1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[letterpaper,10pt,twocolumn]{article}
3: \slugcomment{}
4: \usepackage{graphics,graphicx}
5: \setlength{\textwidth}{6.5in}
6: %\setlength{\textheight}{9in}
7: \setlength{\topmargin}{0.8in}
8: \setlength{\oddsidemargin}{0in}
9: \setlength{\evensidemargin}{0in}
10: %\setlength{\headheight}{0in}
11: \setlength{\headsep}{0in}
12: \setlength{\hoffset}{0in}
13: \setlength{\voffset}{0in}
14: %\renewcommand{\section}{\@startsection%
15: %{section}{1}{0mm}{-\baselineskip}%
16: %{0.5\baselineskip}{\normalfont\Large\bfseries}}%
17: \shorttitle{Redshifted Iron Lines}
18: \shortauthors{Turner, Kraemer, Reeves}
19:
20:
21: \begin{document}
22:
23: \title{
24: Transient Relativistically-Shifted Lines as a
25: Probe of Black Hole Systems }
26:
27: \author{ T.\ J.\ Turner\altaffilmark{1,2}, S.\ B.\ Kraemer\altaffilmark{3},
28: J.\ N.\ Reeves\altaffilmark{2} }
29:
30: \altaffiltext{1}{Joint Center for Astrophysics, Physics Dept., University
31: of Maryland Baltimore County, 1000 Hilltop Circle, Baltimore, MD 21250 }
32: \altaffiltext{2}{Laboratory for High Energy Astrophysics, Code 662,
33: NASA/GSFC, Greenbelt, MD 20771}
34: \altaffiltext{3}{Catholic University of America, NASA/GSFC, Code 681,
35: Greenbelt, MD 20771}
36:
37:
38: \begin{abstract}
39:
40: X-ray spectra of Seyfert galaxies
41: have revealed a new type of X-ray
42: spectral feature, one which appears to offer important new insight into the
43: black hole system.
44: {\it XMM} revealed
45: several narrow emission lines redward of Fe K$\alpha$ in
46: NGC~3516. Since that discovery the phenomenon has been observed in
47: other Seyfert galaxies, e.g. NGC 7314 and ESO 198-G24. We
48: present new evidence for a redshifted Fe line in {\it XMM} spectra
49: of Mrk~766. These data reveal the first evidence for a
50: significant shift in the energy of such a line,
51: occurring over a few tens of kiloseconds. This shift may be interpreted as
52: deceleration of an ejected blob of gas
53: traveling close to the escape velocity.
54:
55: \end{abstract}
56:
57: \keywords {galaxies: active -- galaxies: individual (Mrk 766) --
58: accretion: accretion disks }
59:
60:
61: \section{Introduction}
62:
63: Active Galactic Nuclei (AGN)
64: are believed to be powered by accretion of
65: material onto a black hole. UV photons from the disk are thought to
66: be upscattered by relativistic electrons, providing the hard X-ray continuum.
67: These hard X-rays illuminate the disk surface, undergoing photoelectric
68: absorption or Compton scattering. High abundance and fluorescence yield make
69: Fe K$\alpha$ the strongest line in
70: the X-ray band-pass, emitted via
71: fluorescence or recombination processes between 6.4 -- 7.0 keV,
72: depending on the
73: ionization-state of the gas. This line is commonly observed
74: in AGN \citep{n97} with both narrow
75: and broad components. The former has long been thought to be
76: dominated by contributions from cool material at or beyond
77: the optical broad-line-region
78: while the latter is thought to originate
79: close to the black hole. Lines emitted very close to the black hole
80: will show shifted, broadened and skewed profiles due to the combination
81: of Doppler and relativistic effects (see \citealt{fab2000} and
82: references therein). Infalling material which is not part of the
83: disk structure will
84: also suffer the effects of strong gravity, and may achieve
85: velocities which are a significant fraction of the speed of light, causing
86: significant displacement of
87: emission lines from their rest energy. Many astrophysical sources
88: show evidence for emission (SS~433, e.g. \citealt{mcs02}) and absorption from
89: material traveling at relativistic velocities (e.g. Broad Absorption Line
90: Quasars; \citealt{wey97}).
91:
92: Overlapping {\it Chandra} and {\it XMM-Newton}
93: observations from 2001 November revealed unexpected
94: narrow
95: emission lines at $\sim$ 5.6 and 6.2 keV in the Seyfert galaxy
96: NGC~3516 \citep{t02}.
97: Narrow emission lines, redshifted relative to Fe K$\alpha$ had never
98: before been seen in the X-ray spectrum of an AGN.
99: The high-throughput of {\it XMM} and the excellent energy-resolution
100: afforded by the {\it Chandra} HETG made possible the first unambiguous
101: detection of these weak lines.
102: With hindsight {\it ASCA} data show the appearance of narrow
103: iron lines to be a frequently
104: recurring phenomenon in NGC~3516 \citep{n99}
105: over a wide range in source luminosity.
106: \citet{t02} found variability in the flux of the new lines
107: on timescales of a few tens of ksec.
108: The apparent variability of the 5.6 keV component was most interesting
109: as it could be attributed either to the line flux varying or the
110: line energy shifting from 5.4 to 5.6 keV.
111: A similar line at 5.86 keV has now been detected in NGC~7314
112: \citep{y03} and at 5.7 keV in
113: ESO 198-G24 \citep{g03} -- the phenomenon
114: may be a characteristic of AGN.
115: Here we present data from an {\it XMM} observation of Mrk~766 which shows
116: characteristically similar redshifted emission to that observed in NGC~3516,
117: offering new insight into the origin
118: of these lines.
119:
120: \section{The case of Mrk 766}
121:
122: {\it XMM} has three CCD detectors (EPIC)
123: operating in the $\sim 0.3 - 10$ keV band. The collecting area
124: of the EPIC detector and telescope combination
125: is approximately an order of magnitude higher than the
126: CCDs flown on {\it ASCA}. The energy resolution is
127: $\Delta E \sim 120$ eV (FWHM) at 6.4 keV. This combination of capabilities
128: has opened up a new regime of Fe K$\alpha$
129: profile studies. High S/N spectra can be
130: accumulated in much shorter integration times than ever before, allowing
131: significant progress in studying variable spectral features and in
132: detection of detailed structure to Fe line profiles.
133: We analyzed the archival {\it XMM} data for Mrk 766.
134: Two {\it XMM} observations of Mrk 766 are available in the archives,
135: from 2000 May 20 and 2001 May 20-21. EPIC pn data were screened using
136: SAS v5.3 software to
137: select only events with patterns in the range 0 - 4. We also applied
138: energy cuts to discard data below 0.2 keV and above 15.0 keV.
139: The pn data yielded a useful exposure of 130 ksec for the 2001 observation and
140: 60 ksec for the 2000 data. The 2001
141: observation utilized the Small Window mode and there was no
142: region for extraction of MOS2 background spectra, unfortunately
143: this background level
144: becomes significant at and above the Fe K$\alpha$ emission line. The
145: 2001 MOS1 data were taken in timing mode, with the central chip
146: off. In timing mode, spectra must be extracted based upon the histogram
147: of counts versus RAWX coordinate. This allows the user to extract a spectrum
148: for a given range of RAWX. While the histogram allows selection of
149: source and background spectra, the source spectrum has to include
150: some background dominated channels in the compressed spatial dimension.
151: For weak sources such as Mrk~766 the resulting
152: background-subtracted spectrum is
153: somewhat inaccurate, especially at high energies. We extracted the
154: timing-mode spectra for MOS1 and confirmed that those data were
155: consistent with our PN results, but we do not present those data here.
156:
157: In the 2000 observation pileup hardened the MOS spectra but did not affect
158: the pn data. Thus the MOS data were not used from either observation.
159: Source spectra were extracted from both epochs using circular extraction
160: cells of $\sim 45''$ radii centered on Mrk 766. Background spectra were
161: taken from an offset position close to the source.
162: Source spectra were grouped to
163: have at least 20 counts per spectral bin for fitting using
164: the $\chi^2$ statistic. Response matrices were generated for
165: the source spectra using SASv5.3.
166:
167: \begin{figure}[!h]
168: \includegraphics[scale=0.45,angle=0]{f1.eps}
169: \caption{\it Mrk 766: Light curve from the pn data taken during
170: 2001 May 20-21, accumulated from the source extraction cell
171: in 200 second bins. The high-state corresponds to data from
172: the first 100,000 seconds, and the dip refers to the last 30 ks of
173: the observation where the source flux drops notably.
174: }
175: \end{figure}
176:
177: Figure 1 shows the light curve from 2001 May, as first presented by
178: \citet{p03}. We confirm the general findings of \citet{p03} regarding
179: the mean spectrum, i.e.
180: that the spectrum is steep
181: $\Gamma \sim 2.2$ and
182: shows a broad component of Fe K$\alpha$ at rest-energy
183: $\sim 6.7$ keV, with a narrower component at 6.4 keV.
184: Our initial spectral fits
185: parameterized the hard X-ray continuum
186: in the 3-11.0 keV range using a powerlaw continuum form,
187: but ignoring data in the
188: 5-7 keV band where the Fe K$\alpha$ line resides. We assumed
189: a redshift z=0.012929 for Mrk~766 \citep{mcm}.
190:
191: \begin{figure}[!h]
192: \includegraphics[scale=0.45,angle=0]{f2.eps}
193: \caption{\it Data/model of Fe K$\alpha$ emission in Mrk 766, where
194: the model is the local powerlaw continuum, as detailed by \citet{p03}.
195: The green line represents the high-state and the red line data from the dip. }
196: \end{figure}
197:
198: We also performed time-resolved
199: spectroscopy, accumulating spectra
200: before (high-state) and after (the dip) a drop in source flux (Fig.~1).
201: We adopted the general model of a powerlaw with
202: broad and narrow components to model the complex
203: Fe K$\alpha$ line which is evident in the 6.4 -- 7.0 keV
204: band. In addition,
205: we find a significant feature at 5.60 keV, not discussed by
206: \citet{p03}.
207:
208: Figure 2 shows the data/model line ratios
209: for the two sections of data. An emission line is detected at 5.60 keV
210: in the high-state spectrum and at 5.75 keV in the dip-state.
211: Both lines are detected at $ > 99\%$ confidence.
212: The lines at 5.60 and 5.75 keV had photon fluxes
213: $n=0.50^{+0.26}_{-0.26} \times 10^{-5}$ photons cm$^{-2}$ s$^{-1}$
214: and
215: $n=1.04^{+0.42}_{-0.42} \times 10^{-5}$ photons cm$^{-2}$ s$^{-1}$,
216: respectively (errors represent 90\% confidence). Subdivision of
217: the high-state data into six time-resolved spectra of equal duration
218: and the low-state into two halves
219: revealed no other significant changes in line energy within each
220: flux-state.
221:
222: We also
223: compared the spectra of these two `states'
224: with that from an {\it XMM} observation 1 year earlier. Figure 3
225: shows the unfolded spectra from these three
226: data sections. The apparent energy shift of the 5.60 keV feature
227: is again evident by comparison of the top (green) and center
228: (red triangles) lines. For comparison we
229: include the 2000 May data (bottom line, black crosses), catching
230: the source at a flux state
231: similar to that denoted the ``dip'' in 2001 May data. Those earlier data
232: do not show a significant feature in the range 5.60 - 5.75 keV.
233: The 2000 May spectrum yields an
234: upper limit on the flux of a line at 5.75 keV
235: $n < 0.2 \times 10^{-5}$ photons cm$^{-2}$ s$^{-1}$. Thus the source is
236: inconsistent with the presence of the same line in both 2000 May
237: and 2001 May spectra
238: accumulated during the low flux-state.
239:
240: \begin{figure}[!h]
241: \includegraphics[scale=0.45,angle=0]{f3.eps}
242: \caption{\it Unfolded spectra from two XMM observations of Mrk 766.
243: The top line (green) shows data from the high-state during the
244: long observation of 2001 May 20-21. The red triangles
245: show data from the dip-state of the same observation.
246: A shift is evident in one line from
247: approximately 5.60 keV to 5.75 keV within a few tens of ksec.
248: The black crosses
249: show data from the short observation of 2000 May 2000.
250: }
251: \end{figure}
252:
253:
254: To further investigate the narrow redshifted lines
255: we examined the significance of
256: features in the Fe K$\alpha$ regime. We slid a Gaussian model template
257: across data between 3.0 - 11.0 keV, testing for an improvement to the fit
258: at every resolution element, compared to a powerlaw model for the
259: underlying continuum. The background spectrum was subtracted before
260: the test, so features in Figure 4 are attributable to Mrk 766.
261:
262: \begin{figure}[!h]
263: \includegraphics[scale=0.45,angle=0]{f4.eps}
264: \caption{\it Mrk~766 -
265: Change in fit statistic versus energy for a narrow Gaussian model line
266: which was stepped across the data. The comparison model was
267: a simple absorbed powerlaw continuum, with parameters allowed to
268: float freely. The green line represents the high-state
269: data from 2001 May, the red line is data from the dip and the black line is
270: data from the earlier observation in 2000 May. The strongest Fe line component is also evident at 6.4 keV, the Y-axis is truncated to
271: emphasize the weaker line. }
272: \end{figure}
273:
274: In the resulting fit-statistic, $\Delta \chi^2$ of 9.21 represents
275: a feature which is significant at 99\% confidence. Fig.~4
276: illustrates
277: the newly discovered lines and also traces out the stronger Fe lines
278: reported by \citet{p03}.
279: The green line represents the 2001 high-state, the red line 2001 dip data
280: and the black line the 2000 May observation.
281: Fig. 4 provides
282: supporting evidence that the peak
283: of the emission
284: shifts from 5.60 to 5.75 keV in the rest-frame, and that there is no
285: significant feature in the 2000 May spectrum.
286: Figure 5
287: shows the contours for the energy and intensity of this line
288: in the high and dip states.
289: An apparent shift in line energy is significant at 95\% confidence (while
290: the apparent change in line intensity turns out to be insignificant).
291: Our analysis of a 90 ks {\it Chandra} HEG observation of Mrk~766 from another
292: epoch did not show
293: significant redshifted, narrow Fe lines.
294:
295: \begin{figure}[!h]
296: \includegraphics[scale=0.40,angle=0]{f5.eps}
297: \caption{\it Contours showing the 90\% (black) and 95\% (red) confidence
298: levels for values of energy and intensity for the new line.
299: The left-hand contours are from the high-state and the right-hand contours
300: are from the dip, both from 2001 May.}
301: \end{figure}
302:
303:
304:
305: \section{Discussion}
306:
307: Several cases of narrow and redshifted lines
308: have now been found in Seyfert galaxies and
309: these features may be a characteristic
310: of black hole systems, currently detectable
311: only in the relatively bright Seyfert type AGN.
312:
313: Spectra of Mrk~766 show a narrow, redshifted Fe line which
314: changes from 5.60 to 5.75 keV within a few tens of kiloseconds.
315: These data provide the first ever observation of a significant shift
316: in energy for a line of this type.
317: We dismiss the possibility that the observed change in energy is due to
318: an increase in ionization-state of the emitter. In that picture an
319: increase in line energy should be correlated with an increase,
320: rather than a decrease in central luminosity.
321:
322: Spallation, which is the breakdown of Fe atoms into lower Z metals
323: by energetic protons,
324: has been disfavored as an explanation of the lines
325: in NGC~3516 \citep{t02} on the basis of line ratios and rapid
326: line variability. The energy shift detected in the line from Mrk~766
327: seems inconsistent with the spallation scenario, pushing it further from favor.
328:
329: In the case of NGC 3516 \citet{t02} discuss an origin
330: from narrow radii which are sporadically illuminated
331: on the surface of the accretion disk. The
332: ``Thundercloud'' model may be applicable, where the
333: accretion disk is illuminated in patches from magnetic reconnection events
334: \citep{mf01}.
335: If the redshifted lines observed in Mrk~766 are red horns
336: of disklines then peak energies in the range 5.60-5.75 keV
337: would occur from a radius $R \sim 30$ gravitational radii.
338: For a disk inclined at
339: 30 degrees to our line-of-sight there should be an associated
340: blue horn at $\sim 6.7$ keV, which would be $\sim 70\%$
341: stronger than the red horn due to relativistic effects
342: (\citealt{schwarz}, \citealt{laor}).
343: This is true for both the Schwarzschild and
344: Kerr metrics, which produce similar lines at such radii.
345: If the disk is inclined at 60 degrees then the blue horn would be
346: expected at 7.25 keV with twice the strength of the red horn. The data
347: are inconsistent with the latter line.
348: Without an independent constraint on disk inclination
349: it is difficult to make a conclusive comment on the consistency of the data
350: with a blue horn and thus with the disk model.
351: For the case of a disk inclined close to face-on
352: the blue horn would be lost in the broad line component noted by
353: \citet{p03}, or in the wings of the strong
354: Fe K$\alpha$ component at 6.4 keV.
355: A disk origin remains an interesting possibility.
356: Shifts in line energy might be expected
357: to be correlated with flux as new magnetic reconnections would likely
358: light up different parts of the disk.
359: Alternatively, the
360: observation of a rapid energy shift for the emission line in Mrk~766
361: is suggestive of
362: a decelerating blob of gas, and we now explore this possibility
363: in detail.
364:
365:
366: \subsection{Outflow?}
367:
368: That the line is redshifted could indicate either an origin in
369: material expelled from the nucleus, in which case we are seeing
370: material on the far side of the black hole, or material falling into
371: the nucleus, where we are viewing gas on the near side of the black hole.
372: The observation that the
373: line is getting less redshifted with time is indicative of
374: deceleration of the emitter. Deceleration might be expected as expelled
375: material would be likely slowed by gravity.
376: While there may exist deceleration mechanisms
377: which can slow infalling material (e.g. magnetic field pressure or
378: drag forces) the outflow scenario
379: seems a likely model
380: and so we restrict ourselves to a discussion of this possibility.
381: NGC 3516 may show a similar deceleration, as there is some suggestion of
382: a line which moves from 5.4 to 5.6 keV in a few tens of kiloseconds
383: \citep{t02}.
384: Understanding gas inflow/outflow is to get insight into
385: the accretion/expulsion processes, two of the most fundamental questions
386: regarding black hole systems.
387: Disk winds are a natural consequence of the
388: accretion process, and so it is particularly compelling to look at these
389: new observations in light of such models.
390:
391: One can use the observed shift in line energy to constrain the parameters of
392: gas ejected ballistically from the disk surface.
393: Consider a blob of gas propelled outward just above the plane of the disk.
394: The propulsion could be driven in an number of ways, magnetically,
395: or by the pressure gradient which must exist in the radial direction.
396: The material is expelled and then decelerates under gravity.
397:
398: If the material is decelerating under gravity then
399: $\Delta V/\Delta t = GM/R^2$, where $\Delta t$ is the time under which
400: the deceleration was observed, and $R$ is the radius at which the material
401: is situated at that time. Assuming the line
402: originates from neutral iron then the implied initial velocity of the
403: emitting blob is $38000$ km/s for the
404: line at 5.60 keV in Mrk 766.
405: The observed change in velocity
406: is $\Delta V \sim 7000$ km/s in $\Delta t \sim 65000 s$.
407:
408: Taking an estimate of the mass of the black hole to be
409: $10^7 M_{\odot}$ for Mrk 766 \citep{w02}
410: yields an estimate of the radius to be
411: $R \sim 3.5 \times 10^{14}$ cm. The escape velocity at this
412: radius is 28000 km/s,
413: close to the (ejection) velocity observed. This ties in
414: nicely with the suggested outflow picture as blobs would have to achieve the
415: escape velocity in order to be ejected and blobs not achieving this
416: velocity would fall back to the disk. Application of the
417: same arguments to the tentative energy shift in NGC 3516
418: yields a velocity of 47000 km/s, with $\Delta V \sim 9400$ km/s
419: compared to an escape velocity of 41000 km/s at approximately the
420: same radius (and central mass $\sim 2.3 \times 10^7 M_{\odot}$ ).
421:
422: Interestingly, \citet{kp03}
423: note that the outflow velocity for the X-ray absorber in PG 1211$+143$
424: is close to the escape velocity for that inferred radius, suggesting
425: these are related phenomena, where viewing angle determines whether one
426: sees the ejected gas in emission or absorption. We also note
427: evidence for outflow of gas with velocity $\sim 50,000$ km/s in
428: PDS~456, where \citet{reeves} estimated a launch radius of $\sim 40$
429: gravitational radii
430: for the absorbing gas, if the outflow velocity is approximately equal to the
431: escape velocity.
432:
433: The apparent link between the energy of the redshifted line, and the source
434: flux-state during the 2001 observation
435: is intriguing. However, the lack of a comparable feature in the earlier
436: observation (2000 May) indicates there is no long-term relation
437: between the line and
438: source luminosity. One possible explanation
439: is that the emitting cloud responds to
440: luminosity changes on short timescales, but that blobs do not exist on
441: long timescales. This conjures a slightly more complex picture
442: than noted above. The blob may be driven outwards
443: by radiation pressure and maintain a constant velocity so long as
444: that pressure is balanced by gravity.
445: When the radiation field drops to half the
446: previous value we obtain $\Delta V/\Delta t = GM/2R^2$ yielding
447: $R \sim 1.75 \times 10^{14}$ cm, and an escape velocity of 40,000 km/s.
448: In this picture the blob is initially
449: traveling outwards at a constant velocity
450: approximately equal to the escape velocity. The source flux then
451: drops by a factor 2 and the blob decelerates, dropping
452: below the escape velocity. The evolution of the blob then depends on how the
453: central luminosity continues to vary. If the source does not
454: brighten again, the blob will fall back onto the black hole.
455: This picture is appealing, but does not explain why the flux changes
456: {\it within} the high-state did not show significant line shifts,
457: leaving it unclear whether this picture is indeed more compelling
458: than the simpler case of a blob driven magnetically or by pressure
459: gradients.
460: It is clear from the above estimates that
461: observations with a long baseline which allow us to monitor these lines
462: across several large changes in central luminosity could allow us
463: to distinguish between radiative and other mechanisms driving the
464: ejecta.
465:
466: While an ejection model might be expected to yield both blue and
467: redshifted emission lines, assuming gas is symmetrically ejected
468: towards and away from us, the blue lines will be harder to observe
469: in available data. The blueshifted lines from a blob on the near-side
470: of the black hole, traveling outward at the same velocity, would
471: be observed at $\sim 7.3$ keV. Interestingly the data
472: show a couple of high points at 7.3 keV (Figs.~2, 3) however
473: there are no significant lines present (Fig.~4).
474: Upper limits on the flux of a narrow line
475: are $n< 9.0 \times 10^{-7}$ photons cm$^{-2}$ s$^{-1}$
476: ($EW < 5$ eV) and $n< 3.5 \times 10^{-6}$ photons cm$^{-2}$ s$^{-1}$
477: ($EW < 28$ eV) for the high and low states, respectively.
478: The emitting gas is likely to have a column
479: $\sim 10^{24} {\rm cm^{-2}}$ \citep{fgf}
480: and hence will absorb much of the fluorescent line emitted on the
481: illuminated side of the blobs on the near side of the hole, whereas we
482: view the illuminated face of the receding blobs directly.
483: This makes it difficult to rule out symmetric outflow.
484: Interestingly,
485: lines have been observed blueward of 6.4 keV in sources such as
486: NGC 7314 \citep{y03}.
487: In that case there was ambiguity as to whether the
488: lines were redshifted
489: H-like and He-like species of Fe or blueshifted lines from
490: neutral Fe.
491:
492: In Mrk 766 the line at $\sim 5.60$ keV has mean equivalent
493: width $15^{+6}_{-5}$ eV. The strongest component
494: of Fe K$\alpha$, at 6.4 keV, likely has
495: contributions from the optical broad line clouds
496: with $\Omega/2\pi \sim
497: 0.05-0.3$ in Mrk~766 \citep{nl93}. Adding
498: contributions from the narrow-line region and torus yields
499: $\Omega/2\pi \sim 0.4$. The
500: ratio of strengths of the components at 6.4 and 5.6 keV
501: is $\sim 2.5$.
502: Thus the redshifted line likely has
503: covering fraction $\Omega/2\pi \sim 0.15$.
504: Taking our derived radii and column densities, a mass
505: $\sim 10^{-4} M_{\odot}$ is lost in the blob.
506: The kinetic energy
507: in the outflow is approximately $10^{48} {\rm ergs}$, a few
508: percent of the gravitational potential
509: energy available in the disk at these radii.
510:
511: \citet{p03} found a similar covering factor for the variable warm absorber
512: in Mrk~766, and suggested that gas could contribute to the He-like
513: Fe K$\alpha$ emission. We prefer a neutral origin for our emission line
514: (if the line originated in He-like Fe then we would expect some H-like
515: Fe and so would expect a pair of redshifted lines). However, in
516: general the \cite{p03} picture of flare ejecta is consistent
517: with the new result presented here. The variable absorption may
518: be seen in the component of gas on the near side of the flow, and emission
519: is observed from the component on the far side.
520:
521: \citet{lev} discuss line photon propagation in bulk outflows from AGN
522: where Thomson scattering in the divergent flow redshifts photons to
523: produce the broad red wing observed on the line profile of most AGN.
524: Similar arguments could be applied to produce the narrow redshifted lines
525: observed in NGC 3516, Mrk 766 and other AGN, although in the latter case
526: the rapid shift
527: in line energy may require a more complex explanation than the
528: decelerating emitter, suggested above.
529:
530: \section{Conclusions}
531:
532:
533: Data from several Seyfert galaxies show
534: narrow, redshifted emission most likely
535: explained as Fe K$\alpha$ lines, shifted by relativistic effects.
536: Mrk~766 provides the first detection of a significant shift in the energy of
537: such a line.
538: The timescale of the energy shift is
539: a few tens of kiloseconds, indicating a possible origin in
540: blobs of gas expelled from the nucleus and then
541: gravitationally decelerated.
542: The nature of the ejection mechanism is unknown at this time, but
543: this may be a newly detectable signature of black hole systems.
544:
545: \section{Acknowledgments}
546:
547: We thank Ian George and the anonymous referee for useful
548: comments which improved the paper.
549: T.J.\ Turner acknowledges support from NASA
550: grant NAG5-7538.
551:
552:
553: \clearpage
554: \begin{thebibliography}{}
555:
556:
557: \bibitem[Fabian et al.(1989)]{schwarz}
558: Fabian, A.C., Rees, M.J., Stella, L, White, N.E.,
559: 1989, \mnras, 238, 729 \\
560: \bibitem[Fabian et al.(2000)]{fab2000}
561: Fabian, A.C., Iawsawa, K., Reynolds, C.S., Young, A.J.,
562: 2000, PASP, 112, 1145 \\
563: \bibitem[George \& Fabian(1991)]{fgf}
564: George, I. M., Fabian, A. C., 1991, \mnras, 249, 352 \\
565: \bibitem[Guainazzi(2003)]{g03}
566: Guainazzi, M., 2003, A\&A 401, 903 \\
567: \bibitem[Laor(1991)]{laor}
568: Laor, A., 1991, \apj, 376, 90 \\
569: \bibitem[King \& Pounds(2003)]{kp03}
570: King, A.R. \& Pounds, K.A., 2003, \mnras, in press \\
571: \bibitem[Marshall, Canizares \& Schulz(2002)]{mcs02}
572: Marshall, H.L., Canizares, C.R., \& Schulz, N.S., 2002,
573: \apj, 564, 941 \\
574: \bibitem[McMahon et al.(2002)]{mcm}
575: McMahon, R. G., White, R., L., Helfand, D. J., Becker, R. H.,
576: 2002, \apjs, 143, 1 \\
577: \bibitem[Merloni \& Fabian(2001)]{mf01}
578: Merloni, A. \& Fabian, A.C., 2001, \mnras, 328, 958 \\
579: \bibitem[Miller et al.(2002)]{m02}
580: Miller, J.M. et al. 2002, \apj, 578, 348 \\
581: \bibitem[Nandra et al.(1997)]{n97}
582: Nandra, K. George, I.M., Mushotzky, R.F.,
583: Turner, T.J., Yaqoob, T. 1997,
584: ApJ, 477, 602 \\
585: \bibitem[Nandra et al.(1999)]{n99}
586: Nandra, K., George, I.M., Mushotzky, R.F., Turner, T.J., Yaqoob, T.
587: 1999, \apj, 523, 17 \\
588: \bibitem[Netzer \& Laor(1993)]{nl93}
589: Netzer, H., Laor, A., 1993, \apjl, 451, 51 \\
590: \bibitem[Pounds et al.(2003)]{p03}
591: Pounds, K.A., Reeves, J.N., Page, K.L., Wynn, G.A.,
592: O'Brien, P.T., 2003, \mnras, 342, 1147 \\
593: \bibitem[Reeves, O'Brien \& Ward(2003)]{reeves}
594: Reeves, J. N., O'Brien, P. T., Ward, M. J., 2003, astro-ph/00307127 \\
595: \bibitem[Titarchuk, Kazanas \& Becker(2003)]{lev}
596: Titarchuk, L., Kazanas, D., Becker, P. A., 2003, astro-ph/0307486 \\
597: \bibitem[Turner et al.(2002)]{t02}
598: Turner, T.J., et al. 2002, ApJL 574, 123 \\
599: \bibitem[Wandel(2002)]{w02}
600: Wandel, A., 2002, ApJ, 565, 762 \\
601: \bibitem[Yaqoob et al.(2003)]{y03}
602: Yaqoob, T, George, I.M., Kallman, T.R., Padmanabhan, U.,
603: Weaver, K.A., Turner, T.J., 2003, submitted to ApJ \\
604: \bibitem[Weyman(1997)]{wey97}
605: Weyman, R.J. 1997
606:
607: \end{thebibliography}
608:
609: \end{document}
610: