1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{aastex}
3: \usepackage{emulateapj5}
4: \topmargin 20pt
5:
6:
7: \shorttitle{X-ray Spectral variability of NGC~6251}
8: \shortauthors{Gliozzi et al.}
9: \def\ngc{NGC~6251}
10: \def\pks{PKS~0558--504}
11: \def\oiii{[{\sc O\, iii}]}
12: \def\oii{[{\sc O\, ii}]}
13: \def\mgii{[{\sc Mg\, ii}]}
14: \def\feka{Fe K$\alpha$}
15: \def\chandra{{\it Chandra}}
16: \def\xmm{{\it XMM-Newton}}
17: \def\asca{{\it ASCA}}
18: \def\hst{{\it HST}}
19: \def\rxte{{\it RXTE}}
20: \def\sax{{\it BeppoSAX}}
21: \def\ginga{{\it Ginga}}
22: \def\rosat{{\it ROSAT}}
23: \def\ein{{\it Einstein}}
24: \def\integral{{\it Integral}}
25: \def\merlin{{\it MERLIN}}
26: \def\lum{erg s$^{-1}$}
27: \def\flux{erg cm$^{-2}$ s$^{-1}$}
28: \def\nh{cm$^{-2}$}
29: \def\arcsec{$^{\prime\prime}$}
30:
31: \def\ltsima{$\; \buildrel < \over \sim \;$}
32: \def\simlt{\lower.5ex\hbox{\ltsima}} % < over ~
33: \def\gtsima{$\; \buildrel > \over \sim \;$}
34: \def\simgt{\lower.5ex\hbox{\gtsima}} % > over ~
35:
36: \begin{document}
37: \title{Long-Term X-Ray monitoring of NGC~6251:\\
38: Evidence for a jet-dominated radio galaxy}
39:
40: \author{M. Gliozzi}
41: \affil{George Mason University, 4400 University Drive, Fairfax, VA 22030}
42:
43: \author{I.E. Papadakis}
44: \affil{Physics Department, University of Crete, 710 03 Heraklion,
45: Crete, Greece}
46:
47: \author{R.M. Sambruna}
48: \affil{NASA Goddard Space Flight Center, Code 661, Greenbelt, MD 20771}
49:
50:
51: \begin{abstract}
52: We present the first X-ray monitoring observations of the X-ray bright
53: FR~I radio galaxy NGC~6251 observed with \rxte\ for 1 year.
54: The primary goal of this study is to shed light on the origin of
55: the X-rays, by investigating the spectral variability with model-independent
56: methods coupled with time-resolved and flux-selected spectroscopy.
57: The main results can be summarized as follows: 1) Throughout
58: the monitoring campaign, NGC~6251 was in relatively high-flux state
59: with an average 2--10 keV absorbed flux of the order of 4.5$\times 10^{-12}
60: ~{\rm erg~cm^{-2}~s^{-1}}$ and a corresponding intrinsic luminosity of
61: 6$\times 10^{42}~{\rm erg~s^{-1}}$.
62: 2) The flux persistently changed with fluctuations of the order
63: of $\sim$2 on time scales of 20-30 days.
64: 3) When the hardness ratio is plotted against the average count rate, there
65: is evidence for a spectral hardening as the source brightens; this finding
66: is confirmed by a flux-selected spectral analysis.
67: 4) The fractional variability appears to be more pronounced in the hard
68: energy band (5--12 keV) than in the soft one (2.5--5 keV).
69: 5) 2-month averaged and flux-limited energy spectra are adequately fitted
70: by a power law.
71: A Fe K$\alpha$ line is never statistically required, although the presence of
72: a strong iron line cannot be ruled out,
73: due to the high upper limits on the
74: line equivalent width.
75: The inconsistency of the spectral variability behavior of \ngc\
76: with the typical trend observed in Seyfert galaxies and the
77: similarity with blazars lead support to a jet-dominated scenario
78: during the RXTE monitoring campaign.
79: However, a possible contribution from a disk-corona system cannot be
80: ruled out.
81: \end{abstract}
82:
83: \keywords{Galaxies: active --
84: Galaxies: jets --
85: Galaxies: nuclei --
86: X-rays: galaxies
87: }
88:
89: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
90: %%%%%%%%%%%% 1. INTRODUCTION %%%%%%%%%%%%%%%%%%%%
91: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
92: \section{Introduction}
93: Non-blazar radio-loud active galactic nuclei (AGNs) --objects with jets forming
94: large viewing angles to the line of sight--
95: are traditionally divided into two classes:
96: Fanaroff-Riley II (FR~II), with 178 MHz powers $> 2
97: \times 10^{25}$ W/Hz and edge-darkened radio morphologies, and FR~I
98: galaxies with lower powers and more compact morphologies (Fanaroff \& Riley
99: 1974).
100: For the same host galaxy optical
101: magnitude, FRIs produce about one order of magnitude less optical line
102: emission than FRIIs (Baum et al. 1995) and have fainter or
103: negligible UV continuum fluxes (Zirbel \& Baum 1995, Ho 1999).
104:
105: While earlier models for the origin of the FRI/II dichotomy focused
106: mainly on accounting for their large-scale radio morphologies, more
107: recently new ideas have emerged concerning the nature of the central
108: engine in the two types of radio galaxies in an attempt to explain the
109: nuclear properties. One school of thought is that the nuclear X-ray
110: properties of FR~I and FR~II are related to a different accretion rate
111: onto the central supermassive black hole, with FRII being dominated by
112: relatively large values of $L/L_{Edd}$, while FRI would be accreting at
113: sub-Eddington rates, $L/L_{Edd} \ll 10^{-3}$ (e.g., Reynolds et
114: al. 1996; Ghisellini \& Celotti 2001).
115:
116: A fundamental step to gain insight into the nature of the central engine
117: in radio-loud AGNs is to understand whether the X-ray radiation is produced
118: by disk/corona systems as in Seyfert galaxies or by jets as in blazars.
119: While the time-averaged spectroscopy (due to
120: spectral degeneracy) and the pure temporal analysis (due to the fact
121: that both radio-quiet and radio-loud show strong variability) cannot
122: firmly discriminate between the two competing scenarios, their
123: combination, i.e., time-resolved spectral analysis and
124: energy-selected temporal analysis, offers in principle a better way to
125: distinguish between accretion-dominated and jet-dominated systems.
126: This conclusion is supported by
127: the strikingly different spectral variability behavior
128: shown by Seyfert-like objects (e.g., Papadakis et al. 2002; Markowitz \&
129: Edelson 2001) and by blazars (e.g., Zhang et al. 1999;
130: Fossati et al. 2000; Gliozzi et al. 2006).
131:
132: Here, we concentrate
133: on the X-ray nuclear properties of the nearby radio galaxy
134: NGC~6251 ($z$=0.024), which is a giant elliptical galaxy hosting a
135: supermassive black hole with mass $M_{\rm BH}\sim 4-8 \times10^8
136: M_{\odot}$ (Ferrarese \& Ford 1999). Based on its radio power at
137: 178 MHz, NGC~6251
138: is classified as an FR~I (e.g., Owen \& Laing 1989), whereas in the optical,
139: it is classified as a type-2 AGNs (e.g., Shuder \& Osterbrock
140: 1981).
141:
142: Despite the intensive study of this source at all wavelengths,
143: the nature of the accretion process in NGC~6251 is still a matter of
144: debate. Based on the radio-to-X-ray spectral energy distribution, Ho
145: (1999) suggested that an Advection-Dominated Accretion Flow (ADAF) is
146: present in the nucleus of NGC~6251. On the other hand, Ferrarese \&
147: Ford (1999) and Melia et al. (2002) favored a standard accretion
148: disk. Finally, Mukherjee et al. (2002) and Chiaberge et al. (2003)
149: advocate a jet origin for the
150: broad-band emission, based on the possible association of \ngc\ with
151: the {\it EGRET} source 3EG~J1621+8203 and on
152: its spectral energy distribution, respectively.
153:
154: In the X-ray band, NGC~6251 has been previously
155: observed with various satellites. For example,
156: \rosat\ showed the presence of an unresolved nuclear source embedded in a
157: diffuse thermal emission with temperature $kT \sim 0.9$ keV,
158: associated to the galaxy's halo (Birkinshaw \& Worrall 1993). In
159: addition, the existence of a correlation between soft X-rays and radio
160: fluxes prompted Hardcastle \& Worrall (2000) to hypothesize a jet
161: origin for the soft X-rays. On the other hand, the prominent \feka\ line
162: (EW$\simeq$ 400--500 eV)
163: detected by \asca\ in 1994 (Turner et al. 1997; Sambruna et al. 1999)
164: suggested a standard Seyfert-like scenario with accretion-dominated
165: X-rays. \sax\ observed \ngc\ in July 2001 during a high-flux state
166: (the 2--10 keV flux, $F_{\rm X}=4.7\times 10^{-12}$ \flux,
167: was $\sim$3 times larger than the value measured
168: by \asca\ 7 years earlier). Based on the absence of a prominent \feka\ line
169: (EW $<$ 100 eV) or other reprocessing features,
170: Guainazzi et al. (2003) proposed a scenario
171: with two main spectral components: a blazar-like spectrum dominating
172: the high-flux state and a Seyfert-like spectrum emerging during the
173: low-flux state. More recently, in March 2002, \ngc\ was observed with
174: \xmm\ when the 2--10 keV flux was $\sim$15\% lower than the \sax\
175: value. Despite the relatively high flux, the \xmm\ spectral results seem
176: to support the
177: picture emerged from the \asca\ observation with the
178: presence of a prominent (EW$\sim$220 eV) and possibly broad ($\sigma>0.3$ keV)
179: \feka\ line (Gliozzi et al. 2004).
180: However, the existence of a broad \feka\ line in the \xmm\
181: spectrum is still a matter of debate (see Evans et al. 2005 and
182: Gonz\'alez-Mart\'in et al. 2006 for a discording and a supporting view,
183: respectively).
184:
185: The controversial results derived from previous X-ray studies
186: highlight the impossibility of firmly determining the origin
187: of the X-rays in NGC~6251 based solely on time-averaged spectral
188: results. Here, we present the results from a systematic study of the
189: long-term X-ray flux
190: and spectral variability of \ngc\ using one year long proprietary Rossi X-ray
191: Timing Explorer (\rxte) observations in the 2-12 keV range.
192: We use model-independent methods and time-resolved spectroscopy to study
193: the X-ray temporal and spectral properties of this source.
194: The main purpose of this analysis is to shed light on the origin
195: of the X-rays and in particular on the role played by a jet in the X-rays.
196: Once (if) the jet contribution is properly
197: assessed, the physical parameters characterizing the accretion process onto the
198: supermassive black hole can be better constrained, and hence it is
199: possible to discriminate between competing theoretical models for the
200: accretion process.
201:
202: The outline of the paper is as follows. In $\S~2$ we describe the
203: observations and data reduction. The main characteristics of the
204: X-ray light curve are described in $\S~3$. In $\S~4$ we study the X-ray
205: spectral variability of \ngc\ with two model-independent methods.
206: In $\S~5$ we describe the results of a time-resolved and flux-selected
207: spectral analyses.
208: In $\S~6$ we discuss the results and their implications.
209: In $\S~7$ we summarize the main results and conclusions.
210:
211: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
212: %%%%%%%%%%%% 2. OBS & DATA REDUCTION %%%%%%%%%%%%%%%%%%%%
213: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
214: \section{Observations and Data Reduction}
215:
216: We use proprietary \rxte\ data of \ngc\ that was regularly observed for
217: $\sim$ 1000 s once every 4 days between 2005 March 4 and 2006 February 27.
218: The observations were carried out
219: with the Proportional Counter Array (PCA; Jahoda et al. 1996), and the
220: High-Energy X-Ray Timing Experiment (HEXTE; Rotschild et al. 1998) on \rxte.
221: Here we will consider only PCA data, because the signal-to-noise of the HEXTE
222: data is too low for a meaningful analysis.
223:
224: The PCA data were screened according to the following acceptance criteria: the
225: satellite was out of the South Atlantic Anomaly (SAA) for at least 30 minutes,
226: the Earth elevation angle was $\geq 10^{\circ}$, the offset from the nominal
227: optical position was $\leq 0^{\circ}\!\!.02$, and the parameter ELECTRON-2 was
228: $\leq 0.1$. The last criterion removes data with high particle background rates
229: in the Proportional Counter Units (PCUs). The PCA background spectra and light
230: curves were determined using the ${\rm L}7-240$ model developed at the \rxte\
231: Guest Observer Facility (GOF) and implemented by the program {\tt pcabackest}
232: v.3.0. This model is appropriate for ``faint'' sources, i.e., those
233: producing count rates less than 40 ${\rm s^{-1}~PCU^{-1}}$.
234:
235: All the above tasks were carried out using the {\tt FTOOLS} v.6.2 software
236: package and with the help of the \verb+REX+ script provided by the \rxte\
237: GOF. Data were initially extracted with 16 s time resolution and subsequently
238: re-binned at different bin widths depending on the application. The current
239: temporal analysis is restricted to PCA, STANDARD-2 mode, 2--12.5 keV, Layer 1
240: data, because that is where the PCA is best calibrated and most sensitive. PCUs
241: 0 and 2 were turned on throughout the monitoring campaign. However, since the
242: propane layer on PCU0 was damaged in May 2000, causing a systematic increase
243: of the background, we conservatively use only PCU2 for our analysis (see
244: below). All quoted count rates are therefore for one PCU.
245:
246: The spectral analysis of PCA data was performed using the {\tt XSPEC v.12.3.1}
247: software package (Arnaud 1996). We used PCA response matrices and effective
248: area curves created specifically for the individual observations by the program
249: {\tt pcarsp}, taking into account the evolution of the detector properties.
250: All the spectra were re-binned so that each bin contained enough counts for the
251: $\chi^2$ statistic to be valid. Fits were performed in the energy range
252: 2.5--12.5 keV, where the signal-to-noise ratio is the highest.
253:
254: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
255: %%%%%%%%%%%% 3. X-RAY LIGHT CURVES %%%%%%%%%%%%%%%%%%%%%%
256: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
257: \section{The X-ray Light Curves}
258: Although \ngc\ is generally considered an X-ray bright source -- its
259: average flux is of the order of $4\times 10^{-12}$ \flux\ with a corresponding
260: luminosity L$_{\rm 2-10~keV} \sim 5\times10^{42}$ \lum\ that is nearly a
261: factor 4
262: larger than the typical values observed in low-power radio galaxies
263: (Donato et al. 2004) -- it is rather weak for the \rxte\ capabilities.
264: Therefore, before starting a detailed analysis of temporal
265: properties, it is necessary to demonstrate that the variability observed
266: cannot be ascribed to uncertainties in the \rxte\ background or to other
267: artifacts. To this end, we have performed the following
268: test: We have compared the background-subtracted light curves obtained
269: using PCU2 layer 1 and PCU2 layer 3. Since the genuine signal in layer
270: 3 is quite small, its light curve can be used as a proxy to check how
271: well the background model works. If the latter light curve is
272: significantly variable with a pattern similar to the one produced
273: using layer 1, then the variability is simply due to un-modeled
274: variations of the background. Conversely, if the PCU2 layer 3 light
275: curve does not show any pronounced variability or if the flux changes
276: are uncorrelated with those observed in the layer 1 light curve, we
277: can safely conclude that the variability detected in \ngc\
278: is real.
279:
280: The two light curves in the 2--10 keV range (where the
281: background PCA model is better parameterized; see Jahoda et al. 2006
282: for more details) are shown in Figure~\ref{figure:fig1}. A visual inspection
283: of this figure suggests that the variability in the layer 1 light curve is
284: much
285: more pronounced than the one observed in layer 3. Indeed, the mean count rate
286: of layer 3 is consistent with zero, indicating that on long timescales
287: the background model works adequately. However, statistically speaking both
288: light curves are considered variable: $\chi^2=$ 229.5 (layer 1) and 128.3
289: (layer 3) for 78 degrees of freedom (hereafter dof). On the other hand,
290: a formal analysis based on the excess variance (i.e. the variance corrected
291: for statistical errors), indicates that variability associated with the
292: layer 1 light curve is more than one order of magnitude larger than that
293: associated with layer 3 ($\sigma_{\rm xs}= 4\times10^{-3}$ and
294: $3\times10^{-4}~
295: {\rm s^{-2}}$, respectively). Further support to the fact that the
296: variability associated with layer 1 is genuine comes from a correlation
297: analysis: When layer 3 is plotted versus the layer 1 count rate (see
298: Figure~\ref{figure:fig2}), no correlation is observed, as confirmed by
299: a least square linear fit analysis, $y=0.002(\pm0.015)-0.005(\pm0.048)x$.
300: We therefore conclude that most of the count rate changes observed in layer 1
301: are associated with genuine intrinsic variations of the X-ray source in \ngc.
302:
303: \subsection{PCU0 versus PCU2}
304: In order to maximize the signal-to-noise (S/N) ratio of the light curves,
305: one can combine the 2 PCUs at work during
306: the monitoring campaign (i.e., PCU0 and PCU2), provided that they are
307: consistent with each other. In particular, it is necessary to test
308: whether PCU0, partially damaged since May 2000, is compatible with PCU2.
309: Figure~\ref{figure:fig3} shows the background-subtracted, 2--12 keV light
310: curves of PCU2 (top panel), PCU0 (middle panel), each with superimposed the
311: light curve of the other PCU (represented by continuous lines in top and
312: middle panels), and their ratio (bottom panel)
313: with a time bin of 4 days. A visual inspection of this
314: figure suggests that the PCU0 and PCU2 light curves are not fully consistent
315: with each other.
316: Indeed, a formal check based on a $\chi^2$ test indicates that the ratio
317: PCU0/PCU2 is not consistent with the hypothesis of constancy ($\chi^2=160.5$
318: for 54 dof). Since the propane layer on PCU0 was
319: damaged a few years ago, we decided to work with the PCU2 data only. For
320: completeness, we have also performed the same data analysis on the PCU0 data,
321: and the results are in general agreement with those from the PCU2.
322:
323: Whereas for the temporal analysis the PCU0 data need to be necessarily
324: excluded because they appear to be inconsistent with the PCU2 data, PCU0
325: spectral data should not be a priori ruled out for an analysis averaged over
326: two-month periods. Unfortunately, also the 2-month averaged PCU0 spectra
327: appear to be at odds with the PCU2 spectral results, with unphysically large
328: changes for the photon index (with a seesaw trend between $\Gamma\sim 2$ and
329: $\Gamma\sim 3$) and very large uncertainties.
330: Therefore, also for the spectral analysis (see $\S5$) we use only PCU2 data.
331:
332:
333: Using data from PCU2 only, we constructed background subtracted light curves
334: in two energy bands, namely the ``soft" (2.5--5 keV) and the ``hard"
335: (5--12 keV) band. We
336: show them in Figure~\ref{figure:fig4}, together with the hardness ratio
337: (hard/soft) plotted versus time.
338: Time bins are 4 days. The soft and hard band light curves are
339: significantly variable ($\chi^2=113.1/78$ and $\chi^2=175.6/78$, respectively)
340: and appear to be qualitatively similar. This
341: is formally confirmed by the hardness ratio light curve, which appears to be
342: constant ($\chi^2=70.7/78$).
343:
344: We conclude
345: that the flux variations observed in \ngc\
346: on timescales of weeks-months are not associated with spectral variations.
347:
348: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
349: %%%%%%%%%%%% 4. SPECTRAL VARIABILITY %%%%%%%%%%%%%%%%%%%%
350: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
351: \section{Spectral variability: model-independent analysis}
352: For this study we use simple methods such as hardness ratio versus count rate
353: plots and the fractional variability versus energy plots. These can provide
354: useful
355: information without any a priori assumption regarding the shape of the X-ray
356: continuum spectrum. Thus, the results from the study of these plots
357: can be considered as ``model-independent".
358:
359: Figure~\ref{figure:fig5} shows the Hard/Soft X-ray color
360: (5--12 keV)/(2.5--5 keV) plotted versus the
361: total count rate (2.5--12 keV) for un-binned (smaller, gray symbols) and
362: binned data (larger, darker symbols), respectively.
363: A visual inspection of this figure suggests the presence of a positive
364: trend between $HR$ and the count rate (i.e., the spectrum hardens
365: as the flux increases), although within a large scatter.
366: This positive trend is apparently confirmed by the linear fit of the data,
367: $y=(0.2\pm0.2)+(2.3\pm0.5)x$ with $\chi^2/dof=43.9/77$, which suggests
368: the presence of a positive correlation at 4.6$\sigma$ confidence level. This
369: analysis was performed
370: using the routine \verb+fitexy+ (Press et al. 1997) that accounts
371: for the errors not only on the y-axis but along the x-axis as well.
372:
373: Better insight into the presence of correlations between $HR$ and total count
374: rate can be obtained by investigating the binned data. To this end, we have
375: binned the data, shown in Fig.~\ref{figure:fig5}, using count rate bins of
376: fixed size (0.05 counts/s), and
377: computed the weighted mean (along both the y and x-axis) of all the points
378: which fall into a bin. The error on the weighted mean is computed
379: following Bevington's prescriptions (Bevington 1969). Only bins with at least
380: 5 data points have been plotted in Figure~\ref{figure:fig5} and considered
381: for the linear fit. In this case, the best linear fit,
382: $y=(0.4\pm0.2)+(1.6\pm0.6)$ with $\chi^2/dof=1.44/4$, (shown in
383: Fig.~\ref{figure:fig5} with its
384: 1$\sigma$ uncertainties), indicates that,
385: if a few outliers with large count rate (and error-bar) are neglected,
386: the significance of a positive correlation reduces to 2.67$\sigma$. This is
387: slightly lower than the 3$\sigma$ level, which is generally accepted for
388: significant variations, but it still implies a positive correlation at
389: a confidence level of $\sim$ 99\%.
390:
391: Another simple way to quantify the variability properties of \ngc\, without
392: considering the time ordering of the values in the light curves, is based on
393: the fractional variability parameter $F_{\rm var}$. This is a commonly
394: used measure of the intrinsic
395: variability amplitude relative to the mean count rate, corrected for the effect
396: of random errors, i.e., \begin{equation} F_{\rm
397: var}={(\sigma^2-\Delta^2)^{1/2}\over\langle r\rangle} \end{equation}
398: \noindent
399: where $\sigma^2$ is the variance, $\langle r\rangle$ the unweighted mean count
400: rate, and $\Delta^2$ the mean square value of the uncertainties associated with
401: each individual count rate. The error on $F_{\rm var}$ has been estimated following Vaughan et al. (2003).
402: We computed $F_{\rm var}$ on selected energy bands, with
403: mean count rates similar and sufficiently
404: high.
405:
406: The results are plotted in Figure~\ref{figure:fig6} and
407: suggest that, over the 3--8 keV energy range, a weak positive trend seems
408: to be present. However, the large uncertainties on $F_{\rm var}$ due to the
409: low count rate in the narrow energy-selected bands hamper this kind of
410: analysis. Indeed, statistically speaking, the trend shown in
411: Figure~\ref{figure:fig6} is consistent with the hypothesis of $F_{\rm var}$
412: being constant in the energy range probed by \rxte. If only two
413: broader energy bands are used for this analysis, the resulting fractional
414: variability in the hard energy band 5--12 keV, $F_{\rm var,hard}=(20\pm3)\%$
415: appears to be larger than the value, $F_{\rm var,soft}=(10\pm5)\%$, obtained
416: in the soft band 2.5--5 keV. However, the $F_{\rm var,hard}-
417: F_{\rm var,soft}=(10\pm6)\%$ difference is significant at the 1.7$\sigma$ level
418: only.
419:
420:
421: In summary, the results from the study of the $HR-ct$ plots show
422: evidence for a positive correlation between hardness ratio and
423: total count rate. Similarly, an analysis of the fractional variability
424: suggests that $F_{\rm var}$ is more pronounced in the hard energy band
425: than in the soft one.
426: However, due to the rather limited signal-to-noise ratio of our data these
427: results are significant at just the $2.7$ and $1.7\sigma$ level, respectively.
428:
429:
430:
431: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
432: %%%%%%%%%%%% 5. SPECTRAL ANALYSIS %%%%%%%%%%%%%%%%%%%%
433: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
434: \section{Spectral Analysis}
435:
436: \subsection{Time-resolved Spectroscopy}
437: Given that the data consist of short snapshots spanning a long temporal
438: baseline,
439: they are in principle well suited for monitoring the spectral variability of
440: \ngc. However, due to the limited S/N of the data, the spectral slope
441: $\Gamma$, measured
442: from spectra of individual observations, cannot be adequately constrained
443: and hence the spectral variability cannot be investigated in such a way. Indeed,
444: if we plot the values of $\Gamma$ versus time, no variations are detected due
445: to the large errors on $\Gamma$. This is fully consistent with the results
446: from the $HR$-time plot described in $\S4$. In order to increase the S/N
447: and investigate the presence of possible spectral variations, we use spectra
448: averaged over two-month intervals. This
449: choice is a trade-off between the necessity of accumulate sufficient counts
450: for a reliable spectral analysis and the need to use limited temporal intervals
451: to minimize the effects of the slow drift in the detector gain.
452:
453:
454:
455: We fitted each two-month spectrum with a simple power-law (PL) model
456: absorbed by Galactic $N_{\rm H}$ ($5.65\times10^{20}{\rm~cm^{-2}}$).
457: The model fits all the data reasonably well, as indicated by
458: Fig.~\ref{figure:fig7} that shows a typical spectrum fitted with a
459: simple power law.
460: The best-fit results are listed in Table 1
461: and can be summarized as follows. A simple PL model provides
462: an acceptable parametrization for all spectra. The photon indices are all
463: rather steep ($\Gamma\sim2.5$) and consistent with each other within the
464: errors. In other words, our results suggest that the source's spectrum
465: does not vary significantly on timescales longer than two months. The
466: weighted spectral slope mean is $2.5\pm0.1$.
467: Adding a Gaussian line
468: at 6.4 keV to the PL continuum model does not improve the fit significantly
469: in any of the six spectra, but the 90\% confidence upper limits on the
470: equivalent
471: width are relatively high (EW$\sim$200-700 eV).
472:
473: In order to better constrain $\Gamma$ and the line EW,
474: we have tried a comparison between the first and the last 6 months
475: of monitoring campaign, by fitting together 3 two-month spectra at a time.
476: The results -- $F_{\rm 2-10~keV}=4.3\times 10^{-12}$ \flux,
477: $\Gamma=2.53\pm0.18$, $EW < 212$ eV during the first half, and
478: $F_{\rm 2-10~keV}=4.5\times 10^{-12}$ \flux,
479: $\Gamma=2.46\pm0.17$, $EW < 144$ eV during the second half --
480: are fully consistent with each other.
481:
482: Since there are no indications for long-term spectral variability, we
483: have fitted the 6 two-month spectra together.
484: This yielded $\Gamma=2.5\pm0.1$ and a significantly smaller upper limit
485: on the line strength:
486: EW$<$104 eV. In the 2--10 keV energy
487: band, we obtained
488: an average absorbed flux of 4.4$\times 10^{-12}$ \flux, and a corresponding
489: intrinsic luminosity of $6.3\times 10^{42}$ \lum,
490: assuming $H_0=71{\rm~km~s^{-1}~Mpc^{-1}}$, $\Omega_\Lambda=0.73$ and
491: $\Omega_{\rm M}=0.27$ (Bennet et al. 2003).
492:
493: \subsection{Flux-selected Spectroscopy}
494: In order to verify the presence of a direct correlation between $HR$ and
495: count rate derived in $\S4$, we performed a flux-selected spectral analysis.
496: To this end we divided the 94 individual spectra into 5 bins according to
497: their average count rate (namely, $<0.25,~,0.25-0.30,~0.30-0.35,~0.35-0.40,
498: ~{\rm and}~>0.40$ c/s). In each bin, the individual spectra were fitted
499: simultaneously with a PL model (absorbed by Galactic $N_{\rm H}$) with their
500: photon indices linked together and their respective normalizations free to
501: vary.
502:
503: The results are summarized in Table 2, where the reported errors on
504: $\Gamma$ and flux are respectively 1$\sigma$ and $\sigma/\sqrt N$,
505: with $N$ being the number of individual spectra per bin. The upper limits
506: on EW correspond to the 90\% confidence level.
507: Table 2 indicates that the 2.5--12.5
508: keV spectra harden as the average flux increases. This is clearly shown in
509: Figure~\ref{figure:fig8}, where the values of $\Gamma$ for each count rate bin
510: have been plotted against their respective flux values. The dashed line
511: represents the best linear fit, $y=3.6\pm0.3-(0.24\pm0.08)x$, which reveals
512: that the inverse correlation is significant at 3-$\sigma$ level.
513:
514: For comparison, in Fig.~\ref{figure:fig8} we have also plotted the
515: values corresponding to the \asca, \sax, and \xmm\ observations. To this end,
516: we have used {\tt PIMMS} to convert the observed flux into the \rxte\ energy
517: range,
518: assuming the best-fit spectral parameters reported in Guainazzi et al.
519: (2003) for \sax\ and \asca, and Gliozzi et al. (2004) for \xmm. The \asca,
520: \sax, and \xmm\ photon
521: indices seem to follow the same inverse trend shown by \rxte, becoming flatter
522: as the flux
523: increases. However, their values appear to be significantly smaller than those
524: obtained from the flux-selected \rxte\ spectral analysis. This apparent
525: discrepancy may be probably reconciled (at least for the \sax\ and \xmm\
526: data) by bearing in mind that: 1) Past studies have shown that photon
527: indices measured with the \rxte\ PCA are systematically steeper than those
528: measured
529: by other X-ray satellites (e.g., Yaqoob 2003). 2) The error-bars for the \rxte\
530: data in Fig.~\ref{figure:fig8} are at 1$\sigma$ level. On the other hand,
531: it appears more difficult to reconcile the \asca\ value with the extrapolation
532: of the \rxte\ best-fit trend, suggesting that the source was in a different
533: physical state (for example, it lacked a jet contribution).
534:
535:
536: We have also added a Gaussian line to the PL continuum (see Table 2 last
537: column). Only for the bin with the lowest flux, the line is very marginally
538: significant (the value of $\chi^2$ decreases by 1.2 for one additional dof)
539: with EW$\sim$400 eV and large uncertainties.
540: For bins with larger flux values, only upper limits on EW can be derived.
541: These findings indicate that a \feka\ line is never statistically required,
542: although its presence cannot be completely ruled out when the source flux
543: is at the lowest level measured by \rxte.
544:
545: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
546: %%%%%%%%%%%% 6. DISCUSSION %%%%%%%%%%%%%%%%%%%%
547: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
548: \section{Discussion}
549:
550: We have undertaken the first X-ray monitoring study of the FR~I galaxy \ngc,
551: investigating the temporal and spectral variability as well as time-averaged
552: and flux-selected spectral results.
553:
554: By comparing the X-ray fluxes measured by different satellites over nearly
555: a decade of observations, it is clear that \ngc\ shows large flux changes
556: on long timescales.
557: For example,
558: in October 1994 \asca\ measured a flux of $\sim 1.4\times 10^{-12}$ \flux\
559: in the 2--10 keV energy band,
560: whereas \sax\ observed a flux of $\sim 4.7\times 10^{-12}$ \flux\ in July
561: 2001, and \xmm\ of $\sim 4\times 10^{-12}$ \flux\ in March 2002. The high
562: throughput of the EPIC pn camera aboard \xmm\ also revealed the presence of
563: low-amplitude flux changes (of the order of $\sim$5\%) on timescales of a
564: few ks.
565:
566: The variability behavior detected in the current \rxte\ observations is in
567: agreement with the
568: sparse evidence that had been gathered in the previous years, and show
569: conclusively that \ngc\ is a persistently variable source in X-rays. The
570: frequent \rxte\ observations, spread over a period of one year, indicate
571: that this
572: radio galaxy is characterized by persistent variability in the total
573: (2.5--12 keV),
574: soft (2.5--5 keV), and hard (5--12 keV) energy bands. Throughout the
575: observation, the flux appears to randomly change by a factor of $\sim$2
576: on timescales of a few weeks.
577:
578: Previous X-ray studies, mostly based on the time-averaged spectral
579: analysis, have revealed a dual behavior for \ngc: sometimes the source
580: appears to behave like a Seyfert galaxy with a strong \feka\ line (e.g.,
581: Turner et al. 1997; Sambruna et al. 1999; Gliozzi et al. 2004), whereas in
582: other occasions \ngc\ seems to be more consistent with a blazar-like
583: behavior, showing a featureless X-ray spectrum (e.g., Guainazzi et al. 2003).
584: The densely
585: sampled, year-long \rxte\ observations and the investigation of the flux and
586: spectral variability properties offer an alternative and model-independent
587: way to shed some light on this issue.
588:
589: \subsection{Evidence from the flux variability properties}
590:
591: Unlike the brightest blazars frequently monitored by
592: \rxte\ (e.g., Kataoka et al. 2001, Cui 2004, Xue \& Cui 2005), \ngc\ does
593: not show any prominent flare on any observable timescale, nor does it show
594: the large amplitude variability typically observed in several Seyfert-like
595: objects during yearly-long monitoring campaigns (e.g., Markowitz \& Edelson
596: 2001).
597:
598: On one hand, the apparent inconsistency with the large variability
599: observed in
600: Seyfert galaxies can be explained by the lower values of black hole mass
601: in the latter objects, which is typically one order of magnitude lower than
602: \ngc. On the other hand, the lack of prominent flares in the NGC6251 light
603: curve, which instead
604: characterize the blazar light curves, can be understood by keeping in mind
605: that the blazars monitored by RXTE are the brightest members of
606: this AGN class and that the observations are often triggered only during
607: their flaring activity. Nonetheless, blazar monitoring campaigns with
608: baselines covering several years reveal that also the brightest blazars
609: alternate prominent flaring activity with ``quiescent periods'' that are
610: characterized by moderate flux variations.
611: Indeed, prolonged periods of moderate variability have been detected in several
612: blazars (e.g., B\"ottcher et al. 2005; Marscher 2006).
613:
614: A model-independent study of the \rxte\ variability properties of the
615: prototypical blazar Mrk~501, similar to the one performed on \ngc\ in this
616: work, offers the best opportunity for a more quantitative comparison
617: of \ngc\ with a typical blazar behavior (Gliozzi et al. 2006).
618: For instance, Mrk 501, which showed a large outburst in 1997, underwent a
619: progressive decrease of
620: its activity in the following years, resulting in a lower mean count rate
621: accompanied by lower variability. Specifically, in 1999,
622: when Mrk 501 reached a minimal flux value, the light curve was characterized by
623: variations of the order of $\sim$2 on timescales of 20--30 days, which are
624: fully consistent with those detected in \ngc.
625:
626: Before proceeding further, it must be pointed out that the blazars regularly
627: monitored over the past years with \rxte\ (including Mrk~501) are basically
628: all High-peaked blazars
629: (HBLs), with the synchrotron component peaking in the X-ray range and a
630: second spectral component, generally attributed to inverse Compton scattering,
631: peaking
632: at TeV energies. On the other hand, in the blazar framework, the broadband SED
633: of \ngc\ is consistent Low-peaked blazars (LBLs), with the inverse Compton
634: component peaking in the X-rays (Chiaberge et al. 2003; Guainazzi et al. 2003).
635: As a consequence, a formal comparison of the X-ray properties of \ngc\ should
636: be in principle performed using the TeV properties of Mrk~501. However,
637: detailed studies
638: of X-ray and TeV emissions in HBLs have demonstrated the existence of a tight
639: correlation between these energy bands, indicating that the X-ray and TeV radiation follow the same variability trend (e.g., Fossati et al. 2004; Gliozzi
640: et al. 2006). Further support to this conclusion comes from recent
641: investigations of the TeV properties of Mrk~501 and Mrk~421 carried out
642: with the MAGIC and Whipple telescopes (Blazejowki et al. 2006;
643: Albert et al. 2007). These studies demonstrate that also in the $\gamma$-ray
644: energy band HBLs show the typical blazar spectral variability behavior
645: observed in X-rays, that is a spectral hardening when the
646: source brightens and a fractional variability more pronounced at
647: higher energies (e.g., Zhang et al. 1999; Fossati et al. 2000; Gliozzi et al.
648: 2006; Rebillot et al. 2006 and references therein). We therefore conclude
649: that it is appropriate to compare the X-ray properties of Mrk~501 with those
650: of \ngc.
651:
652:
653:
654: Another relevant finding from the flux variability analysis of \ngc\ is that
655: the fractional variability appears to be more pronounced in the hard than
656: in the soft energy band: $\Delta F_{\rm var}\equiv F_{\rm var,hard}-
657: F_{\rm var,soft}=(10\pm6)\%$. This result can be quantitatively compared
658: with similar studies carried out on Seyfert galaxies monitored with \rxte\
659: for several months or years (Markowitz \& Edelson 2004). Specifically, all
660: the Seyfert galaxies with long \rxte\ monitoring yield negative values of
661: $\Delta F_{\rm var}$ ranging between -3.5 and -24.8, with a mean
662: value of $-10.3\pm1.5$. On the other hand, a similar study, carried out
663: using \rxte\ monitoring data of the blazar Mrk~501 between 1997 and 2000,
664: yielded positive values of $\Delta F_{\rm var}$ ranging between 5 and
665: 42, with a mean value of $15\pm1$.
666:
667: We can therefore conclude that the limited amplitude of the \ngc\ flux
668: variability can be explained equally well in both Seyfert-like and
669: blazar-like scenarios. However, the fractional variability behavior
670: is inconsistent with the trend generally observed in Seyfert galaxies,
671: but fully consistent with the typical blazar behavior observed in Mrk~501.
672:
673: \subsection{Evidence from the spectral properties}
674: A comparative analysis of the
675: spectral variability of \ngc\ and the typical behaviors observed
676: in both radio-quiet AGNs (where the X-rays are thought to be produced by
677: Comptonization in the corona that is closely connected with the accretion
678: disk) and radio-loud jet-dominated AGNs (whose radiation over the entire
679: energy range is ascribed to jet emission) can help us understand whether
680: the X-ray radiation from \ngc\ is dominated by jet or accretion-related
681: emission.
682:
683: A flux-selected spectral analysis (in agreement with
684: the $HR-ct$ plot) has shown that the X-ray spectrum of \ngc\
685: hardens as the source brightens, following the linear correlation
686: $\Gamma\propto -0.24(\pm0.08)\times F_{\rm X}$. A direct comparison of
687: this result with a similar spectral study, carried out on 4 Seyfert galaxies
688: monitored with \rxte, indicate that the latter have always positive slopes
689: in the $\Gamma-F_{\rm X}$ plane ranging between 0.05 and 0.15 with a
690: mean value of $0.09\pm0.01$ (Papadakis et al. 2002). As a consequence,
691: the spectral variability behavior of \ngc\ appears to
692: be inconsistent
693: with the typical Seyfert-like trend. On the other hand,
694: the $\Gamma-F_{\rm X}$ slope measured for Mrk~501 during the weakly variable
695: period of 2000,
696: $-0.37\pm0.06$, appears to be fully consistent with the behavior of \ngc\
697: during the \rxte\ monitoring campaign.
698:
699: The existence of a Seyfert-like component suggested by previous X-ray studies
700: was essentially based on the presence of a strong \feka\ line. Unfortunately,
701: the low S/N spectra obtained during the \rxte\ monitoring coupled with the
702: poor spectral resolution of the \rxte\ PCA hampers a detailed investigation
703: of this issue. Indeed, if the 2-month averaged PCA spectra are fitted with
704: a model including a power law and a Gaussian line with spectral parameters
705: fixed at the best-fit values obtained during the \asca\ observation (we
706: conservatively assumed the best-fit parameters reported in Guainazzi et al.
707: 2003), the results are statistically
708: indistinguishable from the fits obtained
709: using a simple power law. This indicates that \rxte\ is unable to confirm or
710: refute the presence of a \feka\ line, and suggests that only a relatively long
711: exposure of \xmm\
712: with its superior capabilities is able to detect the presence of the
713: line when the average flux of \ngc\ is relatively high.
714:
715: \subsection{The origin of X-rays in \ngc}
716: The primary goal of this work is to investigate the origin of the X-rays
717: in \ngc\ and in particular to assess the role played by the putative jet.
718: At first sight, the possibility that the jet may dominate the radiative
719: output of a radio galaxy may be surprising, and even more so for NGC6251,
720: which has a Mpc radio jet (and hence appears forming a large viewing angle)
721: and Seyfert-like emission during the low flux state.
722: However, under specific circumstances, the jet-dominance hypothesis becomes
723: plausible. This is the case when the base of the jet is not well collimated
724: and the X-rays are produced by a part of the outflowing material that points
725: towards the observer. Indeed, this is the framework proposed to explain the
726: TeV emission detected in another radio galaxy M87 (Aharonian et al. 2006).
727: Alternatively, if the
728: base of the jet is tilted with respect to the large scale jet
729: and forms a small viewing angle, the possible dominance of the jet in the
730: 2--12 keV range can be naturally explained. In fact, this is the
731: scenario put forward by Jones \& Wehrle (2002) to explain
732: the large jet/counterjet brightness ratio on parsec scales inferred for
733: \ngc\ from VLBA observations.
734:
735: By comparing the spectral variability properties of \ngc\ with those of
736: radio-quiet
737: AGNs and blazar objects, we find that they are certainly inconsistent with
738: a typical Seyfert-like behavior but
739: fully consistent with blazars.
740: Combining pieces of information from the model-independent analysis with the
741: findings from the flux-selected spectral analysis, we are led to the conclusion
742: that, during the \rxte\ monitoring campaign, the bulk of the hard X-ray
743: radiation from \ngc\ was dominated by the emission from the unresolved base of
744: the jet. Nonetheless, the presence of a disk-corona component, detected in
745: previous observations with \asca\ and \xmm, cannot be ruled out: the upper
746: limits measured on the equivalent width of the \feka\ line are indeed fully
747: consistent with the values measured by \asca\ and \xmm, but
748: the low S/N of the spectra hampers a more quantitative analysis.
749:
750:
751: In conclusion, we can try to exploit the main results from this work
752: (i.e., the jet
753: dominance in the high flux state, with a possible contribution from
754: a Seyfert-like component emerging
755: at low flux values) to derive some constraints on the accretion process
756: at work in \ngc. Assuming that an accretion-related component is
757: always present in \ngc\ and dominates during the low flux state, we can use
758: the average flux measured in the low count rate bin (see Table 2) to compare
759: the corresponding 2--10 keV luminosity -- $L_{\rm X}=6.6\times 10^{42}$ \lum --
760: to the Eddington value readily derived from the black hole mass estimate.
761: The relatively high value derived, $L_{\rm X}/L_{\rm Edd} > 5\times 10^{-5}$,
762: confirms that \ngc\ is a bright FR~I galaxy that is close to the FR~I/FR~II
763: dividing line in terms of power of the central engine. However, it is not
764: possible to derive any firm conclusion on the nature of the accretion process
765: given the unknown contribution of the jet in the low-flux state.
766:
767:
768:
769:
770:
771:
772:
773:
774: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
775: %%%%%%%%%%%% 7. CONCLUSIONS %%%%%%%%%%%%%%%%%%%%
776: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
777: \section{Summary and Conclusions}
778:
779: We have used data from a year-long \rxte\ monitoring campaign to study the
780: spectral variability of \ngc\, following model-independent and spectral model
781: fitting methods. The main results can be summarized as follows:
782:
783: \begin{itemize}
784:
785: \item Throughout the monitoring campaign and especially during the last 4
786: months, \ngc\ was in a relatively high-flux state, with values of the
787: 2--10 keV absorbed flux comparable to that observed by \sax\ in 2001.
788:
789: \item The light curves show persistent
790: variations by a factor of 2 on timescales weeks/months in the total
791: (2.5--12 keV), soft (2.5--5 keV), and hard (5--12 keV) energy bands.
792:
793: \item The fractional variability, computed over the soft and hard energy
794: bands, reveals an enhanced variability at harder energies ($F_{\rm var}
795: \sim$20\%) compared to the soft band ($F_{\rm var}\sim$10\%).
796:
797: \item There is evidence of a positive trend in the $HR-ct$ plot (or,
798: analogously, of a negative trend in the $\Gamma$-flux plot); in other
799: words the spectrum hardens as the flux increases.
800:
801: \item The 2-month averaged spectra are well fitted by a power-law model,
802: with $\langle\Gamma\rangle\simeq2.5$ and no indication for an \feka\
803: line. Combining all the 2-month averaged spectra yielded EW$<$ 104 eV.
804: Only for the lowest flux-selected spectrum, there is marginal
805: evidence for a \feka\ line. However, the low S/N does not allows one to
806: put reliable constraints on the putative Seyfert-like component.
807: \end{itemize}
808:
809: The inconsistency of the spectral variability behavior of \ngc\
810: with the typical trend observed in Seyfert galaxies and the
811: similarity with blazars lead support to a jet-dominated scenario.
812: However, based on the \rxte\ observations,
813: a substantial contribution from a disk-corona system cannot be ruled out.
814:
815: \begin{acknowledgements}
816: We thank the referee for the
817: comments and suggestions that improved the clarity of the paper.
818: MG acknowledges support by the RXTE Guest Investigator Program
819: under NASA grant 200858. Funds from the NASA LTSA grant
820: NAG5-10708 are also gratefully acknowledged.
821: \end{acknowledgements}
822:
823:
824: \begin{thebibliography}{}
825:
826: \bibitem[Aharonian et al. 2006]{ahar06} Aharonian, F., et al. 2006, Science,
827: 314, 1424
828:
829: \bibitem[Albert et al. 2007]{alb07} Albert, J., et al. 2007, ApJ, 669, 862
830:
831: \bibitem[Arnaud 1996]{arn96} Arnaud, K. 1996, in ASP Conf. Ser. 101, Astronomical
832: Data Analysis Software and Systems V, ed. G. Jacoby \& J. Barnes (San
833: Francisco: ASP), 17
834:
835:
836: \bibitem[Baum et al. 1995]{baum95} Baum,S.A., Zirbel, E.L., \& O'Dea, C.P. 1995,
837: ApJ, 451, 88
838:
839: \bibitem[Bennet et al. 2003]{ben03} Bennet, C.L. et al. 2003, ApJS, 148, 1
840:
841: \bibitem[Bevington 1969]{bevi69} Bevington, P.R. 1969. Data Reduction and Error Analysis for the Physical Sciences. McGraw-Hill, New York
842:
843: \bibitem[Birkinshaw \& Worrall 1993]{Bir} Birkinshaw, M., \& Worrall, D.M. 1993, ApJ, 412, 568
844:
845: \bibitem[Blazejowski et al. 2005]{blaz05} Blazejowski, M., et al. 2005, ApJ,
846: 630, 130
847:
848: \bibitem[B\"ottcher et al. 2005]{boet05} B\"ottcher, M., Harvey, J., Joshi, M. et al. 2005, ApJ, 631, 169
849:
850: \bibitem[2003]{Chiab} Chiaberge, M., Gilli, R., Capetti, A., \& Macchetto, F.D.
851: 2003, ApJ, 597, 166
852:
853:
854: \bibitem[Cui 2004]{cui04} Cui, W. 2004, ApJ, 605, 662
855:
856: \bibitem[Donato et al. 2004]{} Donato, D., Sambruna, R.M., \& Gliozzi, M. 2004,
857: ApJ, 617, 915
858:
859: \bibitem[Evans et al. 2005]{eva05} Evans, D.A., Hardcastle, M.J., Croston, J.H., Worrall, D.M.,
860: \& Birkinshaw, M. 2005, MNRAS, 359, 363
861:
862: \bibitem[Fanaroff \& Riley 1974] {fana74} Fanaroff, B.L. \& Riley, J.M. 1974,
863: MNRAS, 167, 31
864:
865:
866: \bibitem[1999]{Fer} Ferrarese, L., \& Ford, H.C. 1999, ApJ, 515, 583
867:
868: \bibitem[Fossati et al. 2000]{foss00} Fossati, G., et al. 2000, ApJ, 541, 153
869:
870: \bibitem[Fossati et al. 2004]{foss04} Fossati, G., Buckley, J., Edelson, R.A.,
871: Horns, D., \& Jordan, M. 2004, NAR, 48, 419
872:
873: \bibitem[Ghisellini \& Celotti 2001]{ghis01} Ghisellini, G. \& Celotti, A. 2001,
874: A\&A, 379, L1
875:
876:
877: \bibitem[Gliozzi et al. 2003]{glioz03} Gliozzi M., Sambruna, R.M., \&
878: Eracleous, M. 2003, ApJ, 584, 176
879:
880: \bibitem[Gliozzi et al. 2004]{glioz04} Gliozzi M., Sambruna, R.M., Brandt, W.N., Mushotzky, R., \&
881: Eracleous, M. 2004, A\&A, 413, 139
882:
883:
884: \bibitem[Gliozzi et al. 2006]{glioz06} Gliozzi M., et al. 2006, ApJ, 646, 61
885:
886: \bibitem[Gliozzi et al. 2007]{glioz07} Gliozzi M., Sambruna, R.M.,
887: Eracleous, M., Yaqoob, T. 2007, ApJ, 664, 88
888:
889: \bibitem[Gonz\'alez-Mart\'in et al. 2006]{gonz06} Gonz\'alez-Mart\'in, O., Masegosa, J.,
890: M\'arquez, I., Guerrero, M.A., \& Dulzin-Hacyan, D. 2006 A\&A, 460, 45
891:
892: \bibitem[Guainazzi et al. 2003]{Guai03} Guainazzi, M., Grandi, P., Comastri, A., \& Matt, G.
893: 2003, A\&A, 410, 131
894:
895: \bibitem[2000]{har} Hardcastle, M.J., \& Worrall, D.M. 2000, MNRAS, 314, 359
896:
897: \bibitem[Ho 1999]{ho99} Ho, L. 1999, ApJ 516, 672
898:
899: \bibitem[Jahoda et al. 1996]{jaho96} Jahoda, K., Swank, J., Giles, A.B., et al. 1996,
900: Proc.SPIE, 2808, 59
901:
902: \bibitem[Jones \& Wehrle 2002]{jon02} Jones, D.L. \& Wehrle, A.E. 2002, ApJ,
903: 580, 114
904:
905: \bibitem[Kataoka et al. 2001]{kata01} Kataoka, J., et al. 2001, ApJ, 560, 659
906:
907:
908: \bibitem[Markowitz \& Edelson 2001]{marko01} Markowitz, A. \& Edelson, R. 2001, ApJ, 547, 684
909:
910: %\bibitem[Markowitz et al. 2003]{marko03} Markowitz, A., Edelson, R., \&2 Vaughan, S. 2003, ApJ, 598, %935
911:
912: \bibitem[Markowitz \& Edelson 2004]{marko04} Markowitz, A. \& Edelson, R. 2004, ApJ, 617, 939
913:
914:
915: \bibitem[Marscher 2006]{marsch06} Marscher, A.P. 2006, ChJAS, 6, 262
916:
917: \bibitem[2002]{meli} Melia, F., Liu, S., \& Fatuzzo, M. 2002, ApJ 567, 811
918:
919:
920: \bibitem[Mukherjee et al. 2002]{Muk02} Mukherjee, R., Halpern, J., Mirabal, N.
921: , \& Gotthelf, E.V. 2002, ApJ, 574, 693
922:
923:
924: \bibitem[1989]{owe} Owen, F.N., \& Laing R.A. 1989, MNRAS, 238, 357
925:
926: \bibitem[Papadakis et al. 2002]{papa02} Papadakis, I.E., Petrucci, P.O., Maraschi, L., et al. 2002, ApJ, 573, 92
927:
928:
929: \bibitem[Press et al. 1997]{press97} Press, W.H., Teukolsky, S.A., Vetterling, W.T., \& Flannery, B.P. 1997, Numerical Recipes (Cambridge: Cambridge Univ. Press)
930:
931: \bibitem[Rebittlot et al. 2006]{reb06} Rebillot, P.F., et al. 2006, ApJ, 641, 740
932:
933: \bibitem[Reynolds et al. 1996]{reyn99} Reynolds, C.S., Di Matteo, T., Fabian, A.C.,
934: Hwang, U., \& Canizares, C.R. 1996, MNRAS, 283, 111;
935:
936:
937: \bibitem[Rotschild et al. 1998]{rot98} Rotschild, R.E., Blanco, P.R., Gruber, D.E., et
938: al. 1998, ApJ, 496, 538
939:
940: \bibitem[Sambruna et al. 1999]{Sam1} Sambruna, R.M., Eracleous, M., \& Mushotzky, R. 1999, ApJ, 526, 60
941:
942:
943: \bibitem[Shuder \& Osterbrock 1981]{Shu} Shuder,J.M. \& Osterbrock, D.E. 1981, ApJ, 250, 55
944:
945:
946: \bibitem[Turner et al. 1997]{Tur} Turner, T.J., George, I.M., Nandra, K., \& Mushotzky, R.F.
947: 1997, ApJS, 113, 23
948:
949:
950: \bibitem[Vaughan et al. 2003]{vaugh03} Vaughan, S., Edelson, R., Warwick, R.S.,
951: \& Uttley, P. 2003, MNRAS, 345, 1271
952:
953:
954:
955: \bibitem[Xue \& Cui 2005]{xue05} Xue, Y. \& Cui, W. 2005, ApJ, 622, 160
956:
957: \bibitem[Yaqoob et al. 2003]{yaq03} Yaqoob, T., et al. 2003, ApJ, 546, 759
958:
959: \bibitem[Zhang et al. 1999]{zhang99} Zhang, Y.H., et al. 1999, ApJ, 527, 719
960:
961:
962: \bibitem[Zirbel \& Baum et al. 1995]{zirb95} Baum,S.A. \& Zirbel, E.L. 1995, ApJ, 448, 521
963:
964: \end{thebibliography}
965:
966: %%--------------FIG1-------lc layer1 and layer3-------------------------
967: \begin{figure}
968: \begin{center}
969: \includegraphics[bb=85 15 475 405,clip=,angle=0,width=9cm]{f1.eps}
970: \end{center}
971: \caption{Top panel: \rxte\ PCA light curve in the 2--10 keV energy
972: band, using PCU2 layer 1. Bottom panel: PCU2 layer 3 light curve.
973: Time bins are 4 days. The dashed lines are the average count rate
974: level.
975: }
976: \label{figure:fig1}
977: \end{figure}
978:
979: %%--------------FIG2------layer1 vs layer3-------------------------
980: \begin{figure}
981: \begin{center}
982: \includegraphics[bb=30 30 360 300,clip=,angle=0,width=9cm]{f2.eps}
983: \end{center}
984: \caption{Layer 3 plotted versus the layer 1 count rate.
985: The dashed line represents the best linear fit.}
986: \label{figure:fig2}
987: \end{figure}
988:
989: %%--------------FIG3-------PCU0 vs PCU2-------------------------
990: \begin{figure}
991: \begin{center}
992: \includegraphics[bb=100 25 430 460,clip=,angle=0,width=12cm]{f3.eps}
993: \end{center}
994: \caption{Top panel: PCU2 light curve of \ngc\ in the 2--10 keV range; the
995: solid line represents the PCU0 light curve.
996: Middle panel: PCU0 light curve of \ngc\ in the 2--12 keV range; the
997: solid line represents the PCU2 light curve.
998: Bottom panel: PCU0/PCU2 light curve; the dashed line represents the average
999: value of the ratio PCU0/PCU2. Time bins are 4 days.}
1000: \label{figure:fig3}
1001: \end{figure}
1002:
1003: %%--------------FIG4-------X-ray soft, hard, HR lc-------------------------
1004: \begin{figure}
1005: \begin{center}
1006: \includegraphics[bb=105 25 425 455,clip=,angle=0,width=12.cm]{f4.eps}
1007: \caption{Top panel: Soft (2.5--5 keV) light curve of \ngc\ using PCU2 data only.
1008: Middle panel: Hard (5--12 keV) light curve.
1009: Bottom panel: Hardness Hard/Soft ratio light curve; the solid line represents the
1010: average value of the hardness ratio. Time bins are 4 days.}
1011: \label{figure:fig4}
1012: \end{center}
1013: \end{figure}
1014:
1015: %%--------------FIG5-------HR-ct-------------------------
1016: \begin{figure}
1017: \begin{center}
1018: \includegraphics[bb=45 32 355 300,clip=,angle=0,width=9.cm]{f5.eps}
1019: \caption{Hardness ratio (5--15 keV/2.5--5 keV) plotted versus the total count
1020: rate. The gray (light blue in color) small symbols correspond to the un-binned
1021: data points, whereas the darker, larger symbols refer to the weighted mean of
1022: data points in bins of fixed size of 0.05 c/s along the x-axis.
1023: The continuous line represents the best linear fit from the binned data,
1024: whereas the dashed lines are the $\pm 1\sigma$ departure from the best fit. }
1025: \label{figure:fig5}
1026: \end{center}
1027: \end{figure}
1028:
1029: %%--------------FIG6-------fractional variability vs Energy PCU2------------
1030: \begin{figure}
1031: \centering
1032: \includegraphics[bb=65 35 345 290,clip=,angle=0,width=9.cm]{f6.eps}
1033: \caption{Fractional variability amplitude as a function of
1034: energy for \ngc.
1035: The error-bars along the x axis simply represent
1036: the energy band width. The error bars along the y axis are computed following
1037: Vaughan et al. 2003.
1038: }
1039: \label{figure:fig6}
1040: \end{figure}
1041:
1042: %%--------------FIG7-------2-month spectrum-------------------------
1043: \begin{figure}
1044: \includegraphics[bb=65 5 575 715,clip=,angle=-90,width=16.cm]{f7.eps}
1045: \caption{PCA spectrum of \ngc\ during the third 2-month interval obtained
1046: using PCU2 data only. The model used is a simple power law absorbed by
1047: Galactic $N_{\rm H}$.}
1048: \label{figure:fig7}
1049: \end{figure}
1050:
1051:
1052: %%--------------FIG8-------Gamma vs flux-------------------------
1053: \begin{figure}
1054: \begin{center}
1055: \includegraphics[bb=45 32 355 300,clip=,angle=0,width=9cm]{f8.eps}
1056: \end{center}
1057: \caption{$\Gamma$, obtained from the spectral analysis of flux-selected
1058: intervals, plotted against the average flux in the 2.5-12.5 keV band.
1059: The error-bars on $\Gamma$ are 1-$\sigma$ errors.
1060: The dashed line represents the best linear fit $y=3.6\pm0.3-(0.24\pm0.08)x$.
1061: For comparison, we have also plotted the
1062: values corresponding to the \asca, \sax, and \xmm\ observations. To this end,
1063: we have used {\tt PIMMS} to convert the observed flux into the \rxte\ energy
1064: range,
1065: assuming the best-fit spectral parameters reported in Guainazzi et al.
1066: (2003) and Gliozzi et al. (2004).}
1067: \label{figure:fig8}
1068: \end{figure}
1069:
1070:
1071: \input{tab1}
1072: \input{tab2}
1073:
1074: \end{document}
1075:
1076: