alternatively, if the gas is
70: local to the Milky Way, within a 100 kpc region, the baryonic mass is two
71: orders of magnitude lower and it would not dominate the local baryon
72: content.
73:
74: The tracers for this hot gaseous medium are the lines from common ions
75: that occur in the 10$^{{\rm 6}}$ - 10$^{{\rm 7}}$ K range, most
76: notably \ion{O}{7} and \ion{O}{8}. These
77: X-ray lines have been observed in absorption \citep{nica02, rasm03} with
78: both {\it Chandra\/} and {\it XMM\/}, and in emission with a
79: quantum-microcalorimeter spectrograph on a {\it Shuttle\/} payload, DXS
80: \citep{mcca02}. In addition, the \ion{O}{6} ion, most common in 10$^{{\rm 5.5}}$ K gas,
81: has been studied in over 100 sightlines with {\it FUSE\/}
82: \citep{semb03,wakk03}.
83:
84: The location and mass of the hot gas depends strongly on the assumed
85: relationships between the ions. If \ion{O}{6}, \ion{O}{7}, and \ion{O}{8} are cospatial,
86: \citet{nica02} cannot find a pure collisional model at a single temperature.
87: However, he can fit his data with a low density medium where photoionization is important,
88: with a temperature of 2.5$\times$10$^{{\rm 5}}$ K, a gas density of
89: n$_{{\rm e}}$ $\approx$ 6$\times$10$^{{\rm -}{6}}$ cm$^{{\rm -}{3}}$, and
90: a pathlength of 3 Mpc, corresponding to a gas mass
91: of 1.7$\times$10$^{{\rm 13}}$ M$_{{\rm \odot}{}}$. This gas mass is
92: considerably greater than the mass of all stars and cold gas within this
93: volume, which is about 2$\times$10$^{11}$ M$_{\odot}$ and
94: 1.2$\times$10$^{10}$ M$_{\odot}$, respectively \citep{braun91,cox00,vand00,cari06}.
95: This ``Local Group Model'' would help explain the missing
96: baryons in the local universe.
97:
98: Alternatively, \citet{rasm03} assume that only the X-ray lines are
99: cospatial, primarily \ion{O}{7} and \ion{O}{8}, leaving out \ion{O}{6}. They fix the
100: temperature from the \ion{O}{8}/\ion{O}{7} ratio, and, including the emission
101: observations, find that a collisional ionization model is possible with a
102: length scale of 0.15-1 Mpc. Their result implies a Local Group medium
103: with masses of 10$^{{\rm 13}}$ M$_{{\rm \odot}{}}$ for the 1 Mpc
104: length scale, and 10$^{{\rm 11}}$ M$_{{\rm \odot}{}}$ for the
105: 0.15 Mpc length scale, still more gas than the collective ISM of Local
106: Group galaxies. Recently, \citet{fang05b} pointed out the sensitivity of
107: this result to temperature and for a somewhat lower permitted
108: temperature, smaller length scales are allowed, typical of a Galactic halo
109: (50 kpc), and leading to considerably reduced gas masses.
110:
111: In yet another approach, one might assume that the \ion{O}{7} emission and
112: absorption are from the same gas, with some of the \ion{O}{8} and \ion{O}{6}
113: originating in other material. The X-ray emission line data was obtained
114: with the inaugural flight of a quantum calorimeter rocket flight by
115: \citet{mcca02} where they detected individual spectral lines responsible
116: for much of the soft X-ray background, leading to an emission measure
117: and metallicity. With an emission measure (n$^{{\rm 2}}${\it L\/}) and an absorption column
118: (n{\it L\/}), the size of the region {\it L\/} can be obtained and
119: this leads to {\it L\/} $\approx$ 50 kpc
120: \citep{sand02}, although this is a lower limit if one allows for filling
121: factors below unity. For filling factors near unity, this solution describes
122: a halo around the Galaxy where the gas density is
123: n$_{{\rm e}}$ $\sim$ 10$^{{\rm -}{4}}$ cm$^{{\rm -}{3}}$ and the
124: hot gas mass is $\sim$10$^{{\rm 9}}$-10$^{{\rm 10}}$ M$_{{\rm \odot}{}}$,
125: 2-3 orders of magnitude lower than the
126: solutions of Nicastro or Rasmussen. In this ``Galactic Halo'' solution, the
127: \ion{O}{6} is not cospatial, as argued by \citet{sav05a} and \citet{will05}, but might arise
128: nearby as the \ion{O}{7} gas cools and recombines or if there is a conductive
129: interface in a HVC.
130:
131: The ``Local Group'' and ``Galactic Halo'' models represent limiting cases.
132: The hot gas cannot be too small (e.g., in the disk) or it would
133: overproduce the soft X-ray background. Also, the absorbing region
134: cannot be larger than the virial radius of the Local Group or the gas
135: would not have a potential well to fall into and become heated. To
136: distinguish between these possibilities, one can search for an angular
137: signature to the absorption line strengths. In particular, for a Local
138: Group model, the absorption should be more pronounced for sight lines
139: closer to the Milky Way - M31 axis. For a Galactic Halo model, the
140: absorption should be more pronounced for sight lines across the Galaxy,
141: compared to those away from the galaxy. Searching for such signatures
142: requires a statistically significant number of sightlines with measured
143: X-ray absorption lines, and for this purpose, the \ion{O}{7} resonance absorption
144: line is the easiest to detect. There are relatively few {\it Chandra\/}
145: observations of this line toward bright sources, in part because the
146: gratings are not frequently used (but they have still proven quite useful;
147: \citealt{mcker04,fang05b,yao05,wang05}). However, for {\it XMM\/} observations, the combination
148: of the spectrograph ({\it RGS\/}) always operating plus the larger collecting
149: area has led to a significant number of viable targets for \ion{O}{7} absorption
150: measurements. Here we present the sample of the brightest AGNs that
151: have been observed by {\it XMM\/} and were available through the archive as of
152: April 2006. We show that this sample leads to a compelling case for the
153: Galactic Halo model of X-ray absorbing gas.
154:
155: \section{Object Selection and Data Analysis}
156:
157: We wish to sample a path length as large as the Local Group, which
158: implies the selection of extragalactic sources, with AGNs being the most
159: suitable. The only AGNs that can be used for X-ray absorption line
160: studies are bright and have been observed for a sufficiently long time. We
161: constructed a list of the brightest AGNs from the archival data obtained
162: by {\it ROSAT\/} and {\it ASCA\/}. This is a valuable starting point, but AGNs are
163: variable, so some sources that were fainter in the past can be brighter
164: now. Therefore, using the archive, we inspected the {\it XMM-Newton RGS\/}
165: data for dozens of targets and chose the most suitable objects. The
166: quantitative goal was to find sources for which the resulting uncertainty
167: in the \ion{O}{7} equivalent width is about 10 \AA\ or less, since the typical
168: absorption line strength was seen to be about 20 \AA\ in previous studies.
169: This led to a sample of 25 AGNs, plus we also included a source in the
170: LMC, LMC X-3 (Table 1).
171:
172: Many of these sources have been observed multiple times over the
173: lifetime of XMM and so the data can be combined to improve the
174: signal-to-noise. Unfortunately CCD 4 in the RGS2 failed shortly after
175: launch and since it covers the wavelength range 20.1$-$23.9 \AA\ only the
176: RGS1 data is useful. The data was reduced in the usual manner, using
177: the XMM SAS software (version 6.5), and was cleaned using an interative three sigma
178: clipping of a lightcurve from a background region on CCD 9, to remove
179: periods of high background that would degrade the signal-to-noise.
180: First order source and background spectra were produced from the
181: cleaned events using the SAS task RGSSPECTRUM and response matrices
182: were generated using RGSRMFGEN.
183: The RGS1 first order spectra from the observations of each source were
184: background subtracted, and after each spectrum was corrected for the
185: effective area, they were combined to produce a single spectrum for each source.
186: We note that the instrumental redistribution of photons in channel space is not corrected
187: for. This is considered below.
188:
189: To fit absorption lines, we excluded regions of known detector features
190: and regions of intrinsic AGN absorption and emission lines. For the \ion{O}{7}
191: resonance line at 21.603 \AA , there is a detector feature at about 21.82 \AA\
192: when the data are reduced with the standard SAS, so the 21.77-21.84 \AA\ region
193: was excluded in the fits (blank regions in our figures). Many of these bright
194: objects are nearby Seyfert galaxies, which have their own \ion{O}{7} lines,
195: some of which occur in absorption and some in emission. In most cases,
196: the lines are sufficiently redshifted that they do not contaminate the local
197: \ion{O}{7} absorption, but they restrict the region redward of 21.6 \AA\ that
198: can be used to define the baseline. The redshifted lines are marked on the
199: figures of the spectra when relevant: MCG-6-30-15, Mkn766, NGC 3783,
200: NGC 4051, NGC 5548, NGC 3516, NGC 4593, and IC4329a (Figures 1-6). In the few cases where
201: there are no useful continuum regions redward of 21.6 \AA , we used a flat
202: continuum to the blue side of the line. There were objects for
203: which the intrinsic absorption or emission line is badly blended with the
204: Galactic \ion{O}{7} absorption line, such as NGC 4151. Although it is suitably bright,
205: it has a broad \ion{O}{7} resonance emission line that overlaps with the z = 0 \ion{O}{7}
206: feature. The local absorption line, although present, lies in the sharply
207: rising wing of the NGC 4151 emission line, so the equivalent width of the
208: absorption feature cannot be measured reliably. NGC 4151, and other similar
209: objects were not included in the sample.
210:
211: When there are no intrinsic AGN features near 21.6 \AA , we fit a linear
212: continuum in the range 21.2 - 21.96 \AA , the upper limit being set by the
213: need to avoid Galactic \ion{O}{6} K$\alpha$ absorption at 22.05 \AA\ \citep{prad03}.
214: In addition to the continuum, we fit a Gaussian absorption line for the
215: Galactic \ion{O}{7} absorption feature. The least-squares minimization has
216: been performed by both the Simplex method and using a Levenberg-Marquardt
217: minimization and we arrive at the same result. In practice, there are five parameters
218: in this model: the continuum level; the slope of the continuum; the central
219: wavelength of the absorption line; the Gaussian $\sigma$ of the absorption line;
220: and the area of the line (the line depth), which can be negative or positive
221: (absorption and emission are permitted). Certain constraints can be applied
222: to some of these parameters. For example, the Gaussian fit cannot be narrower than the
223: instrumental width ($\sigma$ = 0.03 \AA\, or FWHM = 0.06 \AA ) and it probably will not be wider than
224: a velocity corresponding to the escape velocity of the Local
225: Group, which we estimate to be about 700 km s$^{-1}$, or $\sigma$ = 0.05 \AA .
226: Similarly, the location of the line center should not be outside of sensible
227: Local Group velocities, which sets a displacment of about 500 km s$^{-1}$,
228: or 0.036 \AA . In carrying out our fits, we calculate an uncertainty for each
229: parameter. Only in a few cases are the uncertainties for the line
230: center and line width smaller than the allowed range,
231: and for the five highest S/N sources, Mkn 421, PKS 2155-304, 3C273, MCG-3-30-15, and LMC X-3,
232: the median values of the central wavelength and line width are 21.605 \AA\,
233: and $\sigma$ = 0.081 \AA, with small ranges. Consequently, for fainter sources
234: where the central wavelength and line width cannot be accurately determined, we
235: fix the central wavelength at 21.603 \AA\ and the line width at $\sigma$ = 0.08 \AA.
236: When we relax these constraints, the values for the equivalent widths do not vary
237: greatly, although the uncertainty will naturally be larger. We note that the uncertainty
238: determined from photon counting statistics for the bins of the combined spectra
239: appear to overestimate the true
240: rms since the least square fits have a $\chi$$^{{\rm 2}}$ that is too small (one
241: can verify this by inspection in that a fit goes through nearly every error bar).
242: To achieve a typical $\chi$$^{{\rm 2}}$ for an acceptable fit, we estimate that the
243: uncertainty per point would need to be reduced by 25-80\% , depending on
244: the object. This is due to a misalignment between the raw channels and the flux
245: binning, introduced in the creation of the combined spectra, resulting in a lack of
246: independence between adjacent flux bins. To verify that neither the non-independence of
247: the flux bins nor the lack of treatment of the instrumental redistribution are
248: significantly affecting our results, we have simultaineously fitted
249: the individual spectra in XSPEC for some test cases.
250:
251: Upon examining the spectra, there were no absorption features that deviated from
252: expectations, with the exception of NGC 7469, where there is a possible
253: absorption feature at -900 km s$^{{\rm -1}}$ (98\% confidence), a velocity shift is
254: much greater than the instrumental uncertainty. However, with 26
255: objects, there is a near-unity chance of finding one such feature over a
256: range of $\pm$ 1000 km s$^{{\rm -1}}$, so we do not consider this a real detection.
257:
258: The resulting line strengths and statistical uncertainties are given in Table 2 and the
259: spectra are given in Figures 1-6. The objects are ordered by decending values
260: of their continuum S/N, which is a figure of merit for the quality to which
261: uncertainties on the equivalent widths can be obtained. Of the 26 objects, 10 have detections
262: above a 3$\sigma$ threshold, six are at the 2-3$\sigma$ level and the remainder are below 2$\sigma$.
263: Even non-detections are important, provided that the uncertainties are
264: adequately small. For this sample, 18 objects have uncertainties to the
265: equivalent widths of 10 \AA\ or less, and this forms the most useful subset of
266: our data. There happens to be a decrease in quality of the spectra after object 18,
267: but we also have used the remaining objects in some tests (although as those tests
268: are weighted by the errors, these last eight objects do not have a large influence
269: on the results).
270:
271: A few sources were observed with both {\it XMM\/} and {\it Chandra\/}, and in nearly
272: every case, the {\it XMM\/} data is of higher S/N. The exception is the source
273: LMC X-3, where the {\it XMM\/} and {\it Chandra\/} data \citep{wang05} are of
274: comparable S/N and their equivalent widths are nearly identical.
275: The line strengths measured by others for the {\it XMM\/} data give very
276: similar results (e.g., \citealt{rass06}), the differences being consistent
277: with different fitting methods. Other approaches to fitting the same data
278: (e.g., within XSPEC or IRAF) yielded nearly identical results.
279:
280: There are other, weaker absorption lines present, although there are not
281: enough of them to form a useful set for angular studies. The most
282: relevant is the \ion{O}{8} K$\alpha$ line, which can be
283: used with the \ion{O}{7} K$\alpha$ line to
284: constrain the temperature. The only useful objects are Mkn 421, PKS
285: 2155-304, and 3C273, reported on by \citet{rasm03}, so we use their
286: results for the \ion{O}{8} to \ion{O}{7} line ratios.
287:
288: \section{Analysis of the Data}
289:
290: The primary goal is to determine whether the absorption lines have an
291: angular signature that correlates with either the Local Group structure, or
292: various Milky Way properties. Given the limits of the data, it is
293: impractical to fit multi-parameter models to the data, so we choose simple
294: representative models, although we have explored a wide range of
295: models.
296:
297: \subsection{Local Group Models}
298:
299: We begin with the model for the Local Group, and as there are no widely
300: accepted models for the distribution of diffuse hot gas, here we assume
301: that the diffuse baryons are aligned along the axis defined by the Milky
302: Way and M31, which are the two dominant mass components of the
303: Local Group. If the center of mass of the diffuse baryons lies halfway between
304: M31 and the Milky Way, then the greatest column density should occur in
305: a sightline toward M31 and the smallest column would occur in the anti-M31
306: direction. Consequently, a general test is whether the column
307: density decreases with increasing angle from M31. This
308: prediction is not confirmed by the data (Figure 7), where the equivalent
309: widths weakly increase with increasing angle from M31, rather than decrease.
310:
311: With a more detailed model, one can predict the expected distribution of
312: the column densities for comparison to the data. The specific model that
313: we adopt is where M31 and the Milky Way lie at the two foci of an
314: ellipsoid in which the gas density is uniform. The model is completely
315: defined by choosing a semi-major axis, and we show the results for a
316: model where the semi-major axis is twice the separation between M31
317: and the Milky Way (semi-major axis of 1.54 Mpc). The normalization of
318: the column density of the model is a remaining parameter, which we
319: choose to minimize the $\chi$$^{{\rm 2}}$. Inspection of the model fit (Figure 8)
320: indicates it is a poor fit, and as the reduced $\chi$$^{{\rm 2}}$ = 5.5,
321: it is an unacceptable fit. For semi-major axis from 0.8-2 Mpc, all models produce
322: unacceptable fits with no positive correlation coefficients, showing that it
323: is unlikely that the \ion{O}{7}-absorbing gas is distributed through the Local
324: Group.
325:
326: We have also compared the location of the absorbers with the individual
327: Local Group galaxies to determine if they have a specific signature. For
328: example, if individual galaxies were to have their own extended gaseous
329: halos, there would be stronger absorption lines for sight lines near those
330: galaxies (Figure 9). If galaxy gaseous halos scale with galaxy size, then the largest
331: external galaxy halo should be around M31, where the closest sightline is
332: Akn 564, 29$\degr$ away, or a closest approach of 380 kpc. The
333: equivalent width in this sight line (12.3 $\pm$ 4.6 m\AA ) appears to be less and
334: is certainly not substantially greater than the median equivalent width.
335: This implies that if M31 has a significant gaseous halo, it is less than 380
336: kpc in radius.
337:
338: The next largest galaxy with a nearby sightline is M33, where 3C 59
339: is 9$\degr$ away, or a 130 kpc impact parameter. The \ion{O}{7} equivalent
340: width, 60.9 $\pm$ 19.2 m\AA , has a rather large uncertainty, so it is difficult
341: determining whether there a hot halo exists around M33. Two dwarf
342: galaxies have nearby sightlines: DDO 210, where Mrk 509 is 3$\degr$
343: away (42 kpc); and Leo I \& II, where Ton 1388 is 3.5$\degr$ away (14
344: kpc). The equivalent width for Ton 1388 is above average but fairly uncertain,
345: 34.5 $\pm$ 15.7 m\AA , and is within 1$\sigma$ of the median value for the sample.
346: For Mrk 509, the detection is quite good (25.9 $\pm$ 7.3 m\AA ) and it also
347: is within 1$\sigma$ of the median value. So for Leo I \& II, there
348: appears to not be a gaseous halo either, although this dwarf pair does
349: not have HI nor does it have star formation \citep{mateo98}. The
350: observations, which are only of high quality for M31 and Leo I \& II,
351: argue against substantial halos with a size an order of magnitude larger
352: than the optical galaxy. The data do not rule out smaller galactic halo
353: that are comparable or a few times larger than the host galaxy.
354:
355: There is one useful source within the LMC, LMC X-3 and it lies in the
356: disk, 6.9 $\deg$ (6.7 kpc) from the center of the LMC. If a LMC halo
357: surrounds the galaxy, this sightline would sample half of the LMC halo
358: plus the intervening gas between it and the Sun. The source LMC X-3 has
359: an equivalent width that is very close to the median value (21.0 $\pm$
360: 5.0 m\AA ) and the line width is no wider than average (0.063 $\pm$ 0.012 \AA ),
361: so there is no indication of additional absorption toward this source
362: caused by a LMC halo. Finally, the line center (21.592 $\pm$ 0.016 \AA ) is
363: consistent with the velocity of the Milky Way, but 2$\sigma$ lower in redshift
364: than the LMC, which would have produced a line at 21.623 \AA . This is
365: similar to the conclusions of \citet{wang05}, who used {\it Chandra\/} data
366: to observe LMC X-3.
367:
368: \subsection{Galactic Halo Models}
369:
370: Another model that we consider is for a hot Galactic Halo, where the
371: simplest assumption is that a hot gas halo is centered on the Galaxy
372: center. In this picture, the column density of hot gas should be a
373: decreasing function of angle away from the Galactic center, and as shown
374: in Figure 10, there appears to be the expected inverse
375: correlation. An unweighted linear
376: y-on-x fit shows that there is an inverse correlation with angle at the 95\%
377: confidence level, supporting the Galaxy Halo model. To be more
378: quantitative, yet to keep the model simple, we calculate relative column
379: densities for a uniform sphere of radius R$_{{\rm H}}$, which is the only free
380: parameter (the equivalent width is constrained to be zero for zero path
381: length); we use 8.5 kpc as the distance of the Sun from the center of the
382: Galaxy. A fit of similar significance is found (Figure 11), where a best-fit
383: $\chi$$^{{\rm 2}}$ model leads to R$_{{\rm H}}$ = 20.6 kpc, and the 95\%
384: confidence bounds are 15-110 kpc. For the best-fit model, the $\chi$$^{{\rm 2}}$ is
385: unacceptably large, which can occur if the statistical uncertainty in the
386: equivalent widths are smaller than the variation due to structure in the
387: absorbing medium. To obtain an acceptable value of $\chi$$^{{\rm 2}}$, we added
388: an intrinsic scatter in the equivalent widths attributable
389: to the structure of the absorbing gas of 4 m\AA , or 23\% of the
390: median value of the equivalent width.
391:
392: The above model represents a spherical halo where the density {\it n\/} $\propto$
393: {\it r$^{{\rm 0}}$\/}, but we can also consider models where the density is a decreasing
394: function of radius, such as {\it n\/} $\propto$ $(1+(r/r_{c})^{2})^{-3 \beta /2}$, where $r_{c}$ is the core
395: radius. The value of $r_{c}$ in most other galaxies with hot gas is less than 1
396: kpc, which is significantly smaller than the radius of closest approach for
397: all objects in the sample. Therefore, the value of $r_{c}$ is not relevant
398: for our sample and we can write the density function as a power-law, {\it n\/}
399: $\propto$ $r^{-3 \beta\ }$, and we choose to examine {\it n\/} $\propto$ $r^{-3/2}$, as this is similar to
400: that found for the hot gas in ellipticals ($\beta$=1/2 in the ``beta'' models used
401: for ellipticals, e.g., \citealt{osull03}). While this leads to a gas mass that
402: increases with radius as r$^{3/2}$ and an emission measure that increases
403: logarithmically with radius, the column density is convergent at large radius.
404: There is only a modest correlation between the theoretical and observed
405: equivalent widths, significant at the 92\% confidence level (Figure 12). This model is closely
406: associated with the Galaxy as well.
407: From the available data, we cannot distinguish between the uniform model
408: and this ``beta'' model. More complicated models can
409: certainly be constructed, but given the quality of the data, we do not
410: believe it is warranted at this time.
411:
412: If this absorption is associated with a Galactic Halo, it might be correlated
413: with the soft X-ray background of the Galaxy. The 1/4 keV X-ray
414: background is dominated by the gas within a few hundred pc of the Sun,
415: plus a flattened hot halo of height 2 kpc \citep{snow97}.
416: This is not the X-ray emission that we expect the absorption features to be related to and
417: there is no correlation between the 1/4 keV surface brightness and the
418: \ion{O}{7} equivalent widths. At higher energies, such as the 3/4 keV band,
419: the emission is a mixture of emission from hot Galactic gas plus emission
420: from unresolved AGNs \citep{gilli01}. Also, the absorption of the X-rays
421: by neutral Galactic gas is less of a problem at these energies. It is
422: probably the best energy at which to detect diffuse Galactic emission and
423: it has been known that there is enhanced soft X-ray emission in the
424: general direction of the bulge \citep{snow00}. A correlation
425: exists between R45 and the \ion{O}{7} equivalent width
426: at the 99\% confidence level (weighted or unweighted fit), without
427: requiring that the fit go through the origin (Figure 13).
428: This relationship is better than the correlation found
429: with angle from the Galactic Center or with either of the above Galactic
430: Halo models. For this fit, the $\chi$$^{{\rm 2}}$ = 45 (16 dof), which is unacceptable at
431: the 99\% confidence level, so we added a additional
432: rms in the equivalent width, as described above. For an additional rms of
433: 3.5 m\AA\ (19\% of the median equivalent width), the $\chi$$^{{\rm 2}}$ is reduced to
434: an acceptable 28.5
435: and the probability of a weighted fit is reduced to the 98\% confidence level.
436: The high initial $\chi$$^{{\rm 2}}$ is due to one low point and if this
437: outlier were removed, the quality of the fit would improve.
438: By using the entire sample and creating error-weighted bins (Figure 14), we
439: obtain a similar level of significance, although a somewhat lower
440: additional rms is required in this case (3.0 m\AA ) to obtain
441: acceptable values of $\chi$$^{{\rm 2}}$.
442:
443: The R45 band emission contains not only large-scale Galactic emission,
444: but individual bright features due to multiple supernovae, such as the
445: North Polar Spur. We examined the location of the objects to determine
446: if they lie in bright regions of these features, but there are hardly any
447: coincidences. The quasar 3C273 is located near a fainter part of the
448: North Polar Spur, but it's \ion{O}{7} equivalent width is not remarkable (24.6
449: $\pm$ 3.3 m\AA ). The strongest component of the R45 emission is the \ion{O}{7}
450: emission triplet, so the relationship between the R45 emission and the
451: absorption equivalent widths suggests a correlation between the \ion{O}{7}
452: emission and absorption. This implies that the \ion{O}{7} emission and
453: absorption is largely cospatial.
454:
455: Another piece of evidence for the Galactic Halo picture is the mean
456: redshift of our four highest S/N AGNs. There is a random uncertainty
457: in the wavelength calibration of 0.008 \AA (XMM-Newton Users's Handbook), or 111 km s$^{-1}$, so it
458: is helpful to average together several objects. Using these four AGNs,
459: the unweighted mean central wavelength is 21.608 \AA (+70 km s$^{-1}$), and the inferred
460: standard deviation of a single point is 0.006 \AA , but to be conservative, we assume that
461: it is 0.010 \AA , as this is the median value for the uncertainty in the
462: central wavelength from individual fits. Then, the uncertainty in the mean
463: central wavelength is 0.005 \AA (55 km s$^{-1}$), which places this mean about 1$\sigma$ of
464: the Milky Way (z = 0, or 21.603 \AA ). The Local Group is probably at
465: -250 km s$^{-1}$, a wavelength of 21.585 \AA , 4.6$\sigma$ different from the
466: measured mean value. This suggests that the absorption is most likely
467: associated with the Milky Way and not the Local Group.
468:
469: The conversion of the \ion{O}{7} equivalent width to a hydrogen density,
470: which will yield a length scale as well, requires knowledge of the oxygen
471: column density, the oxygen abundance, and the fraction of oxygen in the
472: form of \ion{O}{7}. When the optical depth of the absorption line is low, there is
473: a simple linear relationship between the equivalent width and the column
474: density, N(\ion{O}{7}) = 3.48$\times$10$^{{\rm 14}}$ EW, where the equivalent width (EW)
475: has units of m\AA\ and the column density, N(\ion{O}{7}) has units of cm$^{-2}$.
476: If the lines are resolved or if the gaseous medium is turbulent at the sound
477: speed, then the optical depth is small, but if the lines have a Doppler width
478: for oxygen (no turbulent or significant ordered motion), the optical depth
479: at the line center can be 0.1-0.4. As the optical depth rises, the higher order
480: lines, such as \ion{O}{7} K$\beta$, become of equal strength to the K$\alpha$ line.
481: We examined the strength of the K$\beta$ for the highest S/N spectrum, of Mkn 421,
482: and find a value of 3.6 $\pm$ 1.1 m\AA , although an instrumental feature
483: contaminates the red side of the line. The K$\beta$ to K$\alpha$ ratio is
484: 0.30 $\pm$ 0.11, which is within 1$\sigma$ of the ratio of the f values, 0.21.
485: This is consistent either with a small or modest optical depth. Here we will
486: calculate column densities in the limit of low optical depth.
487:
488: The ionization fraction can be determined by forming the ratio of \ion{O}{7} with
489: an adjacent ionization state, such as \ion{O}{6} or \ion{O}{8}, provided that the
490: different ionization states are cospatial. The relationship between \ion{O}{7}
491: and \ion{O}{6} is poor and both \citet{sav05a} and \citet{will05} argue that the two are probably not
492: closely related. It is tempting to relate \ion{O}{8} to \ion{O}{7}, and a correlation
493: between these two quantities would lend support to this possibility.
494: Unfortunately, there are only a few objects for which we can determine
495: useful measurements of \ion{O}{8} absorption, the errors are significant, and
496: there is no statistically meaningful correlation between the two quantities \citep{rasm03}.
497: It is possible to have a multi-temperature medium where the hotter
498: component leads to \ion{O}{8} while a slightly cooler component dominates
499: the \ion{O}{7}. As the above study showed that \ion{O}{7} emission and absorption
500: are correlated, we will use these two measurements to obtain densities
501: and length scales, later considering whether it is consistent with the \ion{O}{8}
502: to \ion{O}{7} column density ratio.
503:
504: For a Galactic Halo model, we expect the metallicity to be similar to the
505: value in the ISM and the Sun. For oxygen, the ``Solar'' abundance has
506: changed considerably in the past two decades and is still not a settled
507: issue. The most commonly cited Solar abundances come from \citet{ande89},
508: where log{\it N$_{{\rm O}}$\/} = 8.93, but \citet{holw01} recommends a value of
509: 8.736, a 35\% reduction from the older value. \citet{aspl05} advocates a
510: value at least 20\% lower, but here we adopt the value 8.74 for the oxygen
511: abundance. For an ion fraction {\it f\/} = 0.5, this leads to an electron column of
512: {\it N$_{{\rm e}}$ = n$_{{\rm e}}$L = 5.8$\times$10$^{{\rm 19}}$
513: (N$_{OVII}$/10$^{{\rm 16}}$ cm$^{{\rm -2}}$) (f/0.5)$^{{\rm -1}}$\/}
514: cm$^{-2}$, where {\it n$_{{\rm e}}$\/} is the
515: electron density and {\it L\/} is the path length for a constant density slab. The
516: emission measure is available from only one line of sight, toward {\it l,b\/} = 90$\degr$,
517: 60$\degr$ and with a 1 sr field of view \citep{mcca02}. Toward this general
518: location, R45 is typically 200 cnt s$^{-1}$ arcmin$^{-2}$, above the mean of the lines of sight we
519: have studied. For sight lines at similar values of R45, the typical
520: equivalent width is 25 m\AA , or a column density of {\it N$_{OVII}$\/}
521: = 8.7$\times$ 10$^{{\rm 15}}$ cm$^{-2}$ and
522: {\it N$_{{\rm e}}$= 5.0$\times$10$^{{\rm 19}}$\/} cm$^{-2}$. The emission measure at Solar
523: metallicity is EM = n$_{{\rm e}}$$^2$L = 0.009 cm$^{-6}$ pc \citep{mcca02}, so one can use
524: this, adjusted to our oxygen abundance, and the column density to
525: solve for the length scale {\it L\/}, and we find {\it L\/} =
526: 19 kpc and {\it n$_{{\rm e}}$\/} = 9$\times$10$^{{\rm -4}}$ cm$^{-3}$.
527: This length scale supports the Galactic Halo model for the hot
528: gas distribution. Aside from the temperature dependence,
529: {\it L\/} $\propto$ {\it f\/}$^{{\rm -2}}$ {\it (O/H)\/}$^{{\rm -1}}$,
530: so for {\it f\/} = 1, the halo could be as small as 5 kpc.
531: A value smaller than 5 kpc is ruled out because it would have the consequence
532: of raising the emission measure above the observed value.
533:
534: The \ion{O}{8} to \ion{O}{7} column density ratio is consistent with a gas whose
535: temperature is logT = 6.2-6.3. This column density ratio changes rapidly
536: in this temperature range, so even a modest amount of temperature variation along
537: the line of sight (e.g., 20-30\%) can make a difference in this ratio by a
538: factor of several, as \citet{fang05b} points out. The analysis of the emission
539: in the R45 band indicates a temperature of logT = 6.0-6.1 \citep{snow00,mcca02},
540: so the \ion{O}{7} absorption may also be occurring in a
541: temperature a bit too low to produce the \ion{O}{8} absorption as well.
542: Temperature variation at this level is seen in some early-type galaxies
543: (e.g., NGC 4636; \citealt{jones02}), and if a positive temperature gradient
544: is present in the Milky Way halo, there could be radial regions where
545: \ion{O}{7} is dominant and more distant regions where \ion{O}{8} is dominant. A
546: radial decomposition of the \ion{O}{7} to \ion{O}{8} ionization fraction ratio is
547: possible in principle, but the present data are inadequate for such work.
548:
549: \section{Summary and Final Comments}
550:
551: We sought to determine whether the X-ray absorption, primarily in the
552: \ion{O}{7} resonance line, were due to a Local Group medium or a hot
553: Galactic Halo. It is possible to address this question by searching for the
554: spatial signature of these two models, which are distinct. The Local
555: group is expected to be elongated along the axis defined by the Milky
556: Way and M31. For our model, where the two galaxies are placed at the
557: foci of an ellipsoid of rotation about the MW-M31 axis, the greatest
558: column would result toward M31 and the smallest in the opposite
559: direction. For the Milky Way Halo model where the halo is comparable
560: to the size of the Galaxy, the greatest columns would occur when looking
561: across the Galaxy, in the general direction of the bulge. By good fortune,
562: these two models are nearly orthogonal in that the minimum direction for
563: the Local Group model, the anti-M31 direction, is at {\it l, b\/} =
564: 301$\degr$, 21.6$\degr$, where the Galactic Halo model would predict
565: an excess compared to the mean sightline.
566:
567: The comparison between the absorption line data and the models
568: shows that the Galactic Halo model is preferred.
569: The data do not permit us to accurately determine the properties of the halo
570: so we adopted the simplest possible model, a uniform spherical halo.
571: In the context of this model, the characteristic radial size lies in
572: the range 15-100 kpc. The redshift of the
573: lines is consistent with a Galactic origin rather than a Local Group origin.
574: The absorption line data correlates best with the R45 (0.4-1 keV) emission,
575: for which the biggest contribution is the \ion{O}{7} emission triplet. This
576: suggests that the \ion{O}{7} emission and absorption are cospatial to some
577: degree, and for a cospatial model, we derive an independent measure of
578: the size of the halo, which is about 20 kpc. Constraints on the
579: temperature may be provided by using the \ion{O}{8} absorption as well,
580: which leads to a temperature of logT = 6.1, as discussed previously
581: \citep{mcca02, fang05b}. These results are in good agreement with
582: the work of \citet{fang05b}, who also argue for a Galactic origin based
583: on {\it Chandra\/} obervations and on the absence of X-ray absorption
584: by other groups.
585:
586: As an example of a possible halo, when we consider a
587: uniform Galaxy halo of radius 20 kpc, the mass for
588: the hot gas is about 4 $\times$10$^{{\rm 8}}$ M$_{\odot}$, which is
589: considerable, but only a small fraction of the gaseous content of the
590: Galaxy, and orders of magnitude less than the values suggested for the
591: Local Group models \citep{nica02}. This amount of mass is similar to that found in the
592: less X-ray luminous early-type galaxies, and it is comparable to the amount of
593: mass suggested to be present due to Galactic Fountain models \citep{
594: rosen95,avill04}. Also, this mass is much less than the external baryon mass
595: predicted by in some N-body/gasdynamical simulations \citep{somm06}, which
596: is expected to be comparable to the baryonic mass of the disk plus bulge stars
597: and the gas in the Galaxy (5 $\times$ 10$^{{\rm 10}}$ M$_{\odot}$).
598: While we have assumed a Solar-like metallicity, we note
599: that the oxygen abundance cannot be significantly lower, as it would lead to a
600: higher electron column density that would exceed the values
601: determined from the dispersion measure toward pulsars. At the pole, the
602: pulsar dispersion measure implies an electron column of 5.1 $\times$ 10$^{{\rm 19}}$
603: cm$^{-2}$ \citep{taylor93}, comparable to the column inferred from the
604: \ion{O}{7}. Already, there is hardly room for the contribution from the 10$^{{\rm 4}}$ K gas
605: that is known to exist around the disk \citep{reyn04}, so any decrease in the
606: oxygen abundance or the \ion{O}{7} fraction can be excluded. The constraint
607: provided by the pulsar dispersion measure argues for near-Solar
608: abundances and an \ion{O}{7} ionization near unity.
609:
610: A weak limit on the size of hot halos around other galaxies is provided by
611: the failure to find enhanced absorption toward M31, where the sightline
612: of the nearest object passes 380 kpc from its center. Also, there was no
613: substantial enhancement in the \ion{O}{7} (or \ion{O}{8}) absorption toward LMC
614: X-3, which suggests that there is not a significant 10$^{{\rm 6}}$ K halo around the
615: LMC. This is consistent with the results of \citet{wang05}, who showed
616: that the redshift of the line center, as determined with {\it Chandra\/}, is
617: inconsistent with an LMC origin and consistent with a Galactic origin. A
618: cooler ionized halo is known to exist around the LMC, which is
619: responsible for the enhanced dispersion measure toward pulsars \citep{manc06}.
620:
621: This study does not rule out some component
622: of the \ion{O}{7} emission lying in a Local Group medium. Certainly, the
623: Local Group is not devoid of gas, so it must have some \ion{O}{7} as well.
624: However, as \citet{fang05b} discuss, if galaxy groups commonly had a
625: significant \ion{O}{7} column, it would be detected in sight lines toward
626: background AGNs at a rate in conflict with existing limits.
627:
628: There is more work that can be carried out with {\it XMM\/} data, as new
629: observations are improving the S/N for some AGNs and additional AGNs
630: are being observed for significant amounts of time. However, {\it XMM\/} and
631: {\it Chandra\/} will not be able to provide us with $\sim$ 10$^{{\rm 2}}$ sight lines with high
632: S/N line measurements, which would be needed to map out the structure
633: of the hot halo, much in the same way as has been accomplished for lower
634: temperature gas with {\it IUE\/}, {\it HST\/}, and {\it FUSE\/} spectra. An improvement in
635: the X-ray capabilities of an order of magnitude would provide the
636: necessary data, and this will become possible with {\it Constellation-X\/}.
637:
638: \acknowledgements
639: The authors would like to thank Andy Rasmussen, Wilt Sanders, Jimmy Irwin,
640: and Renato Dupke for their comments and insight. We are indebted to those who
641: have developed and maintained the XMM archive, upon which this work was based.
642: Also, we would like to acknowledge support from NASA for these activities,
643: through the Long Term Space Astrophysics grant NAG5-10765, and also NAG5-13137.
644:
645: \begin{thebibliography}{}
646:
647: \bibitem[de Avillez \& Breitschwerdt(2004)]{avill04} de
648: Avillez, M., \& Breitschwerdt, D.\ 2004, \apss, 292, 207
649:
650: \bibitem[Anders \& Grevesse(1989)]{ande89} Anders, E., \&
651: Grevesse, N.\ 1989, \gca, 53, 197
652:
653: \bibitem[Asplund et al.(2005)]{aspl05} Asplund, M., Grevesse,
654: N., \& Sauval, A.~J.\ 2005, ASP Conf.~Ser.~336: Cosmic Abundances as
655: Records of Stellar Evolution and Nucleosynthesis, 336, 25
656:
657: \bibitem[Braun(1991)]{braun91} Braun, R.\ 1991, \apj, 372, 54
658:
659: \bibitem[Cagnoni et al.(2003)]{cagn03} Cagnoni, I., Nicastro,
660: F., Maraschi, L., Treves, A., \& Tavecchio, F.\ 2003, New Astronomy Review,
661: 47, 561
662:
663: \bibitem[Carignan et al.(2006)]{cari06} Carignan, C., Chemin,
664: L., Huchtmeier, W.~K., \& Lockman, F.~J.\ 2006, \apjl, 641, L109
665:
666: \bibitem[Cox(2000)]{cox00} Cox, A.~N.\ 2000, Allen's
667: astrophysical quantities, 4th ed.~Publisher: New York: AIP Press; Springer,
668: 2000.~Editedy by Arthur N.~Cox.~ ISBN: 0387987460
669:
670: \bibitem[Fang et al.(2005)]{fang05a} Fang, T., Croft, R.~A.~C.,
671: Sanders, W.~T., Houck, J., Dav{\'e}, R., Katz, N., Weinberg, D.~H., \&
672: Hernquist, L.\ 2005, \apj, 623, 612
673:
674: \bibitem[Fang et al.(2006)]{fang05b} Fang, T., Mckee, C.~F.,
675: Canizares, C.~R., \& Wolfire, M.\ 2006, \apj, 644, 174
676:
677: \bibitem[Fox et al.(2004)]{fox04} Fox, A.~J., Savage, B.~D.,
678: Wakker, B.~P., Richter, P., Sembach, K.~R., \& Tripp, T.~M.\ 2004, \apj,
679: 602, 738
680:
681: \bibitem[Gilli et al.(2001)]{gilli01} Gilli, R., Salvati, M.,
682: \& Hasinger, G.\ 2001, \aap, 366, 407
683:
684: \bibitem[Holweger(2001)]{holw01} Holweger, H.\ 2001, AIP
685: Conf.~Proc.~598: Joint SOHO/ACE workshop ''Solar and Galactic
686: Composition'', 598, 23
687:
688: \bibitem[Jones et al.(2002)]{jones02} Jones, C., Forman, W.,
689: Vikhlinin, A., Markevitch, M., David, L., Warmflash, A., Murray, S., \&
690: Nulsen, P.~E.~J.\ 2002, \apjl, 567, L115
691:
692: \bibitem[Mateo(1998)]{mateo98} Mateo, M.~L.\ 1998, \araa, 36,
693: 435
694:
695: \bibitem[Manchester et al.(2006)]{manc06} Manchester, R.~N.,
696: Fan, G., Lyne, A.~G., Kaspi, V.~M., \& Crawford, F.\ 2006, \apj, 649, 235
697:
698: \bibitem[McCammon et al.(2002)]{mcca02} McCammon, D., et al.\
699: 2002, \apj, 576, 188
700:
701: \bibitem[McKernan et al.(2004)]{mcker04} McKernan, B., Yaqoob,
702: T., \& Reynolds, C.~S.\ 2004, \apj, 617, 232
703:
704: \bibitem[Nicastro et al.(2005)]{nica05b} Nicastro, F., et al.\
705: 2005, \apj, 629, 700
706:
707: \bibitem[Nicastro et al.(2005)]{nica05a} Nicastro, F., et al.\
708: 2005, \nat, 433, 495
709:
710: \bibitem[Nicastro et al.(2002)]{nica02} Nicastro, F., et al.\
711: 2002, \apj, 573, 157
712:
713: \bibitem[O'Sullivan et al.(2003)]{osull03} O'Sullivan, E.,
714: Ponman, T.~J., \& Collins, R.~S.\ 2003, \mnras, 340, 1375
715:
716: \bibitem[Paerels \& Kahn(2003)]{paer03} Paerels, F.~B.~S., \&
717: Kahn, S.~M.\ 2003, \araa, 41, 291
718:
719: \bibitem[Pradhan et al.(2003)]{prad03} Pradhan, A.~K., Chen,
720: G.~X., Delahaye, F., Nahar, S.~N., \& Oelgoetz, J.\ 2003, \mnras, 341, 1268
721:
722: \bibitem[Rasmussen et al.(2003)]{rasm03} Rasmussen, A., Kahn,
723: S.~M., \& Paerels, F.\ 2003, ASSL Vol.~281: The IGM/Galaxy Connection.~The
724: Distribution of Baryons at z=0, 109
725:
726: \bibitem[Rasmussen et al.(2007)]{rass06} Rasmussen, A.~P.,
727: Kahn, S.~M., Paerels, F., Herder, J.~W.~d., Kaastra, J., \& de Vries, C.\
728: 2007, \apj, 656, 129
729:
730: \bibitem[Reynolds(2004)]{reyn04} Reynolds, R.~J.\ 2004,
731: Advances in Space Research, 34, 27
732:
733: \bibitem[Rosen \& Bregman(1995)]{rosen95} Rosen, A., \&
734: Bregman, J.~N.\ 1995, \apj, 440, 634
735:
736: \bibitem[Sanders et al.(2002)]{sand02} Sanders, W.~T., et al.\
737: 2002, Bulletin of the American Astronomical Society, 34, 1178
738:
739: \bibitem[Savage et al.(2005)]{sav05a} Savage, B.~D., Wakker,
740: B.~P., Fox, A.~J., \& Sembach, K.~R.\ 2005, \apj, 619, 863
741:
742: \bibitem[Sembach et al.(2003)]{semb03} Sembach, K.~R., et al.\
743: 2003, \apjs, 146, 165
744:
745: \bibitem[Snowden et al.(1997)]{snow97} Snowden, S.~L., et al.\
746: 1997, \apj, 485, 125
747:
748: \bibitem[Snowden et al.(2000)]{snow00} Snowden, S.~L.,
749: Freyberg, M.~J., Kuntz, K.~D., \& Sanders, W.~T.\ 2000, \apjs, 128, 171
750:
751: \bibitem[Sommer-Larsen(2006)]{somm06} Sommer-Larsen, J.\ 2006,
752: \apjl, 644, L1
753:
754: \bibitem[Taylor \& Cordes(1993)]{taylor93} Taylor, J.~H., \&
755: Cordes, J.~M.\ 1993, \apj, 411, 674
756:
757: \bibitem[van den Bergh(2000)]{vand00} van den Bergh, S.\ 2000,
758: The galaxies of the Local Group, by Sidney Van den Bergh.~Published by
759: Cambridge, UK: Cambridge University Press, 2000 Cambridge Astrophysics
760: Series Series, vol no: 35, ISBN: 0521651816.,
761:
762: \bibitem[Wakker et al.(2003)]{wakk03} Wakker, B.~P., et al.\
763: 2003, \apjs, 146, 1
764:
765: \bibitem[Wang et al.(2005)]{wang05} Wang, Q.~D., et al.\ 2005,
766: \apj, 635, 386
767:
768: \bibitem[Williams et al.(2005)]{will05} Williams, R.~J., et
769: al.\ 2005, \apj, 631, 856
770:
771: \bibitem[Yao \& Wang(2005)]{yao05} Yao, Y., \& Wang, Q.~D.\
772: 2005, \apj, 624, 751
773:
774: \end{thebibliography}
775:
776: \clearpage
777:
778: \begin{deluxetable}{rlrrcrr}
779: \tabletypesize{\scriptsize}
780: %\rotate
781: \tablecaption{Absorption Line Targets: Basic Data}
782: \tablewidth{0pt}
783: \tablehead{
784: \colhead{No.} & \colhead{Name} & \colhead{\it{l}} & \colhead{\it{b\/}}
785: & \colhead{z} & \colhead{N$_{H}$} & \colhead{t$_{exp}$\tablenotemark{a}} \\
786: \colhead{} & \colhead{} & \colhead{(deg)} & \colhead{(deg)} &
787: \colhead{} & \colhead{(10$^{20}$cm$^{-2}$}) & \colhead{(ksec)}
788: }
789: \startdata
790: 1 & Mkn 421 & 179.83 & 65.03 & 0.0300 & 1.38 & 509 \\
791: 2 & PKS 2155-304 & 17.73 & -52.24 & 0.116 & 1.71 & 723 \\
792: 3 & 3C 273 & 289.95 & 64.36 & 0.1583 & 1.79 & 451 \\
793: 4 & MCG-6-30-15 & 313.29 & 27.68 & 0.0077 & 4.06 & 460 \\
794: 5 & LMC X-3 & 273.57 & -32.08 & 0.000927 & 4.74 & 126 \\
795: 6 & 1H1426+428 & 77.49 & 64.90 & 0.129 & 1.38 & 388 \\
796: 7 & Akn 564 & 92.14 & -25.34 & 0.0247 & 6.27 & 130 \\
797: 8 & NGC 4051 & 148.88 & 70.09 & 0.0023 & 1.32 & 174 \\
798: 9 & NGC 3783 & 287.46 & 22.95 & 0.0097 & 8.26 & 316 \\
799: 10 & NGC 5548 & 31.96 & 70.50 & 0.0172 & 1.69 & 160 \\
800: 11 & Akn120 & 201.69 & -21.13 & 0.0327 & 12.2 & 112 \\
801: 12 & PKS 0558-504 & 257.96 & -28.57 & 0.137 & 4.39 & 168 \\
802: 13 & Mkn766 & 190.68 & 82.27 & 0.0129 & 1.71 & 189 \\
803: 14 & NGC 4593 & 297.48 & 57.40 & 0.009 & 2.31 & 115 \\
804: 15 & 3C 390.3 & 111.44 & 27.07 & 0.0561 & 4.24 & 123 \\
805: 16 & NGC 7469 & 83.10 & -45.47 & 0.0163 & 4.87 & 189 \\
806: 17 & Mkn 509 & 35.97 & -29.86 & 0.0344 & 4.11 & 75 \\
807: 18 & 3C 120 & 190.37 & -27.40 & 0.033 & 11.1 & 129 \\
808: 19 & NGC 3516 & 133.24 & 42.40 & 0.0088 & 3.05 & 260 \\
809: 20 & Ton 1388 & 223.36 & 68.21 & 0.1765 & 1.28 & 110 \\
810: 21 & 1H 0414+009 & 191.82 & -33.16 & 0.287 & 10.3 & 79 \\
811: 22 & Mr2251-178 & 46.20 & -61.33 & 0.064 & 2.70 & 65 \\
812: 23 & IC 4329a & 317.5 & 30.92 & 0.0161 & 4.42 & 136 \\
813: 24 & Fairall 9 & 295.07 & -57.83 & 0.047 & 3.19 & 33 \\
814: 25 & MS0737.9+7441 & 140.27 & 29.57 & 0.315 & 3.54 & 86 \\
815: 26 & 3C 59 & 142.04 & -30.54 & 0.1096 & 5.84 & 82 \\
816: \enddata
817: %
818: \tablenotetext{a}{Cleaned exposure time for the {\it XMM RGS}.}
819: \end{deluxetable}
820: \clearpage
821:
822: \begin{deluxetable}{rlrrrrrrrrrrrr}
823: \tabletypesize{\scriptsize}
824: \rotate
825: \tablecaption{OVII Absorption Line Measurements}
826: \tablewidth{0pt}
827: \tablehead{
828: \colhead{No.} & \colhead{Name} & \colhead{AngleGC} & \colhead{AngleM31}
829: & \colhead{R12} & \colhead{R45} & \colhead{R67} & \colhead{F$_{\lambda}$}
830: & \colhead{S/N} & \colhead{${\lambda}_c$} & \colhead{2$\sigma$} & \colhead{EW}
831: & \colhead{err} & \colhead{S/N} \\
832: \colhead{} & \colhead{} & \colhead{(deg)} & \colhead{(deg)} & \colhead{} & \colhead{}
833: & \colhead{} & \colhead{}
834: & \colhead{} & \colhead{(\AA )} & \colhead{(m\AA )}
835: & \colhead{(m\AA )} & \colhead{(m\AA )} & \colhead{}
836: }
837: \startdata
838: 1 & Mkn 421 & 114.97 & 97.42 & 1087 & 105 & 116 & 1.24E-02 & 48.9 & 21.602 & 0.069 & 11.8 & 0.8 & 14.1 \\
839: 2 & PKS 2155-304 & 54.32 & 80.89 & 964 & 145 & 135 & 2.88E-03 & 26.5 & 21.605 & 0.081 & 13.7 & 1.9 & 7.4 \\
840: 3 & 3C 273 & 81.51 & 136.57 & 1374 & 189 & 132 & 1.43E-03 & 14.7 & 21.608 & 0.083 & 24.6 & 3.3 & 7.4 \\
841: 4 & MCG-6-30-15 & 52.61 & 167.41 & 826 & 304 & 189 & 9.42E-04 & 11.2 & 21.617 & 0.081 & 32.6 & 6.8 & 4.8 \\
842: 5 & LMC X-3 & 86.97 & 120.20 & 533 & 160 & 156 & 9.00E-03 & 8.9 & 21.592 & 0.063 & 21.0 & 5.0 & 4.2 \\
843: 6 & 1H1426+428 & 84.73 & 92.73 & 1084 & 124 & 114 & 7.74E-04 & 8.0 & 21.603 & 0.08 & 11.6 & 4.1 & 2.8 \\
844: 7 & Akn 564 & 91.93 & 26.84 & 429 & 115 & 115 & 1.85E-03 & 7.4 & 21.603 & 0.08 & 12.3 & 4.6 & 2.7 \\
845: 8 & NGC 4051 & 106.95 & 93.74 & 1118 & 115 & 118 & 7.20E-04 & 7.4 & 21.603 & 0.05 & 24.6 & 3.1 & 7.9 \\
846: 9 & NGC 3783 & 73.97 & 167.24 & 510 & 143 & 118 & 3.52E-04 & 5.9 & 21.603 & 0.08 & 24.1 & 7.6 & 3.2 \\
847: 10 & NGC 5548 & 73.54 & 110.02 & 1052 & 152 & 139 & 7.79E-04 & 5.7 & 21.603 & 0.08 & 7.0 & 6.8 & 1.0 \\
848: 11 & Akn120 & 150.08 & 74.02 & 444 & 141 & 155 & 7.03E-04 & 5.6 & 21.603 & 0.08 & -6.0 & 5.5 & -1.1 \\
849: 12 & PKS 0558-504 & 100.56 & 114.80 & 500 & 122 & 120 & 6.89E-04 & 5.1 & 21.613 & 0.05 & 21.7 & 7.8 & 2.8 \\
850: 13 & Mkn766 & 97.60 & 108.70 & 887 & 112 & 112 & 8.33E-04 & 5.1 & 21.603 & 0.08 & 0.2 & 6.8 & 0.0 \\
851: 14 & NGC 4593 & 75.61 & 144.07 & 861 & 201 & 153 & 9.46E-04 & 5.0 & 21.603 & 0.08 & 23.4 & 8.5 & 2.8 \\
852: 15 & 3C 390.3 & 108.99 & 49.55 & 512 & 138 & 125 & 4.24E-04 & 4.3 & 21.603 & 0.08 & 27.4 & 7.3 & 3.8 \\
853: 16 & NGC 7469 & 85.17 & 39.15 & 406 & 100 & 117 & 6.96E-04 & 4.2 & 21.603 & 0.08 & 1.6 & 8.9 & 0.2 \\
854: 17 & Mkn 509 & 45.42 & 75.49 & 743 & 220 & 147 & 6.23E-04 & 4.0 & 21.603 & 0.08 & 25.9 & 7.3 & 3.6 \\
855: 18 & 3C 120 & 150.85 & 62.46 & 466 & 111 & 117 & 2.81E-04 & 3.5 & 21.603 & 0.08 & 13.8 & 9.2 & 1.5 \\
856: 19 & NGC 3516 & 120.39 & 64.94 & 831 & 113 & 118 & 9.04E-05 & 2.5 & 21.603 & 0.08 & 22.0 & 13.4 & 1.6 \\
857: 20 & Ton 1388 & 105.66 & 114.48 & 1004 & 121 & 128 & 1.41E-04 & 2.3 & 21.603 & 0.08 & 34.5 & 15.7 & 2.2 \\
858: 21 & 1H 0414+009 & 145.03 & 62.68 & 612 & 141 & 144 & 1.66E-04 & 2.2 & 21.603 & 0.08 & -3.1 & 14.8 & -0.2 \\
859: 22 & Mr2251-178 & 70.60 & 64.00 & 646 & 137 & 126 & 1.37E-04 & 1.8 & 21.603 & 0.08 & 39.8 & 19.6 & 2.0 \\
860: 23 & IC 4329a & 50.77 & 162.66 & 669 & 248 & 155 & 1.85E-04 & 1.7 & 21.603 & 0.08 & 33.8 & 19.3 & 1.8 \\
861: 24 & Fairall 9 & 76.96 & 100.44 & 893 & 122 & 124 & 2.17E-04 & 1.7 & 21.603 & 0.08 & 31.1 & 16.5 & 1.9 \\
862: 25 & MS0737.9+7441 & 131.98 & 54.35 & 548 & 112 & 116 & 1.30E-04 & 1.6 & 21.603 & 0.08 & -13.9 & 20.7 & -0.7 \\
863: 26 & 3C 59 & 132.77 & 20.73 & 493 & 128 & 115 & 9.22E-05 & 1.5 & 21.603 & 0.08 & 60.9 & 19.2 & 3.2 \\
864: \enddata
865: %
866: \tablecomments{AngleGC is the angle from the Galactic Center; AngleM31 is the angle from M31;
867: R12, R45, and R67 are the {\it ROSAT} soft X-ray background values in units of
868: 10$^{-6}$ cnt s$^{-1}$ arcmin$^{-2}$; F$_{\lambda}$ is the continuum flux
869: at 21.60 \AA\ in units of cnt s$^{-1}$ {\AA }$^{-1}$ cm$^{-2}$, followed by its S/N;
870: ${\lambda}_c$ the the line center of the OVII absorption; $\sigma$ is
871: the Gaussian parameter describing the OVII line width;
872: and EW is the equivalent width of the line.
873: }
874: \end{deluxetable}
875:
876: \clearpage
877:
878: \begin{figure}
879: \plotone{f1.eps}
880: \caption{The {\it XMM} flux, corrected for instrumental features, at the
881: location of the \ion{O}{7} line at 21.60 \AA . There is an instrumental
882: feature near 21.82 \AA , which is not used in the fit; the fit is shown by the dotted
883: line. A feature near 22.0 \AA\ is from a Galactic \ion{O}{6} blend. The
884: objects are shown by decreasing S/N of the continuum. For MCG-6-30-15, the
885: location of the resonance and intercombination lines at the redshift of
886: the object is shown.}
887: \end{figure}
888:
889: \clearpage
890:
891: \begin{figure}
892: \plotone{f2.eps}
893: \caption{As above, where the objects are ordered by continuum S/N. The
894: location of the AGN \ion{O}{7} lines in NGC 4051, if present are shown. For this
895: source, the continuum is defined by the data blueward of the 21.60 \AA\ line.}
896: \end{figure}
897:
898: \clearpage
899:
900: \begin{figure}
901: \plotone{f3.eps}
902: \caption{As above, for targets 9-12. NGC 3783 has a possible P Cygni profile
903: in its redshifted resonance \ion{O}{7} line.}
904: \end{figure}
905:
906: \clearpage
907:
908: \begin{figure}
909: \plotone{f4.eps}
910: \caption{As above, for targets 13-16. }
911: \end{figure}
912:
913: \clearpage
914:
915: \begin{figure}
916: \plotone{f5.eps}
917: \caption{As above, for targets 17-20. For most of the analysis, we do not
918: use targets 19 and higher and there is a decrease in S/N such that the
919: uncertainty in the equivalent widths become too large. In NGC 3516, both
920: the redshifted resonance and intercombination lines are present in emission.}
921: \end{figure}
922:
923: \clearpage
924:
925: \begin{figure}
926: \plotone{f6.eps}
927: \caption{For these final six objects, the S/N is rather poor. The
928: faintest source, 3C59, which shows possible absorption, lies within
929: 130 kpc of M33. }
930: \end{figure}
931:
932: \clearpage
933:
934: \begin{figure}
935: \plotone{f7.eps} % Fig 7
936: \caption{The distribution of \ion{O}{7} equivalent widths as a function
937: of angle from M31. In the Local Group model, we would expect a dereasing
938: equivalent width with increasing angle, a prediction not verified. }
939: \end{figure}
940:
941: \clearpage
942:
943: \begin{figure}
944: \plotone{f8.eps}
945: \caption{The observed \ion{O}{7} equivalent widths as a function
946: of the column density of an ellipse of diffuse gas filling the Local Group
947: and oriented along the Milky Way - M31 axis. The data are in conflict with
948: the predicted distribution, shown as a dashed line. }
949: \end{figure}
950:
951: \clearpage
952:
953: \begin{figure}
954: \plotone{f9.eps}
955: \caption{The distribution in the sky of the X-ray absorption sources
956: (black circles are objects 1-18 and are labeled; open circles are objects 19-26)
957: is shown along with the Local Group galaxies (stars; black indicates HI or
958: H$_2$ gas was detected). Some of the prominent galaxies are labeled as are
959: any moderatly close galaxy-AGN pairs; the anti-M31 direction is labeled.}
960: \end{figure}
961:
962: \clearpage
963:
964: \begin{figure}
965: \plotone{f10.eps} % Fig 10
966: \caption{The \ion{O}{7} equivalent widths as a function of the angle from
967: the Galactic Center, which shows a trend that is consistent with the
968: Galactic Halo model. The dotted line is the best-fit, which is significant
969: at the 95\% level. }
970: \end{figure}
971:
972: \clearpage
973:
974: \begin{figure}
975: \plotone{f11.eps}
976: \caption{The \ion{O}{7} equivalent widths as a function of the column density
977: for a Galactic halo of uniform density and radius of 20 kpc as measured from
978: the Galactic Center. The model line passes through the origin and the fit
979: is significant at the 95\% level. }
980: \end{figure}
981:
982: \clearpage
983:
984: \begin{figure}
985: \plotone{f12.eps}
986: \caption{The \ion{O}{7} equivalent widths as a function of the column density
987: for a Galactic halo where the density n $\propto$ r$^{-3/2}$. The fit shown,
988: which is not required to go through the origin is significant at the 95\% level. }
989: \end{figure}
990:
991: \clearpage
992:
993: \begin{figure}
994: \plotone{f13.eps} % Fig 13
995: \caption{The \ion{O}{7} equivalent widths as a function of the R45 intensity in
996: an annulus around each target; the units for R45 are 10$^{-6}$ counts s$^{-1}$ arcmin$^{-2}$.
997: The fit (dashed line) is significant at the 98-99\% confidence level.}
998: \end{figure}
999:
1000: \clearpage
1001:
1002: \begin{figure}
1003: \plotone{f14.eps}
1004: \caption{As above, except that all 26 objects are used to make binned,
1005: error-weighted averages. We have incorporated an additional rms to the
1006: equivalent widths of 3 m\AA\ so that the $\chi$$^{{\rm 2}}$ of the error-weighted fit
1007: is in the acceptable range. The correlation is significant at the
1008: 99\% confidence level.}
1009: \end{figure}
1010:
1011: %\clearpage
1012:
1013: %\begin{figure}
1014: %\plotone{f1.eps}
1015: %\caption{ }
1016: %\end{figure}
1017:
1018:
1019: \end{document}
1020: