1:
2: \documentclass[12pt,preprint]{aastex}
3: \usepackage{emulateapj5}
4:
5: %\lefthead{Dewangan et al.}
6: %\righthead{Testing truncated accretion disk models with Iron K$\alpha$ line}
7: \slugcomment{To appear in ApJ}
8:
9:
10:
11: \def\mes{M{\'e}sz{\'a}ros}
12: \newcommand\asca{{\it ASCA}}
13: \newcommand\sax{{\it BeppoSAX}}
14: \newcommand\chandra{{\it Chandra}}
15: \newcommand\rosat{{\it ROSAT}}
16: \newcommand\rxte{{\it RXTE}}
17: \newcommand\xmm{{\it XMM-Newton}}
18: \newcommand\ftools{{\it FTOOLS}}
19: \newcommand\s{{\rm~s}}
20: \newcommand\kev{{\rm~keV}}
21: \newcommand\ev{{\rm~eV}}
22: \newcommand\kms{\ifmmode {\rm~km\ s}^{-1} \else ~km s$^{-1}$\fi}
23: \newcommand\Hunit{\ifmmode {\rm~km\ s}^{-1}\ {\rm Mpc}^{-1}
24: \else ~km s$^{-1}$ Mpc$^{-1}$\fi}
25: \newcommand\ctssec{\ifmmode {\rm~count\ s}^{-1} \else ~count s$^{-1}$\fi}
26: \newcommand\ergsec{\ifmmode {\rm~erg\ s}^{-1} \else
27: ~erg s$^{-1}$\fi}
28: \newcommand\funit{\ifmmode {\rm~erg\ s}^{-1}\;{\rm cm}^{-2} \else
29: ~ergs s$^{-1}$ cm$^{-2}$\fi}
30: \newcommand\phflux{\ifmmode {\rm~photon\ s}^{-1}\;{\rm cm}^{-2}
31: \else ~photon s$^{-1}$ cm$^{-2}$\fi}
32: \newcommand\efluxA{\ifmmode {\rm~erg\ s}^{-1}\;{\rm cm}^{-2}\;{\rm
33: \AA}^{-1} \else ~erg s$^{-1}$ cm$^{-2}$ \AA$^{-1}$\fi}
34: \newcommand\efluxHz{\ifmmode {\rm~erg\ s}^{-1}\;{\rm cm}^{-2}\;{\rm
35: Hz}^{-1} \else ~erg s$^{-1}$ cm$^{-2}$ Hz$^{-1}$\fi}
36: \newcommand\cc{\ifmmode {\rm~cm}^{-3} \else cm$^{-3}$\fi}
37: \newcommand\FWHM{\ifmmode {\rm~FWHM} \else ${\rm~FWHM}$\fi}
38: \newcommand\Msun{\ifmmode M_{\odot} \else $M_{\odot}$\fi}
39: \newcommand\Lsun{\ifmmode L_{\odot} \else $L_{\odot}$\fi}
40: \newcommand\ltsim{\raisebox{-.5ex}{$\;\stackrel{<}{\sim}\;$}}
41: \newcommand\gtsim{\raisebox{-.5ex}{$\;\stackrel{>}{\sim}\;$}}
42: \newcommand\hbeta{\ifmmode {\rm H}\beta \else H$\beta$\fi}
43: \newcommand\Kalpha{\ifmmode {\rm K}\alpha \else K$\alpha$\fi}
44: \newcommand\NH{\ifmmode N_{\rm H} \else N$_{\rm H}$\fi}
45: \usepackage{graphicx}
46: \usepackage{here}
47:
48:
49: %\received{2002 August 28}
50: %\received{2002 August 28}
51:
52: \begin{document}
53:
54:
55: \title{Iron K$\alpha$ emission from the low-luminosity Active Galaxies M81 and NGC 4579}
56:
57: \author{G. C. Dewangan\altaffilmark{1,2}, R. E. Griffiths\altaffilmark{1}, T. Di Matteo\altaffilmark{3}, \& N. J. Schurch\altaffilmark{1}}
58: \altaffiltext{1}{Department of Physics, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA 15213 USA}
59: \altaffiltext{2}{email: gulabd@cmu.edu}
60: \altaffiltext{3}{Max-Planck-Institut f{\" u}r Astrophysik, Karl-Schwarzschild-St
61: r.~1, 85740 Garching bei M{\" u}nchen, Germany}
62:
63: \begin{abstract}
64: We report on \xmm{} spectroscopy of the
65: low-luminosity active galaxies (LLAGN) M81 and NGC~4579 both of
66: which have known black hole masses and well-sampled spectral energy
67: distributions (SED). The iron K$\alpha$ line profiles from both the
68: LLAGN can be described in terms of two components -- a narrow line
69: at $6.4\kev$ and a moderately broad line (FWHM $\sim 2 \times
70: 10^{4}\kms$) arising from highly ionized, He-like or H-like species
71: ($E \sim 6.8\kev$). We interpret the broad lines arising from an
72: accretion disk the inner edge of which is restricted to large radii
73: ($r_{in} \sim 100 r_g$). However, the Eddington ratio,
74: ${L}/{L_{Edd}}$, of these sources, is 3--4 orders of magnitude lower
75: than that required to photo-ionize a cold disk to He-like iron.
76: We suggest that the
77: lines can be explained as collisionally ionized X-ray lines arising
78: from the transition region between a hot (radiatively inefficient)
79: flow in the inner regions and a cold disk outside $r \sim 100r_g$.
80: The accretion flow geometry probed by our {\it XMM-Newton}
81: observations is consistent with the truncated disk models proposed
82: to explain the SED of LLAGNs.
83:
84: \end{abstract}
85:
86: \keywords{accretion, accretion disks -- galaxies: active --
87: X-rays: galaxies }
88:
89: \section{Introduction}
90:
91: Active galactic nuclei (AGN) typically display a `big blue bump'
92: in their optical/UV spectra
93: interpreted as blackbody emission arising from optically thick,
94: geometrically thin accretion disks. Their hard X-ray power-law
95: spectrum is thought to be produced by Comptonization of the soft disk
96: photons in a hot corona above the accretion disk (e.g.; Haardt \&
97: Maraschi 1991). Low luminosity AGN (LLAGN), emitting well below
98: their Eddington limit (${L}/{{L}_{Edd}} \le 0.01$), typically
99: lack the `big blue bump' (e.g., Ho 1999; Di Matteo et al.~2003 for
100: the case of M87) and for this
101: reason it has been proposed that in these objects the accretion disk
102: consists of of two zones: an outer thin disk that extends from some
103: large radius down to a transition radius and an inner low-radiative
104: efficiency accretion flow close to the black hole.
105:
106: Fluorescence Fe K$\alpha$ line emission can be a very powerful probe for
107: the accretion flow geometry around black holes. Such lines have been
108: interpreted as being produced through X-ray irradiation of the accretion
109: disk; therefore the physical width of the line can serve as a trace of
110: the inner accretion disk extent. Observations of relativistically
111: broadened iron K$\alpha$ lines from some Seyfert galaxies
112: (e.g., MCG-6-30-15 -- Wilms et al 2001; Fabian et al. 2002), which
113: require that the disk must be irradiated from an X-ray source
114: within $6r_g$,
115: suggests that at least in this case the
116: accretion disk extends to the last stable orbit. It is interesting to
117: note that MCG-6-30-15 (FWHM(H$\alpha$)$\sim 2200\kms$; Sulentic et al.
118: 1998) is like narrow-line Seyfert~1 galaxies (NLS1), which are thought to
119: be accreting at relatively high accretion rates.
120:
121: In contrast, Lasota et al. (1996) suggested that the accretion disk in
122: most LLAGN must be truncated. Gammie, Narayan, and Blandford (1999)
123: have shown the broadband spectrum of NGC~4258 can be explained by
124: emission from a geometrically thin accretion disk + ADAF model with a
125: transition radius at $\sim 10-100$ Schwarzschild radii ($r_S$). On the basis of their
126: optical/UV continuum spectra, Quataert et al. (1999), have provided
127: further evidence for such accretion flow geometry and showed that the
128: spectral energy distributions of M~81 and NGC~4579 can be explained by
129: an ADAF + disk with a transition radius at $\simeq 100 r_S$.
130:
131: One prediction from any such models, as discussed in
132: Quataert et al (1999), is the lack of a relativistically broadened
133: iron K$\alpha$ line in the X-ray spectrum of these objects. The
134: presence, or absence, of this feature in X-ray spectra from
135: \chandra{} and \xmm{} can therefore be used to test the accretion
136: flow geometry proposed for LLAGN. Specifically, any detection of
137: relativistically broad iron K$\alpha$ line would strongly argue
138: against truncated disk models.
139:
140: In this paper, we utilize \xmm{} observations of
141: two of the best studied LLAGN, NGC~4579 and M~81, to investigate whether the
142: observed iron K$\alpha$ profiles are consistent with
143: truncated disk models. Both NGC~4579 and M~81 show broad H$\alpha$ line and
144: are classified as type 1 AGN (Barth et al. 2001; Filippenko \& Sargent 1985).
145:
146: \section{Observation and Data Reduction}
147: \xmm{} observed NGC~4579 and M~81 on 12 June 2003 and 22-23 April
148: 2001 respectively, using he European Photon Imaging Camera (EPIC),
149: the reflection grating spectrometer (RGS), and the optical monitor
150: (OM) instruments. The EPIC PN \citep{Struderetal01} and MOS
151: \citep{Turneretal01} cameras were operated in full frame mode using
152: the thin filter for the observation of NGC~4579.
153: For the observation of M 81, the EPIC PN
154: was operated in small window, the MOS1 camera was operated in fast
155: uncompressed mode and the MOS2 camera was operated in full frame
156: mode, with the medium filter in place for all three detectors.
157: We present the spectral data obtained from the EPIC cameras for
158: NGC 4579 and the spectral data from the EPIC PN and MOS2 camera for M~81.
159: The EPIC PN exposures were 22ks and 130ks for NGC~4579 and M~81, respectively.
160:
161: The raw events were processed and filtered using the most recent
162: updated calibration database and analysis software ({\tt SAS v5.4.1})
163: available in August 2003. Events in the bad pixels file and those
164: adjacent pixels were discarded. Only events with pattern $0-4$ (single and
165: double) for the PN and $0-12$ for the MOS were selected.
166: Light curves extracted from source free
167: regions were examined for flaring particle background and affected
168: exposure intervals were excluded from the rest of the analysis. This
169: resulted in `good' exposure time of $15.2{\rm~ks}$ and $ {\rm~ks}$ for
170: NGC~4579 and $71.9{\rm~ks}$ M~81, respectively.
171:
172: Source spectra were extracted from a circular region of radius
173: $50\arcsec$ for both NGC~4579 and M~81. Associated background spectra
174: were extracted from source free regions. The total number of source
175: photons, detected with the PN is $4.1\times 10^4$ for NGC 4579 and
176: $5.44\times10^{5}$ for M~81. The source spectra were grouped to a
177: minimum of 20 counts per spectral channel with appropriate sampling of
178: the data ($\sim 3$ channels per spectral resolution elements) and were
179: analyzed using {\tt XSPEC 11.2.0} (Arnaud 1996). All the errors quoted
180: below were calculated at $90\%$ confidence level for one interesting parameter.
181:
182: \section{Detection of complex iron line profiles}
183: In the following analysis, only the EPIC data above $2.5\kev$ are
184: examined, the results presented in this paper are not critical to the
185: soft X-ray spectrum. The soft X-ray spectrum of M~81 obtained from
186: the RGS observation has been studied by Page et al. (2003b). For each
187: object, at first the MOS and PN spectra were fitted separately to
188: check for the uncertainties due to cross-calibration problems. We
189: found generally good agreement between MOS2 and PN cameras ($\Delta
190: \Gamma \sim 0.03$) in the $2.5-10\kev$ bands. However, the MOS1 data results in
191: a flatter ($\Delta\Gamma \sim 0.15$) spectrum. These discrepancies are
192: consistent with that found by Molendi \& Sembay (2003). Therefore we
193: present the spectral results obtained by fitting the same model
194: jointly to the PN and MOS2 data in the $2.5-12\kev$ band, and allowing
195: the relative normalizations to vary.
196:
197: \subsection{NGC 4579}
198: We fit the PN and MOS2 spectrum of NGC 4579 jointly with a power-law
199: model modified by photoelectric absorption. This fit resulted in a
200: $\chi^2$ of $296.6$ for 253 degrees of freedom (dof). The best-fit
201: absorption column density is consistent with the Galactic column
202: density ($N_{H}=3.1\times10^{20}{\rm~cm^{-2}}$ along the line of sight
203: of NGC~4579. From this simple fit, at least two iron line components
204: are evident in the X-ray spectrum of NGC~4579. Figure~\ref{f1} shows
205: the iron K$\alpha$ line profile of NGC~4579 as the ratio of the
206: observed PN data to the the best-fit continuum model (see below).
207: Addition of a narrow Gaussian line at $\sim 6.4\kev$ improved the fit
208: significantly ($\chi^2=267.1$ for 251 dof). However, excess emission
209: is seen at $\sim 6.7\kev$. The addition of a broad Gaussian line
210: (model 1: power-law + narrow Gaussian + broad Gaussian)
211: improves the joint fit significantly ($\chi^2=242.5$ for $249$ dof),
212: over the best-fit single Gaussian line model and also results in a
213: better fit than the best-fit from a model consisting of an absorbed
214: power-law and two narrow Gaussian lines ($\chi^2=257.0$ for $249$
215: dof). The best-fit parameters are $\Gamma=1.77_{-0.07}^{+0.06}$,
216: narrow Gaussian: $E_{K\alpha,n}=6.40\kev$ (fixed), $EW=106\ev$, broad
217: Gaussian: $E_{K\alpha,b}=6.79_{-0.13}^{+0.11}\kev$,
218: $\sigma=231_{-97}^{+102}\ev$, $EW=287\ev$ (see model 1 in Table~\ref{tab1}).
219:
220: The observed flux of NGC~4579 in the $2.5-12\kev$ band is $3.9\times
221: 10^{-12}{\rm~erg~cm^{-2}~s^{-1}}$, and the corresponding luminosity is
222: $1.2\times10^{40}{\rm~erg~s^{-1}}$ (for a distance of
223: $d=16.7{\rm~Mpc}$ as measured by Pierce \& Tully, 1998). An {\it
224: ASCA} observation of NGC~4579 in 1995 showed a strong (EW $\sim
225: 500\ev$), marginally broad, symmetrical, iron K$\alpha$ line profile
226: at $6.73\pm0.13\kev$, while the previous 1998 observation showed a
227: narrow line at $6.39\pm0.09\kev$ (EW $\sim 260\ev$) with no clear line
228: at $\sim 6.7\kev$. It appears that both the above components observed
229: with \asca{} are clearly detected in the current \xmm{} observation.
230:
231: \subsection{M 81}
232: The PN and MOS2 spectra of M81 were also fit by a power-law
233: modified by photoelectric absorption. The best-fit absorption column
234: density is consistent with the Galactic value
235: ($\NH=4.2\times10^{20}{\rm~cm^{-2}}$). This simple fit is
236: statistically unacceptable ($\chi2$ of $397.5$ for 303 dof) due to
237: significant spectral features in the iron K$\alpha$ region (see
238: Fig.~\ref{f1}). The complex line profile of M~81 is not well
239: described by a narrow Gaussian line with the power-law plus narrow Gaussian model
240: resulting in a poor fit ($\chi^2=363.0$ for 303
241: dof). The addition of a second narrow Gaussian line at $\sim 6.7\kev$
242: improves the fit significantly ($\chi^2 = 318.3$ for 301 dof). Varying
243: the width of this line further improves the fit ($\chi^2 = 299.3$ for
244: 300 dof). The best-fit parameters are $\Gamma_X=1.87\pm0.02$, narrow
245: Gaussian: $E_{K\alpha,n}=6.40\kev$ (fixed), $EW=31\ev$, broad Gaussian:
246: $E_{K\alpha,b}=6.79_{-0.07}^{+0.06}\kev$, $\sigma=188_{-57}^{+76}\ev$,
247: $EW=101\ev$.
248:
249: The observed flux of M~81 in the $2.5-12\kev$ band is $1.2\times
250: 10^{-11}{\rm~erg~cm^{-2}~s^{-1}}$, and the luminosity in the same band
251: is $1.8\times10^{40}{\rm~erg~s^{-1}}$ (for a distance of
252: $d=3.6{\rm~Mpc}$; Freedman et al. 1994).
253:
254: The \xmm{} observations of M~81 confirm the previous \asca{} and
255: \sax{} results that detected an intrinsically broad or complex iron line
256: centered at $\sim6.7\kev$, with an equivalent width of
257: $\sim170\pm60\ev$ (Ishisaki et al.1996). Neither \asca{} nor \xmm{}
258: observations detected an absorption edge at $\sim 8.6\kev$, similar to
259: that observed previously with \sax{} (Pellegrini et al.~2000).
260:
261: \section{Inferences from the iron lines}
262:
263: Both NGC~4579 and M~81 show iron line emission with broad ($FWHM \sim
264: 20000\kms$), symmetrical line profiles, in addition to narrow iron
265: K$\alpha$ lines at $6.4\kev$.
266:
267: The broad iron K$\alpha$ lines cannot be easily interpreted as thermal
268: emission from the host galaxy because the inferred thermal broadening
269: from such a system is too small ($FWHM\sim 350\kms$ for a temperature
270: of $10^8{\rm~K}$) to explain the widths of the observed lines. A
271: thermal origin for the lines is further ruled out by the variability
272: of the lines as observed by \asca{} (Terashima et al. 2000).
273:
274: The observed broad
275: line from NGC~4579 and M~81 may thus arise from the accretion disk, as
276: generally assumed in Seyfert galaxies.
277: The broad Gaussian lines in the current models were replaced with the
278: disk-line model of Fabian et al. (1989) in order to test this
279: hypothesis. This model assumes a Schwarzschild geometry and a disk
280: emissivity that is a power-law as a function of disk radius ($\epsilon
281: \propto r^{-q}$). The emissivity index was fixed at $q=2.5$, the
282: average value for Seyfert~1 galaxies derived by Nandra et al. (1997).
283: The outer disk radius was also fixed at $1000r_g$ and the disk
284: inclination angle $i$ was fixed at $30\deg$. The free parameters of
285: the disk-line model were the line energy, the inner radius of the
286: accretion disk, and the line flux. This model
287: (model 2: power-law + narrow Gaussian + diskline) results in an
288: acceptable fit for both AGN; $\chi^2/dof=240.6/249$ for NGC~4579 and
289: $\chi^2/dof=296.3/300$ for M~81. The continuum and narrow Gaussian
290: line parameters did not change significantly (see Table~\ref{tab1}).
291:
292: Figure~\ref{f2} shows the observed PN data, the best-fit model (which
293: includes an absorbed power law a narrow Gaussian and a diskline) and the
294: deviations (in units of $\sigma$) for NGC~4579 and M~81. The iron
295: K$\alpha$ profiles, shown in Fig.~\ref{f1} as the ratio of observed
296: data and the best-fit continuum model, were obtained by setting the
297: normalizations of the narrow and broad K$\alpha$ lines to be zero.
298: Only the ratios to the PN data are plotted in Fig.~\ref{f1} for clarity.
299: In order to evaluate the
300: dependence of the line properties on disk inclination, we also fit the
301: data for $i=60\deg$. This results in $E_{diskline} =
302: 6.70_{-0.09}^{+0.09}\kev$; $r_{in} = 232_{-117}^{+490} r_g$ for
303: NGC~4579 and $E_{diskline} = 6.82_{-0.10}^{+0.05}\kev$;
304: $r_{in}>160r_g$ for M~81.
305: Fitting accretion disk models to the data suggest that the line
306: emitting region of the disk is truncated at an inner radius
307: of about or greater than $100 r_g$ (the narrow core
308: to the $6.4$ keV line suggests that this component
309: must also originate at a substantial distance from the black hole).
310:
311: Our best-fit model does not show any
312: residuals indicating the presence of a substantial reflection continuum (from
313: either neutral of partially ionized matter from a disk close to the
314: black hole) in the spectral data of either M81 or NGC~4579 (see Fig.
315: 2). The apparently lack of reflection continuum was also confirmed by
316: the detailed spectral analysis of the \sax{} observation of M81
317: (Pellegrini 2000).
318: It is therefore unlikely the the $6.7$ keV emission is produced
319: by reflection from an ionized disk (see also \S 5.2).
320: To assess the possible origin of the $\sim 6.7\kev$ broad feature in
321: the spectrum of NGC~4579 as the Compton reflection emission,
322: we fitted a Compton reflection model ({\tt pexrav}; Magdziarz \&
323: Zdziarski 1995) along with the absorbed power-law and narrow Gaussian
324: line model to the PN spectrum of NGC~4579. The best-fit of this
325: reflection model results in an unrealistically large relative reflection parameter,
326: $R=8.2_{-1.5}^{+8.9}$ (where $R=1$ corresponds to a solid angle of
327: $2\pi$ subtended by the reflector) and the fit worsens
328: ($\Delta \chi^2 = +20.7$ for an additional dof) in comparison to the best-fit
329: obtained with the absorbed power law and double Gaussian line model (see Table~\ref{tab1}).
330: Thus the unphysical value of the reflection parameter and the
331: poor quality of the fit rule out
332: Compton reflection emission in place of the fluorescence iron
333: K$\alpha$ emission.
334:
335:
336: Finally, a combination of multiple narrow Gaussian lines was also tested as a
337: model to fit the complex iron line profiles observed in these sources.
338: Three narrow Gaussian emission lines were added to the absorbed power-law
339: model with fixed line energies of
340: $6.4$, $6.7$ and $7.0\kev$, corresponding to cold, He-line, and H-like
341: iron. The fits based on this model are statistically acceptable fits
342: (NGC~4579, $\chi^2=248.5$ for 250 dof; M~81, $\chi^2=304.1$ for 301
343: dof). These are worse fits compared to the diskline
344: models ($\Delta\chi^2=+7.9$ and $+7.8$ for NGC~4579 and M~81,
345: respectively) at a confidence level of $99.5\%$ based on an F-test.
346: With the current generation of X-ray spectroscopy
347: satellites it is not possible to clearly distinguish between the multiple
348: ($\ge 3$) narrow line and the disk-line models however, observations
349: with, {\it Astro~E2}, and {\it Constellation-X}, with their correspondingly
350: higher spectral resolution and sensitivity will be particularly important to
351: further investigate the complexity of these iron line profiles.
352:
353: \section{Origin of iron line complex}
354: Both NGC~4579 and M~81 show complex Iron K$\alpha$ features which are
355: well described by a combination of a narrow Gaussian and either a
356: broad Gaussian emission line, a disk-line, or a combination of narrow
357: emission lines from multiple ionization states of iron.
358:
359: \subsection{The neutral, narrow iron K$\alpha$ line}
360: Despite the good spectral resolution of the EPIC cameras, the
361: line energies and the widths of the observed narrow K$\alpha$ lines
362: are not well constrained. The line parameters for NGC~4579 are
363: similar to the narrow lines commonly observed from Seyfert galaxies
364: with \chandra{} and \xmm{} (e.g., Yaqoob et al. 2001; Pounds et al.
365: 2001), however, the strength of the narrow K$\alpha$ line is weaker in
366: M~81, in comparison to that of 'typical' Seyfert~1 galaxies (see e.g.
367: Page et al. 2003a). The observed narrow-lines from NGC~4579 and M~81
368: can be produced anywhere in the broad-line region, dusty torus, outer
369: regions of an standard accretion disk or a disk with a flat emissivity
370: law (Yaqoob et al. 2001,2003). The weaker narrow line of M~81 may
371: suggest a smaller covering fraction of the material responsible for
372: the narrow line emission in NGC~4579 than that in M~81.
373:
374: \subsection{The ionized, broad iron K$\alpha$ line}
375:
376: The widths and line center energies of the broad K$\alpha$ lines from
377: NGC~4579 and M~81 are remarkably similar ($E\sim 6.7-6.8\kev$; FWHM
378: $\sim 20000\kms$) suggestive of both an origin in similar regions the
379: AGN and a similarity in the ionization stages (He- or H-like iron) of
380: the emitting material. The observed line energies are higher than
381: those generally observed from Seyfert 1. In general, only narrow-line
382: Seyfert~1 galaxies (NLS1s), with steep X-ray spectra, regularly show
383: strong, highly ionized iron K$\alpha$ lines (Dewangan 2002). In such
384: cases the lines are interpreted as being produced by reflection from
385: material in the disk photo-ionized by the strong X-ray continuum. The
386: physical problem with explaining the lines in M81 and NGC~4579 in a
387: similar way is that we do not expect photoionization to be important
388: in these low luminosity sources. For photo-ionized material, the
389: ionization parameter can be written as $\xi=\frac{L}{nR^2}$, where $L$
390: is the luminosity of ionizing photons, $n$ is the density of the
391: material, and $R$ is the distance of the material from the ionizing
392: source (Kallman \& McCray 1982). For standard thin disks, Matt et al.
393: (1993) have shown that the ionization parameter is strongly dependent
394: on the accretion rate relative to the Eddington rate and hence can be
395: rewritten as, $\xi \propto ({{L}/{L_{Edd}}})^3$. In order to ionize
396: He-like iron, $\xi$ should be at least $\sim500$ or
397: ${L}/{L_{Edd}}\gtrsim 0.2$ (Matt et al. 1993). This result is
398: consistent with the observation of highly ionized iron K$\alpha$ lines
399: from some NLS1s as they are thought to emit a higher fraction of their
400: Eddington luminosity (Pounds, Done, \& Osborne 1995). For NGC~4579,
401: the total nuclear luminosity of $\sim 5\times10^{41}\ergsec$ (Barth et
402: al. 1996) and black hole mass in the range of
403: $1.8-9.3\times10^7{\rm~M\odot}$ (Barth et al. 2001) correspond to
404: ${L}/{L_{Edd}}\sim (0.5-2.2)\times10^{-4}$ while Laor (2003) reported
405: ${L}/{L_{Edd}}\sim 1.35\times10^{-4}$. Ho et al. (1996) found
406: ${L}/{L_{Edd}}\sim (2-10)\times10^{-4}$ for M~81 which has a black
407: hole mass of $4 \times 10^{6} \Msun$. Thus the ionizing luminosity of
408: both the LLAGNs NGC~4579 and M~81 are 3-4 orders of magnitude smaller
409: than that required for a photo-ionized accretion disk with He-like
410: iron. Clearly, the dominant ionization process in the disk of NGC~4579
411: and M~81 cannot be photo-ionization.
412:
413: The widths of the ionized iron K$\alpha$ lines from NGC~4579 and M~81
414: (FWHM$\sim20000\kms$) are not as broad as the line widths predicted by
415: models of an accretion disk extending down to the last stable orbit.
416: Assuming Keplerian motion of the emitting material (i.e., $r\sim
417: ({c^2}/{v_{FWHM}^2}) r_g$) the size of the broad iron K$\alpha$ line
418: emitting material is estimated to be $\sim150 r_g$ for NGC~4579 and
419: $\sim230 r_g$ for M~81, independent of black hole mass. Thus the
420: majority of iron K$\alpha$ photons are not produced near the last
421: stable orbit for a Schwarzschild ($r_{in}=6r_g$) or a Kerr black hole
422: ($1.235r_g\le r_{in} \ 6r_g$). The same conclusion can be drawn from
423: the disk-line fits which resulted in $r_{in}=50_{-21}^{+111}r_g$ for
424: NGC~4579 and $65_{-27}^{+105}r_g$ for M~81.
425:
426: Two possible scenarios that can explain the observed
427: moderate line widths and their $\sim6.7-6.8\kev$ centroid energy are:
428: ($i$) the
429: accretion disk is fully ionized for $r<r_{in}$ so there are no lines
430: arising from within this region, or ($ii$) the accretion disk is truncated at
431: $r\sim r_{in}$. As discussed above, however, the accretion disk
432: cannot be photo-ionized
433: by irradiation, and thus
434: must be intrinsically
435: hot. Both ($i$) and ($ii$) then imply that the inner regions of the
436: accretion flows in M~81 and NGC~4579 LLAGNs are likely to be
437: different from those of typical Seyfert galaxies.
438:
439: \subsection{The truncated disk model}
440: Our results are consistent with the suggestion that LLAGN have
441: accretion disks that are truncated at some large radii and have a hot
442: flow in the inner regions (Lasota et al. 1996; Narayan 1996; Gammie et
443: al.; Li \& Wang 2000). In particular, the inner radius $r_{in}\sim100
444: r_s$ of the accretion disk implied by the observed iron line width is
445: fully consistent with the position of the truncation radius deduced
446: from the modeling of the optical/UV spectra of M~81 and NGC~4579 by
447: Quataert et al. (1999; assuming a geometrically-thin accretion disk
448: and an optically thin, two-temperature ADAF in the inner regions).
449: Note also that in
450: this model, the continuum X-ray emission is produced in the inner
451: region of the hot flow (from $r=1 - 100 r_s$) and is due to
452: Comptonization of synchrotron emission from the flow.
453:
454:
455: We suggest that the relatively broad $6.7$ keV lines may arise from
456: regions of the flow close to or at the transition radius, where the hot,
457: fully ionized flow starts cooling to connect to the outer optically
458: thick, geometrically thin region. Using the dynamical solutions of
459: the transition regions from a thin disk to an ADAF recently derived by
460: Manmoto et al. (2001) we can try to estimate the densities and
461: temperatures typical for this transition region.
462: In order for the hot flow to connect to a
463: standard thin disk the temperature and velocity of the gas must change
464: by orders of magnitude over a small transition region between the ADAF
465: and the thin disk. For a transition radius of $r_{tr} \sim 100 r_s$,
466: the temperature ($T \propto c_s^2$) scales as $\sim r^{-1}$ in the inner
467: regions, and as
468: it approaches the transition radius, it then drops sharply from values
469: of $T\sim10^9$ K at $r\sim50 r_s$ to $\sim10^8$ K at $r\sim100 r_s$
470: and cooler beyond. As pressure balance has to be maintained, the
471: density of the flow increases sharply (by $\sim$2 orders of magnitude)
472: in the same range of radii, reaching values of $n_e\sim2\times10^{12}$
473: cm$^{-3}$ at $r\sim100 r_s$. Even using the values of the flow
474: density derived above, photo-ionization in the transition region is
475: not likely to be important because of the low level of the ionizing
476: hard X-ray flux, in fact, at the transition region we find that the
477: ionization parameter $\xi\sim$1. Instead,
478: %We find $\xi=\frac{L}{n r_{tr}^2}\sim 1$ at the transition region. T
479: the observed He-like and H-like iron lines are likely to represent
480: collisionally ionized X-ray lines produced in a hot, $T\sim10^8$ K,
481: transition region. Specific calculations of the emission line profiles from
482: this region are not available at the moment and would be necessary
483: to fully explore this hypothesis. To
484: test this inference, therefore, we consider a simple
485: thermal plasma model ({\tt mekal};
486: Liedahl, Osterheld, \& Goldstein (1995) and references therein) and
487: fit it in combination with a power-law continuum and a narrow Gaussian
488: emission line at $6.4\kev$ (fixed) to the PN and MOD spectra jointly.
489: The density of the plasma was fixed to $n=2\times10^{12}{\rm~cm^{-3}}$
490: (as derived above) and no intrinsic absorption is required.
491: The mekal model was
492: convolved with a Gaussian in order to mimic the line broadening due to
493: Keplerian motion of the plasma at the transition region. The Gaussian
494: $\sigma$ was fixed at $0.2\kev$, similar to that derived for the broad
495: Gaussian component in the double Gaussian model for the iron line (see
496: Table~\ref{tab1}). The model provided a good fit to the data
497: ($\chi^2/dof= 246.4/249$ for NGC~4579 and $298.0/296$ for M~81),
498: albeit somewhat worse than the models incorporating either a broad
499: Gaussian or diskline. As an example, this model is shown in
500: Figure~\ref{f3} for the PN data of M~81. The best-fit parameters are
501: $\Gamma=1.73_{-0.11}^{+0.06}$, $kT=8.3_{-5.6}^{+7.1}\kev$ for NGC~4579
502: and $\Gamma=1.87\pm0.04$, $kT=8.5_{-3.0}^{+3.2}\kev$ for M~81. The
503: best-fit plasma temperatures for both the AGN are similar to that
504: deduced above for the transition region. The power-law and the mekal
505: fluxes are $2.6\times10^{-12}$ and
506: $1.3\times10^{-12}{\rm~erg~cm^{-2}~s^{-1}}$ for NGC~4579 and
507: $1.0\times 10^{-11}{\rm~erg~cm^{-2}~s^{-1}}$ and $1.3\times
508: 10^{-12}{\rm~erg~cm^{-2}~s^{-1}}$, respectively.
509:
510: The success of this model in reproducing the observed iron K$\alpha$
511: profiles suggests that the observed lines could be produced
512: by collisional excitation in a relatively hot transition
513: region between a hot flow and a thin disk.
514: The truncation radii of $\sim100 r_S$ radii inferred by
515: Quataert et al. (1999) to fit
516: the optical/UV to X-ray SED is somewhat larger than the best-fit inner
517: radii inferred from the diskline model fits to the iron K$\alpha$ line
518: profile. However, is has been noted the if a standard disk is connected to an
519: ADAF across a narrow transition layer, the gas near the transition
520: radius must rotate with a super-Keplerian angular velocity
521: (Abramowicz, Igumenshchev, \& Lasota 1998)
522: This means $v_{FWHM}{\rm (Fe~K\alpha)} > c ({r_g}/{r_{tr}})$ or $r_{tr} >
523: r_{in}\sim ({c^2}/{v_{FWHM}^2}) r_g$. Additionally, the transition region,
524: being hot ($\sim10^8 {\rm~K}$), may not contribute the optical/UV
525: emission which determined the truncation radius quoted by Quataert
526: et al. (1999).
527: Note also that ADAFs are likely to develop strong outflows
528: (Blandford \& Begelman 1999) or convection (Quataert \& Gruzinov 1999;
529: Narayan et al. 2000) which are likely to change the dynamics
530: of the transition region.
531:
532: \subsection{Alternative scenario -- multiple narrow lines?}
533: The broad iron line components of NGC~4579 and M~81 can also be described in
534: terms of multiple narrow emission lines from ionized iron, although with a
535: lesser statistical significance. Such lines can
536: arise from ($i$) hot thermal plasma, if present, outside the accretion disk
537: e.g host galaxy, or
538: ($ii$) a photo-ionized accretion disk with flat emissivity law extending down
539: to the last stable orbit (Yaqoob et al. 2003). The former process has already
540: been ruled out due to the variability of the iron line from NGC~4579
541: (Terashima et al. 2000). The later possibility is unlikely due to a
542: low ionization parameter (see $\S 5.3$). Thus the case for multiple narrow
543: lines for the iron K$\alpha$ complexes of NGC~4579 and M~81 is poor at
544: present.
545:
546:
547: \section{Conclusions}
548: Using the \xmm{} EPIC observations, we detected complex iron line
549: profiles from two LLAGN NGC~4579 and M~81. The line profiles are well
550: described by a combination of a narrow neutral line and a broad
551: Gaussian (FWHM $\sim 20000\kms$) or an accretion-disk line arising
552: from highly ionized (He- or H-like) iron. The line profiles from both
553: the AGN are remarkably similar except that the neutral iron K$\alpha$
554: line from M~81 is much weaker than that that from NGC~4579. The main
555: result of this work is that the broad and highly ionized iron
556: K$\alpha$ lines from two LLAGNs NGC~4579 and M~81 are consistent with
557: accretion disks whose inner edges are restricted to large radii
558: $r_{in} \sim 100r_g$.
559:
560: \acknowledgements We thank an anonymous referee for the comments and
561: suggestions on this paper. This work is based on
562: observations obtained with \xmm{}, an ESA science mission with
563: instruments and contributions directly funded by ESA Member States and
564: the USA (NASA). REG acknowledges NASA award NAG5-9902 in support of
565: his Mission Scientist position on \xmm{}.
566:
567: \clearpage
568:
569: \begin{figure*}
570: \centering
571: \includegraphics[width=6cm,angle=-90]{f1.eps}
572: \caption{Iron K$\alpha$ line profiles of NGC~4579 and M~81 shown as the ratio of the observed PN data and the best-fit continuum model (see text).
573: \label{f1}}
574: \end{figure*}
575:
576:
577: \begin{figure*}
578: \centering
579: \includegraphics[width=6cm,angle=-90]{f2a.eps}
580: \includegraphics[width=6cm,angle=-90]{f2b.eps}
581: \caption{The observed EPIC PN and MOS2 spectra of NGC~4579 and M~81, the best-fit models consisting of an absorbed power-law, a narrow Gaussian, and a diskline (see text). Lower panels show the deviations of the observed data from the best-fit models.
582: \label{f2}}
583: \end{figure*}
584:
585:
586: \begin{figure*}
587: \centering
588: \includegraphics[width=6cm,angle=-90]{f3.eps}
589: \caption{The best-fit model consisting of (i) mekal plasma model, (ii) a narrow Gaussian, and (iii) a power-law derived from the joint fit to the PN and MOS2 spectra of M~81. The model is shown for the PN data only for clarity, the model for the MOS2 data differs in relative normalization by a small fraction ($\sim 0.02\%$). The mekal model was convolved with a Gaussian of $\sigma\sim0.2\kev$ (see text)}.
590: \label{f3}
591: \end{figure*}
592:
593: \clearpage
594:
595: \begin{thebibliography}{}
596: \bibitem[Abramowicz, Igumenshchev, \& Lasota(1998)]{1998MNRAS.293..443A}
597: Abramowicz, M.~A., Igumenshchev, I.~V., \& Lasota, J.\ 1998, \mnras, 293,
598: 443
599: \bibitem[Arnaud et al.(1985)]{Arnaudetal85} Arnaud, K.~A.~et al.\
600: 1985, \mnras, 217, 105
601: \bibitem[Barth et al.(2001)]{2001ApJ...546..205B} Barth, A.~J., Ho, L.~C.,
602: Filippenko, A.~V., Rix, H., \& Sargent, W.~L.~W.\ 2001, \apj, 546, 205
603: \bibitem[Barth et al.(1996)]{1996AJ....112.1829B} Barth, A.~J., Reichert,
604: G.~A., Filippenko, A.~V., Ho, L.~C., Shields, J.~C., Mushotzky, R.~F., \&
605: Puchnarewicz, E.~M.\ 1996, \aj, 112, 1829
606: \bibitem[Dewangan(2002)]{2002ApJ...581L..71D} Dewangan, G.~C.\ 2002, \apjl,
607: 581, L71
608: \bibitem[Di Matteo et al.(2003)]{2003ApJ...582..133D} Di Matteo, T., Allen,
609: S.~W., Fabian, A.~C., Wilson, A.~S., \& Young, A.~J.\ 2003, \apj, 582, 133
610: \bibitem[Freedman et al.(1994)]{1994ApJ...427..628F} Freedman, W.~L.~et
611: al.\ 1994, \apj, 427, 628
612: \bibitem[Gammie, Narayan, \& Blandford(1999)]{1999ApJ...516..177G} Gammie,
613: C.~F., Narayan, R., \& Blandford, R.\ 1999, \apj, 516, 177
614: \bibitem[Fabian et al.(1989)]{Fabianetal89} Fabian,
615: A.~C., Rees, M.~J., Stella, L., \& White, N.~E.\ 1989, \mnras, 238, 729
616: \bibitem[Ishisaki et al.(1996)]{1996PASJ...48..237I} Ishisaki, Y.~et al.\
617: 1996, \pasj, 48, 237
618: \bibitem[Kaspi et al.(2001)]{2001ApJ...554..216K} Kaspi, S.~et al.\ 2001,
619: \apj, 554, 216
620: \bibitem[Liu \& Meyer-Hofmeister(2001)]{2001A&A...372..386L} Liu, B.~F.~\&
621: Meyer-Hofmeister, E.\ 2001, \aap, 372, 386
622: \bibitem[Lasota et al.(1996)]{1996ApJ...462..142L} Lasota, J.-P.,
623: Abramowicz, M.~A., Chen, X., Krolik, J., Narayan, R., \& Yi, I.\ 1996,
624: \apj, 462, 142
625: \bibitem[Liu et al.(1999)]{1999ApJ...527L..17L} Liu, B.~F., Yuan, W.,
626: Meyer, F., Meyer-Hofmeister, E., \& Xie, G.~Z.\ 1999, \apjl, 527, L17
627: \bibitem[Lu \& Wang(2000)]{2000ApJ...537L.103L} Lu, Y.~\& Wang, T.\ 2000,
628: \apjl, 537, L103
629: \bibitem[Matt, Fabian, \& Ross(1993)]{1993MNRAS.262..179M} Matt, G.,
630: Fabian, A.~C., \& Ross, R.~R.\ 1993, \mnras, 262, 179
631: \bibitem[Manmoto, Kato, Nakamura, \& Narayan(2000)]{2000ApJ...529..127M}
632: Manmoto, T., Kato, S., Nakamura, K.~E., \& Narayan, R.\ 2000, \apj, 529,
633: 127
634: \bibitem[Molendi \& Sembay(2003)]{MS03} Molendi, S., \& Sembay, S.\ 2003, XMM-SOC-CAL-TN-0036.
635: \bibitem[Nandra et al.(1997)]{1997ApJ...477..602N} Nandra, K., George,
636: I.~M., Mushotzky, R.~F., Turner, T.~J., \& Yaqoob, T.\ 1997, \apj, 477, 602
637: \bibitem[Page, O'Brien, Reeves, \& Turner(2003)]{2003astro.ph..9394P} Page,
638: K.~L., O'Brien, P.~T., Reeves, J.~N., \& Turner, M.~J.~L.\ 2003a, ArXiv
639: Astrophysics e-prints, 9394
640: \bibitem[Page et al.(2003)]{2003A&A...400..145P} Page, M.~J., Breeveld,
641: A.~A., Soria, R., Wu, K., Branduardi-Raymont, G., Mason, K.~O., Starling,
642: R.~L.~C., \& Zane, S.\ 2003b, \aap, 400, 145
643: \bibitem[Pellegrini et al.(2000)]{2000A&A...353..447P} Pellegrini, S.,
644: Cappi, M., Bassani, L., Malaguti, G., Palumbo, G.~G.~C., \& Persic, M.\
645: 2000, \aap, 353, 447
646: \bibitem[Pounds, Done, \& Osborne(1995)]{1995MNRAS.277L...5P} Pounds,
647: K.~A., Done, C., \& Osborne, J.~P.\ 1995, \mnras, 277, L5
648: \bibitem[Pounds et al.(2001)]{2001ApJ...559..181P} Pounds, K., Reeves, J.,
649: O'Brien, P., Page, K., Turner, M., \& Nayakshin, S.\ 2001, \apj, 559, 181
650: \bibitem[Quataert, di Matteo, Narayan, \& Ho(1999)]{1999ApJ...525L..89Q}
651: Quataert, E., Di Matteo, T., Narayan, R., \& Ho, L.~C.\ 1999, \apjl, 525,
652: L89
653: \bibitem[Str{\" u}der et al. 2001]{Struderetal01} Str{\" u}der,
654: L.~et al.\ 2001, \aap, 365, L18
655: \bibitem[Terashima et al.(1998)]{1998ApJ...503..212T} Terashima, Y.,
656: Kunieda, H., Misaki, K., Mushotzky, R.~F., Ptak, A.~F., \& Reichert, G.~A.\
657: 1998, \apj, 503, 212
658: \bibitem[Terashima et al.(2000)]{2000ApJ...535L..79T} Terashima, Y., Ho,
659: L.~C., Ptak, A.~F., Yaqoob, T., Kunieda, H., Misaki, K., \& Serlemitsos,
660: P.~J.\ 2000, \apjl, 535, L79
661: \bibitem[Turner et al. 2001]{Turneretal01} Turner, M.~J.~L.~et al.\
662: 2001, \aap, 365, L27
663: \bibitem[Yaqoob et al.(2001)]{2001ApJ...546..759Y} Yaqoob, T., George,
664: I.~M., Nandra, K., Turner, T.~J., Serlemitsos, P.~J., \& Mushotzky, R.~F.\
665: 2001, \apj, 546, 759
666: \bibitem[Yaqoob et al.(2003)]{2003ApJ...596...85Y} Yaqoob, T., George,
667: I.~M., Kallman, T.~R., Padmanabhan, U., Weaver, K.~A., \& Turner, T.~J.\
668: 2003, \apj, 596, 85
669: \end{thebibliography}
670:
671: \clearpage
672:
673: \begin{table*}
674: \caption{Best-fit spectral model parameters for NGC~4579 and M~81.}
675: \label{tab1}
676: {\small
677: \begin{tabular}{lccccccc}
678: \tableline\tableline
679: Comp. & Parameter\tablenotemark{a} & \multicolumn{2}{c}{NGC~4579} & \multicolumn{2}{c}{M~81}
680: \\
681: & & model 1\tablenotemark{b} & model 2\tablenotemark{b} & model 1\tablenotemark{b} & model 2\tablenotemark{b} \\ \tableline
682: PL & $\Gamma_X$ & $1.77_{-0.07}^{+0.06}$ & $1.77_{-0.07}^{+0.06}$ & $1.86_{-0.02}^{+0.02}$ & $1.86_{-0.02}^{+0.02}$ \\
683: & $n_{PL}$\tablenotemark{c} & $1.02_{-0.09}^{+0.10}$ & $1.02_{-0.09}^{+0.11}$ & $3.6_{-0.1}^{+0.1}$ & $3.6_{-0.1}^{+0.1}$ \\
684: Gaussian & $E_{K\alpha,n}$($\kev$) & $6.40$(f) & $6.40$(f) & $6.40$(f) & $6.40$(f) \\
685: & $\sigma$ ($\kev$) & 0.00(f) & 0.00(f) & 0.00(f) & 0.00(f) \\
686: & $f\tablenotemark{d}$ & $4.7_{-2.5}^{+2.3}$ & $4.3_{-1.9}^{+2.5}$ & $3.7_{-1.9}^{+1.6}$ & $4.2_{-2.3}^{+1.6}$ \\
687: & $EW$ ($\ev$) & 106 & 96 & 31 & 35 \\
688: Gaussian & $E_{K\alpha,b}$($\kev$) & $6.79_{-0.13}^{+0.11}$ & -- & $6.79_{-0.07}^{+0.06}$ & -- \\
689: & $\sigma$ ($\ev$) & $231_{-97}^{+102}$ & -- & $188_{-57}^{+76}$ & --\\
690: & $f\tablenotemark{d}$ & $9.9_{-3.8}^{+4.4}$ & -- & $10.4_{-2.6}^{+3.2}$ & -- \\
691: & $EW$ ($\ev$) & 287 & -- & 101 & -- \\
692: DL & $E_{diskline}$($\kev$) & -- & $6.78_{-0.09}^{+0.06}$ & -- & $6.84_{-0.07}^{+0.04}$ \\
693: & $q$ & -- & 2.5 & -- & 2.5(f) \\
694: & $r_i/r_g$ & -- & $53_{-22}^{+137}$ & -- & $104_{-61}^{+110}$ \\
695: & $r_o/r_g$ & -- & 1000(f) & -- & 1000(f) \\
696: & $i$ & -- & 30(f) & -- & 30(f) \\
697: & $f\tablenotemark{d}$ & -- & $10.0_{-3.6}^{+3.4}$ & -- & $10.1_{-2.3}^{+3.0}$
698: \\
699: & $EW_D$($\ev$) & -- & 302 & -- & 104 \\
700: Total & $\chi^2_{min}$/dof & 242.5/249 & 240.6/249 & 299.3/300 & 296.3/300 \\ \tableline
701: \end{tabular}}
702: \tablenotetext{a}{(f) indicates a fixed parameter value.}
703: \tablenotetext{b}{model 1: absorbed power law $+$ narrow Gaussian $+$ broad Gaussian; model 2: absorbed power law $+$ narrow Gaussian $+$ diskline.}
704: \tablenotetext{c}{$10^{-3}{\rm~photons~cm^{-2}~s^{-1}~keV^{-1}}$ at $1\kev$.}
705: \tablenotetext{d}{$10^{-6}{\rm~photons~cm^{-2}~s^{-1}}$.}
706: %\tablenotetext{c}{}
707: \end{table*}
708:
709:
710: \end{document}
711: