astro-ph0309799/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass[letterpaper]{article}
3: 
4: \usepackage{emulateapj}
5: \usepackage{onecolfloat}
6: \usepackage{apjfonts}
7: \usepackage{amsmath}
8: \usepackage{natbib}
9: 
10: %\textheight=23.2cm
11: 
12: 
13: \usepackage{graphicx}
14: \usepackage[figuresright]{rotating} %for tables, maybe removed later
15: \usepackage{float}       %% Float Control
16: \usepackage{afterpage}   %% Used to flush out floats
17: %\usepackage{scalefnt}    %% More font choices for tables than default
18: 
19: %%\usepackage{emulateapj5}
20: %%\usepackage{portland}
21: %% /*******************************************************************
22: %% ** Preamble                                                       **
23: %% *******************************************************************/
24: %%\include{preamble}
25: \newcommand {\IN}{$\Gamma$}
26: \newcommand{\rxte}{\emph{RXTE}}
27: \newcommand{\wsim}{\ensuremath{\sim}}
28: \newcommand{\xf}{XTE~J1550$-$564}
29: \newcommand{\xs}{XTE~J1650$-$500}
30: \newcommand{\gs}{GRO~J1655$-$40}
31: \newcommand{\gx}{GX~339$-$4}
32: \newcommand{\fu}{4U~1630$-$47}
33: \newcommand{\reso}{XTE~J1748$-$288}
34: \newcommand{\maso}{XTE~J1859$+$226}
35: \newcommand{\ind}{$\Gamma$}
36: \newcommand{\tin}{$T_{in}$}
37: 
38: \begin{document}
39: 
40: %% /*******************************************************************
41: %% ** The Header                                                     **
42: %% *******************************************************************/
43: 
44: 
45: \twocolumn[
46: \title{A close look at the state transitions of Galactic black hole transients during outburst decay}
47: 
48: \author{E. Kalemci\altaffilmark{1},
49:  J. A. Tomsick\altaffilmark{2},
50:  R. E. Rothschild\altaffilmark{2},
51:  K. Pottschmidt\altaffilmark{3,4},
52:  P. Kaaret\altaffilmark{5}
53: }
54: 
55: \altaffiltext{1}{Space Sciences Laboratory, 7 Gauss Way, University of 
56: California, Berkeley, CA, 94720-7450, USA}
57: 
58: \altaffiltext{2}{Center for Astrophysics and Space Sciences, Code
59: 0424, University of California at San Diego, La Jolla, CA,
60: 92093-0424, USA}
61: 
62: \altaffiltext{3}{Max-Planck-Institut f\"ur Extraterrestrische Physik, Giessenbachstr. 1, 85748 Garching, Germany}
63: 
64: \altaffiltext{4}{\emph{INTEGRAL} Science Data Centre, Chemin d'\'Ecogia 16, 1290 Versoix, Switzerland}
65: 
66: \altaffiltext{5}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
67:  Cambridge, CA, 02138, USA}
68: 
69: %% /*******************************************************************
70: %% ** The Abstract                                                   **
71: %% *******************************************************************/
72: 
73: \begin{abstract}
74: 
75: We characterize the evolution of spectral and temporal properties of several 
76: Galactic black hole transients during outburst decay using the data from
77: well sampled PCA/\rxte\ observations close to the transition to the low/hard 
78: state. We find several global patterns of evolution for spectral and temporal
79:  parameters before, during, and after the transition. We show that the changes
80: in temporal properties (sudden increase or decrease in the rms amplitude of 
81: variability) are much sharper than the changes in the spectral properties, and
82: it is much easier to identify a state transition with the temporal properties.
83:  The spectral index shows a drop 3-5 days before the transition for some of our 
84: sources. The ratio of the power-law flux to the total flux in the 3-25 keV band 
85: increases close to the transition, which may mean that the system must be 
86: dominated by the coronal emission for the transition to occur. We also 
87: show that the power-law flux shows a sharp change along with the temporal 
88: properties during the transitions which may indicate a threshold transition 
89: volume for the corona. The evolution of the spectral and temporal properties 
90: after the transition is consistent with the idea that the inner accretion disk 
91: moves away from the black hole. Based on the evolution of spectral and 
92: temporal parameters and changes during the transitions, we discuss possible 
93: scenarios of how the transition is happening.
94: 
95: \end{abstract}
96: 
97: \keywords{black hole physics -- X-rays:stars}
98: 
99: 
100: ] % twocolumn
101: 
102: %% /*******************************************************************
103: %% ** Introduction                                                   **
104: %% *******************************************************************/
105: 
106: \section{Introduction}\label{sec:intro}
107: 
108: X-ray observations of Galactic black holes (GBH) indicate that they are found 
109: in several distinct spectral states \citep{Tanaka95,McClintock03}. These 
110: states are generally determined by the relative strength of two different 
111: emission components: soft blackbody-like radiation from an optically thick,
112:  geometrically thin accretion disk and a harder component showing a power-law 
113: spectrum believed to originate from Compton upscattering of soft seed photons 
114: from the disk by a hot electron corona. Often a relation between the X-ray 
115: luminosity of the source and spectral states also exists. When the soft 
116: component dominates the spectrum, the 2--10 keV luminosity is relatively high
117:  ($\gtrsim 10^{37}$ ergs/s), and therefore this state is called the 
118: ``high/soft state'' (HS). When the hard component dominates, then usually the 
119: 2--10 keV luminosity is low ($\lesssim 10^{37}$ ergs/s), and this state is 
120: called the ``low/hard state'' (LS). This dependence shows that the mass 
121: accretion rate plays an important role in determining the spectral states. 
122: These states also differ in terms of their short-term timing properties. The 
123: HS is often characterized by lack of or a very low level of variability, 
124: whereas the LS shows very strong variability (\wsim 30\% rms in 0.04--4 Hz).
125:  Recently, changes in the radio properties are also associated with the 
126: spectral states. In the LS, compact, optically thick jets are observed
127:  \citep{Fender01}. During state transitions optically thin outflows are 
128: detected \citep{Corbel01}, and during the HS the radio emission is quenched 
129: \citep{Fender01}. There are also cases for which both spectral components are
130:  present at comparable strength, characterized by parameters that are 
131: intermediate between those of the HS and the LS. For such a case, if the 
132: source flux is lower than the HS flux, the state is often called the 
133: ``intermediate state'' (IS), and if the flux is higher than the HS flux, it is
134:  called the ``very high state'' (VHS). We note that the classification of 
135: these states is not rigorously defined and is still an active topic of debate. 
136: 
137: Most of the GBHs are observed during outbursts caused by a sudden, dramatic 
138: increase of mass accretion rate onto the black hole. Fig.~\ref{fig:asm_many} shows 
139: several examples of outburst light curves discussed in this report (see 
140: \citealt{Kalemci_tez} for all of the ASM light curves.). Because of the
141:  dependence of spectral states on the mass accretion rate, a GBH transient often 
142: follows a specific sequence of spectral states. It is usually observed in the LS at 
143: the beginning of the outburst. As flux increases, it makes a transition to the HS or 
144: the VHS. As the source decays towards quiescence, a transition to the LS is usually 
145: observed. Some transients might follow a more complicated sequence of states, and 
146: some stay in the LS throughout the outburst. It is generally believed that the state 
147: transitions involve large restructuring of the accretion geometry of 
148: GBHs \citep{Esin97,Zdziarski02_2}. Therefore, analysis of these sources during
149: state transitions may probe the dynamics of their accretion structure. Although the
150:  mass accretion rate is a very important parameter determining the spectral states,
151:  it is unlikely that the states and transitions are solely determined by this 
152: parameter. Some sources show hysteresis of the transition luminosities \citep[the 
153: luminosity of transition from LS to HS during the beginning of the outburst is much
154:  higher than the luminosity of the transition from HS to LS during the outburst 
155: decay,][]{Miyamoto95,Nowak02b}, and for some sources, a second, independent 
156: parameter seems to be required to explain the complexity of transitions. It is not
157:  clear what this second parameter is. It can be the position of the inner edge of 
158: the accretion disk, but this parameter may not be completely independent from the 
159: mass accretion rate (e.g. \citealt{Esin97,Meyer00}). Based on the behavior of the 
160: 1998 outburst of \xf, \cite{Homan01} claimed that the second independent parameter 
161: may be the size of the corona. The transitions may also be a result of an overall 
162: change in the type and the geometry of the corona. \cite{Zdziarski02_2} explains 
163: the different states in Cyg~X-1 by the change of an accretion structure which 
164: consists of a hot inner accretion flow surrounded by an optically thick disk 
165: truncated far away from the minimum stable orbit in the LS, to one that 
166: consists of flares and active regions above an accretion disk extending close 
167: to the minimum stable orbit in the so-called ``soft state''. We note that 
168: for Cyg X-1, the soft state is more similar to the IS of GBH transients rather than
169:  the HS. According to the hybrid model of \cite{Coppi00}, the state transitions 
170: may be explained by a change of size and electron energy distribution of 
171: the corona. In the LS, the inner part of the disk puffs up and becomes the 
172: hot corona, dominated by thermal electrons. Non-thermal electrons may also be 
173: present. In the HS, the edge of the cool accretion disk is close to the minimum 
174: stable orbit, and a small, non-thermal corona is also present. In this case, Compton 
175: scattering of disk photons by the non-thermal electrons is responsible for the steep 
176: power-law observed in the spectra of GBHs in the HS. It has also been claimed that 
177: this steep power-law component is a manifestation of Compton up-scattering of disk 
178: photons by a converging inflow of material inside the last stable orbit 
179: \citep[][and references therein]{Laurent01}. Then the transition might be related to 
180: the type of Comptonization: bulk motion in the HS versus thermal (or hybrid) in the 
181: LS. 
182: 
183: \begin{figure}[t]
184: %\centerline{\includegraphics{f1.ps}}
185: \plotone{f1.ps}
186: \caption{\label{fig:asm_many}
187: The \rxte, All Sky Monitor light curves of six outbursts in 1.5 - 12 keV band. The 
188: dashed lines represent the time of transition to the LS during outburst decay.
189: }
190: \end{figure}
191: 
192: The transition to the LS not only results in sharp changes in the X-ray 
193: spectrum, but also creates a physical environment that shows strong broad-band 
194: variability and quasi-periodic oscillations (QPO). The changes in the spectral 
195: parameters close to the state transition might provide important clues to 
196: understand how this transition is occurring and the driving force of the 
197: observed variability. Our group has been observing GBH transients during 
198: outburst decays in X-rays with the \emph{Rossi X-ray Timing Explorer} (\rxte) 
199: and in radio. We have quantified the evolution before, during and after the 
200: transition to the LS for various sources 
201: \citep{Kalemci01,Tomsick01b,Kalemci02,Tomsick03_2}. In 
202: Section~\ref{subsec:evolto}, we combine results of spectral and temporal 
203: analyses of the \rxte\ data from individual sources during outburst decay 
204: using our observations as well as the archival data, and investigate the 
205: evolution of the spectral index (\ind), the inner disk temperature (\tin), the 
206: power-law flux, the disk-blackbody (diskbb) flux and the power-law fraction (PLR, 
207: ratio of the power-law flux to the total flux in 3--25 keV band) to obtain a global 
208: understanding of the physical environment before and during the state transition.
209: 
210: Another important topic is the evolution of spectral parameters after the
211: state transition, which provides an understanding of the dynamics of the accretion 
212: systems in the LS. In addition to the spectral evolution, the evolution of the 
213: temporal parameters is another valuable tool in the LS. Thus, in 
214: Section~\ref{subsec:evolafter}, we analyze the evolution of spectral and temporal 
215: parameters of GBH transients after the state transition. The temporal parameters we 
216: use are the characteristic frequencies of the Lorentzian components in the 
217: power-spectral-density (PSD) fits and the rms amplitude of variability.
218:  
219: %% /*******************************************************************
220: %% ** Observations and Analysis                                      **
221: %% *******************************************************************/
222: 
223: \section{Observations and Analysis}\label{sec:obs}
224: 
225: We analyzed the PCA/\emph{RXTE} (Proportional Counter Array, see 
226: \citealt{Bradt93} for a description of \emph{RXTE}) data from all GBH 
227: transients that have been observed with \emph{RXTE} between 1996 and 2001 that 
228: made a state transition during outburst decay. Eight sources in eleven 
229: outburst decays obey these source selection criteria. These sources and 
230: outburst years are : \xs\ in 2001, GRO~J1655$-$40 in 1996, \reso\ in 1998, 
231: XTE~J1755$-$324 in 1997, \gx\ in 1998 (decayed in 1999), \fu\ in 1998, 1999 
232: and 2001, \xf\ in 1998 and 2000 and \maso\ in 1999. Although XTE~J1755$-$324 obeys 
233: the criteria, it has very poor coverage and results from this source are 
234: not included in this work. This decreases the number of sources to seven, and the 
235: number of outbursts to ten. Note that \maso\ did not make a ``traditional'' 
236: transition to the low/hard state, however it showed timing noise for some 
237: observations during the decay (see \citealt{Kalemci_tez} and references 
238: therein for the properties and observation times of each 
239: source).
240: 
241: For all of the spectral analysis, we fit the data in the 3--25 keV band using 
242: the response matrix and the background model created using the standard FTOOLS 
243: (version 5.2) programs. We added a 1\% systematic error to the spectra to 
244: account for uncertainties in the PCA response. We used a 
245: multi-component spectral model consisting of a power-law, a multi-color 
246: disk blackbody \citep[\textbf{diskbb} in XSPEC, ][]{Makishima86}, a broad 
247: absorption edge \citep[\textbf{smedge} in XSPEC, ][]{Ebisawa94} with 
248: interstellar absorption (\textbf{phabs} in XSPEC). This model has been commonly
249: used for the spectral analysis of GBHs in the LS 
250: \citep{Tomsick00,Sobczak00}. For all observations, the reduced $\chi^{2}$ is 
251: between 0.5 and 1.5. For some observations, in order to reach acceptable 
252: $\chi^{2}$ values, a Gaussian iron line feature was needed. However, the 
253: parameters of the iron line have not been used in the analysis. For some of the
254: outbursts, we used published spectral fit parameters if the same model was applied 
255: for the fit (\citealt{Tomsick00} for \fu\ in 1998, and
256: \citealt{Tomsick01b} for \xf\ in 2000). All PCA fluxes in this report are unabsorbed 
257: model fluxes to remove source to source variations due to different absorption 
258: column densities. The details of spectral fit models and parameters for each source 
259: can be found in \cite{Kalemci_tez}. 
260: 
261: For the temporal analysis, we were as uniform as possible in terms of 
262: choosing energy bands, time resolution and segmentation of light curves for 
263: all of our sources. However, since some of this work is based on the 
264: analysis of the archival data, we were limited to the choice of data modes by 
265: the PI of the original proposal. For most of the observations we used 2--26 keV
266:  energy band, a Nyquist frequency of 256 Hz and 256 s light curve segments.
267:  Although there may be slight differences from source to source in terms of 
268: energy band, and highest and lowest PSD frequencies, these differences are not 
269: critical since we look for trends rather than absolute values.
270: 
271: \begin{figure}
272: \plotone{f2.ps}
273: \caption{\label{fig:lorex}
274: Power spectra of \xs\ in the HS (top panel) and in the LS (bottom panel). The PSD
275: in the LS is fitted with two broad and a narrow Lorentzian. The HS observation has
276: a 2$\sigma$ upper limit of 4\% on the rms amplitude of variability, whereas the LS 
277: observation 2 days later has an rms amplitude of variability of $12.43 \pm 0.55$ \%}
278: \end{figure}
279: 
280:    Historically, the PSD of GBHs during the low state has been modeled by a 
281: broken power-law (or power-laws with more than one break) plus narrow 
282: Lorentzians to fit the QPOs \citep{Nowak99,Tomsick00,Kalemci01}. However, 
283: recent papers successfully fit several GBH and neutron star PSDs with broad 
284: Lorentzians for the continuum and narrow Lorentzians for the QPOs 
285: \citep{Belloni02,vanStraaten01,Pottschmidt02,Kalemci02}. Following this
286: approach, we fit all our PSDs with Lorentzians of the form:
287: \begin{equation}
288: L_{i}(f)\;=\;{{R_{i}^{2} \; \Delta_{i}}\over{2 \, \pi \; [(f-f_{i})^{2}+({1\over{2}}\,\Delta_{i})^2]}}
289: \end{equation}
290: where subscript $i$ denotes each Lorentzian component in the fit, $R_{i}$ is 
291: the rms amplitude of the Lorentzian in the frequency band of -$\infty$ to 
292: +$\infty$, $\Delta_{i}$ is the 
293: full-width-half-maximum, and $f_{i}$ is the resonance frequency. A useful
294: quantity of the Lorentzian is the ``peak frequency'' at which the Lorentzian
295:  contributes maximum power per logarithmic frequency interval:
296: \begin{equation}
297: \nu_{i}\;=\;{f_{i}\;\left({{\Delta_{i}^{2}}\over{4 \, f_{i}^{2}}}+1\right)^{1/2}}
298: \end{equation}
299: Fig.~\ref{fig:lorex} shows an example LS power spectrum of \xs\ in the form of {PSD 
300: $\times$ frequency}, along with broad and narrow Lorentzians fit components. In this 
301: figure, the Lorentzians peak at $\nu_{i}$, demonstrating the easy identification of 
302: characteristic frequencies as peak frequencies of Lorentzian components. Some of our 
303: observations contain a Lorentzian that is narrow (with quality value 
304: $Q_{i} = f_{i}/{\Delta_{i}}>2$, as compared to $Q<1$ for broad Lorentzians) which we 
305: call a QPO. In this work, for each observation, the term ``characteristic 
306: frequency'' represents the resonance frequency of the fundamental QPO\footnote{The 
307: peak frequency and the resonance frequency differ by $<$3\% for QPOs with $Q>$2.} if 
308: present, and otherwise the lowest peak frequency of the broad Lorentzian components 
309: in the PSD\footnote{The lowest peak frequency is nearly equivalent to the 
310: ``break frequency'' if broken power-law modeling is adopted, see \cite{Belloni02} 
311: for a detailed discussion.}. Note that since the QPO frequencies and the Lorentzian 
312: peak frequencies are all shown to be correlated \citep{Belloni02}, it does not 
313: matter which one we use to characterize the evolution. The rms amplitudes are 
314: calculated over a frequency band from zero to infinity.
315: 
316:  Mostly due to our group's monitoring program of GBHs with \rxte\ during 
317: outburst decay, we obtained very good coverage for 3 sources, in 6 different 
318: outbursts (1998, 1999 and 2001 outbursts of \fu, 1998 and 2000 outbursts of
319:  \xf, and the 2001 outburst of \xs\ with almost daily monitoring). 
320:  GRO~J1655$-$40 and \gx\ are excluded from the discussion of the changes before
321: and during the transition in Section~\ref{subsec:evolto} since they have poor 
322: coverage ($\geq$10 days between observations). XTE~J1748$-$288 with 
323: observations every \wsim5 days is included in the discussion of the evolution 
324: to the state transition, however is not included in the discussion of the 
325: changes during the transition. Results from these sources are included in 
326: the discussion of the evolution after the transition for which very good coverage 
327: is not necessary. All the data used in this report are available on-line as a 
328: machine-readable table.
329: 
330: 
331: %% /*******************************************************************
332: %% ** Results                                                        **
333: %% *******************************************************************/
334: 
335: \section{Results}\label{sec:results}
336: 
337: In this work, we define the state transition in terms of sharp changes in 
338: variability properties rather than sharp changes in spectral properties. All 
339: of our sources with good coverage showed sharp, distinct changes in terms of 
340: variability in less than two days, but this was not the case for the spectral 
341: properties. Most of the transitions occurred from the HS to LS, and for those 
342: cases the transition was marked by a very large increase in the total rms 
343: amplitude of variability as shown in Fig.~\ref{fig:evolmix1}a. For most of the 
344: cases, the transition was from a featureless, Poisson noise dominated PSD with 
345: only a few \% rms amplitude upper limit to a PSD showing well defined broad-band 
346: variability. In other words, the variability ``appeared'' on a few days timescale 
347: for these systems. Fig.~\ref{fig:lorex}  represents this appearance of variability, 
348: as the observation in the HS (top panel, with $<$4\% rms upper limit variability) 
349: and the observation in the LS (bottom panel, with $>$12\% rms variability with well
350: defined Lorentzian components and a QPO) in this figure are only two days apart.
351:  Since the HS PSD is often featureless and has very low or no rms amplitude of 
352: variability, it is not possible to discuss the evolution of characteristic 
353: frequencies before the transition for none of our sources except the 2000 outburst 
354: of \xf, which showed a complex pattern. It was in the IS, with rms amplitude 
355: variability of \wsim13\%. During the IS, it showed a QPO with constant frequency at 
356: \wsim 9 Hz. Then it showed a large drop in variability to an amplitude of 
357: \wsim7\%. We marked this change as the transition. Two days later, the rms amplitude 
358: of variability jumped again to levels of 15\%. The morphology of the PSDs was 
359: different in the IS, LS, and during the time that the rms amplitude showed a drop 
360: \citep{Kalemci_tez}. Despite the complexity of some individual cases, it is possible 
361: to infer global patterns of evolution for different parameters before, during and 
362: after the state transition.
363: 
364: \begin{figure}
365: \plotone{f3.ps}
366: \caption{\label{fig:evolmix1}
367: Evolution of (a) the rms amplitude of variability, (b) the spectral index, and
368:  (c) the inner disk temperature. The state transition is assumed to have happened 
369: exactly in between the observations closest to the sharp change observed in 
370: panel (a), and represented by a dashed line. For (a), the values for each 
371: source are normalized with respect to the value just after the state 
372: transition. For both (b) and (c), the values for each source are normalized 
373: with respect to the value just before the state transition. For some points, the
374: 1 $\sigma$ errors are smaller than the plot symbols. The upper limits are 2$\sigma$.}
375: \end{figure}
376: 
377: \subsection{Changes in spectral properties that lead to state transitions}\label{subsec:evolto}
378: 
379: First, we characterize the changes in spectral properties before the state
380: transition. The evolution of the photon spectral index (\ind), and the inner disk 
381: temperature (\tin) close to the transition are shown in Fig.~\ref{fig:evolmix1}b 
382: and c respectively. The evolution of the PLR, power-law flux and the diskbb flux 
383: are shown in Fig.~\ref{fig:evolmix2}. In most sources, \ind\ 
384: (Fig.~\ref{fig:evolmix1}b) is approximately constant until 3-5 days before the 
385: timing transition at which point it begins to decline. For some sources, there is 
386: a distinct drop in \ind\ at the beginning of the decline. Except for the 1998 
387: outburst of \xf, \tin\ (Fig.~\ref{fig:evolmix1}c) either decreases or stays
388:  constant before the transition. The power-law flux (Fig.~\ref{fig:evolmix2}b) 
389: shows a complex behavior (see also Fig.~\ref{fig:evolplf}), and the 
390: diskbb-flux (Fig.~\ref{fig:evolmix2}c) decreases for majority of the outbursts, or 
391: stays constant. The cumulative effect of the power-law and diskbb-flux evolution 
392: for the majority of the outbursts is an increasing PLR (Fig.~\ref{fig:evolmix2}a) 
393: before the transition. 
394: 
395: \begin{figure}
396: \plotone{f4.ps}
397: \caption{\label{fig:evolmix2}
398: Evolution of (a) the PLR, (b) the power-law flux, and (c) the diskbb flux. The
399:  dashed line represents the time of transition. The values for each source are
400:  normalized with respect to the value just before the state transition. A legend 
401: is given in Fig.~\ref{fig:evolmix1}.  For some points, the 1 $\sigma$ errors are 
402: smaller than the plot symbols. The upper limits are 2$\sigma$}
403: \end{figure}
404: 
405: We established that the transition occurs in a few days time scale for GBH 
406: transients. The change in variability properties (usually, the appearance of
407: variability) that defines the transition should be a response to the change of 
408: one or more spectral parameters. To understand which parameter drives the 
409: transitions and appearance of variability, we compared the two observations 
410: just before and after the transition (observations just before and  
411: after the dashed lines in Figs.~\ref{fig:evolmix1}~and~\ref{fig:evolmix2}). 
412: We specify that a spectral parameter is showing a sharp change during the 
413: transition if either the slope of the parameter as a function of time changes
414:  sign or the percentage change in the value of the parameter compared to the 
415: previous observation is at least three times that of the previous observation.
416:  With this definition, two parameters show a sharp change for the majority of 
417: the outbursts: the power-law flux (see Fig.~\ref{fig:evolplf}) and the 
418: PLR\footnote{\gx\ during its recent decay from the outburst showed a large 
419: increase in the PLR, and by using this information, we were able to estimate 
420: when the transition would occur and scheduled a successful \rxte\ observation 
421: to catch the beginning of the transition.}. The PLR is not an independent 
422: parameter, and in this case its increase is mostly driven by the increase in the 
423: power-law flux. For both the 1999 and the 2001 outbursts of \fu, there is no 
424: sharp change in any of the spectral parameters between those two observations,
425:  yet one observation shows no variability and the next one shows variability. 
426: However, for both, the evolution of the power-law flux shows a change in the slope 
427: an observation earlier. The situation is more complicated for the 2000 outburst 
428: of \xf. There is a large increase in the power-law flux and PLR during the 
429: transition, but this change corresponds to a decrease in the rms amplitude 
430: \citep{Kalemci_tez}. 
431: %Fig3. was here
432: 
433: \begin{figure}[t]
434: \plotone{f5.ps}
435: \caption{\label{fig:evolplf}
436: Evolution of the power-law flux before and after the transition. The dashed 
437: line indicates the time of state transition. The values for each source are 
438: normalized with respect to the value just before the state transition. The 1 
439: $\sigma$ errors are smaller than the plot symbols
440: }
441: \end{figure}
442: 
443: Analysis of PCA data from GBH transient \maso\ during its outburst 
444: decay revealed interesting behavior in terms of variability and spectral 
445: evolution and provided another case for the relation between the power-law 
446: flux and variability. \maso\ did not show a traditional transition, but showed 
447: variability for some of the observations during the decay. In 1 day 
448: timescale, the power-law flux almost doubled and variability appeared, and 
449: when the power-law flux dropped below a threshold value, variability disappeared. 
450: Except the PLR, which is tied to the power-law flux, no other spectral 
451: parameter showed a sharp change between observations that show variability, 
452: and those that do not show variability for this source 
453: \citep{Kalemci_tez}.
454: 
455: \subsubsection{A discussion of transition fluxes}
456: 
457: It is important to get an idea of when the transition happens during
458: the decay for effective monitoring of the sources with pointed instruments. 
459: Often the sources are not monitored frequently enough (i.e., once a day) with 
460: pointed observations, but the ASM measures the flux in 3 energy bands very 
461: frequently, which provides an opportunity to detect transitions without 
462: re-pointing the satellite. (See Fig.~\ref{fig:asm_many} for the ASM light curves
463: of some of our sources.) The transition is often detected by a sudden 
464: increase in the hardness ratio 2 (HR2, the ratio of 5--12 keV flux to 1.5--3 
465: keV flux). A few days just before and after the transition may manifest the 
466: most interesting behavior. A relationship between the peak flux and the 
467: transition flux (1.5 -- 12 keV flux for the observation that shows 
468: variability) had been realized earlier (Tomsick, private communication). We 
469: plot this relation for all of the outbursts that show a state transition in
470:  Fig.~\ref{fig:trflux}. Except the 1998 outburst of \xf, which showed one of 
471: the strongest flares in the \rxte\ history, the transition flux increases with 
472: the peak flux. The Spearman's rank order correlation coefficient for this data
473:  set (excluding the 1998 outburst of \xf) is 0.90, pointing out a correlation. 
474: On the other hand, the linear correlation coefficient is 0.52, indicating 
475: a relation which is not strictly linear.
476: 
477: \begin{figure}[t]
478: \plotone{f6.ps}
479: \caption{\label{fig:trflux}
480: Peak flux versus transition flux in 1.5 -- 12 keV band. Fluxes are in units of
481: $\rm 10^{-9}\, ergs \, cm^{-2} \, s^{-1}$. For the points connected with dashed
482:  lines, both peak fluxes and plateau fluxes are shown.
483: }
484: \end{figure}
485: 
486: Some outbursts have the usual ``fast-rise-exponential-decay'' (FRED) shape 
487: (such as \reso), and some have a very complicated shape (such as \fu\ in 2001).
488:  The complicated ones may show a plateau before the final decay (like GRO~J1655$-$40,
489: and \xf\ in 2000 in Fig.~\ref{fig:asm_many}). For these sources, we also plotted the 
490: plateau ASM flux along with the peak flux. Replacing the peak flux by the plateau 
491: flux for these sources improves the linear correlation for the whole sample to $r$ of
492:  0.706. The linear correlation is not very strong, but it allows for an estimate of 
493: when the transition might happen, which is useful for planning observations to study 
494: the transition.
495: %Fig4 was here
496: 
497: \subsection{Evolution of spectral and temporal parameters after the transition}\label{subsec:evolafter}
498: 
499:  The evolution of spectral and temporal parameters after the transition is 
500: shown in Figs.~\ref{fig:evolmix1},~\ref{fig:evolmix2},~\ref{fig:evolplf},~and~
501: \ref{fig:evolcf}. For \xs, \gx\ (not shown in 
502: Figs.~\ref{fig:evolmix1},~\ref{fig:evolmix2},~\ref{fig:evolplf} because of 
503: poor coverage), \fu\ in 1998 and \xf\ in 2000, \ind\ decreases after the 
504: transition. For other outbursts, it is either constant or shows irregular behavior.
505:  The power-law flux decreases for all of the outbursts except \xs\ 
506: (see Figs.~\ref{fig:evolmix2}~and~\ref{fig:evolplf}). For most cases, 
507: the diskbb-flux and \tin\ either decrease, or are unobservable after the 
508: transition. It is hard to constrain the evolution of \tin\ for the 1998 
509: outburst of \xf\ and the 2001 outburst of \fu.  For all outbursts, the diskbb 
510: component is unobservable within 15 days of the state transition. For \xs\ and the 
511: 2000 outburst of \xf, the PLR first increases and then stays constant. For all other 
512: outbursts, the PLR is very close to unity and does not vary. 
513: 
514: \begin{figure}[t]
515: \plotone{f7.ps}
516: \caption{\label{fig:evolcf}
517: Evolution of the characteristic frequencies (the QPO resonance frequency for
518: \xf\ in 2000, \fu\ in 1998 and 1999, and \reso, lowest peak frequency of the 
519: broad Lorentzian components in the PSD for the remaining) after the state 
520: transition. For some points, the 1 $\sigma$ errors are smaller than the plot 
521: symbols.
522: }
523: \end{figure}
524: 
525:  The evolution of the characteristic frequencies after the transition is shown
526: in  Fig.~\ref{fig:evolcf}. Except a single observation at the beginning of the
527: 2000 outburst of \xf, all characteristic frequencies decrease after the 
528: transition. There are only two observations that show variability in the 1999 
529: and 2001 outbursts of \fu, and even for those cases, the later observations 
530: have smaller characteristic frequencies. The behavior of the rms 
531: amplitudes are more complicated (see Fig.~\ref{fig:evolmix1}a). The 1998 
532: outburst of \xf\ and \reso\ show a decreasing trend. The rms amplitudes for 
533: \xs, the 2000 outburst of \xf\ and the 1998 outburst of \fu\ increase and
534:  level off. For GRO~J1655$-$40, the rms amplitudes are consistent with being 
535: constant. For \gx\ on the other hand, the rms amplitude increases 
536: \cite{Kalemci_tez}. 
537: %Fig5 was here
538: 
539: %% /*******************************************************************
540: %% ** Discussion                                                        **
541: %% *******************************************************************/
542: 
543: \section{Discussion}
544: 
545: \subsection{State transition and appearance of variability}
546: 
547:  First, we discuss possible reasons for the state transition and appearance of
548:  variability based on our results above. Suppose almost all of the 
549: variability we observe is due to the photons coming from the power-law
550: component presumably originating in the corona, and the disk photons are 
551: providing mostly Poisson noise. This idea is supported by the fact that 
552: very little or no variability is observed during the HS when the diskbb 
553: dominates, and strong variability is observed when the power-law component 
554: dominates in the LS. The anti-correlation between the diskbb flux and the rms 
555: amplitude observed both in \xs\ \citep{Kalemci02} and the 2000 outburst of 
556: \xf\ \citep{Kalemci01} is also potentially consistent with this scenario. It
557:  is conceivable that the Poisson noise dominated diskbb flux reduces the rms 
558: amplitudes of variability to unobservable levels. Then the PLR would be a good 
559: indicator of when the variability will be observed, since the higher the PLR, the
560: lower the diskbb flux compared to the total flux. For all of our sources, no 
561: variability is observed if the PLR$<$0.45. However, the value of the PLR cannot be
562:  the sole determinant of when the variability will be observed. For most of 
563: the sources the PLR has to be greater than 0.8 for the variability to appear, 
564: although it can be as low as 0.45 for \maso\ \citep{Kalemci_tez}. To test the 
565: argument that the diskbb flux might reduce the rms amplitudes to unobservable 
566: levels, we conducted very simple simulations. We added various levels of 
567: Poisson noise to the light curve of an observation of GRO~J1655$-$40 taken on 
568: August 14, 1997, and compared the resultant PSD with the original PSD. We chose 
569: this observation since it has a complex structure, and no diskbb flux, so all the 
570: emission is coming from the corona. The results are shown in 
571: Fig.~\ref{fig:simmulplot}. Although this figure clearly shows that the rms 
572: amplitude of the PSD depends on the PLR, it also shows that even for a PLR of 0.3,
573:  the variability is clearly observed. There is no large change in the PSD shape. 
574: Moreover, lack of diskbb emission \textbf{does not} guarantee broad-band 
575: variability even if a hard component exists. There is strong emission in the 
576: 6--15 keV band for most of the sources before the transition, and spectral analysis
577:  shows that only a few percent of the emission in this band is from the soft 
578: component. Nonetheless, although completely power-law dominated, these 
579: observations do not show variability. Note that, these simple simulations do not 
580: take into account the effects of strong disk emission on the corona. The two 
581: emission components are not independent of each other, because Compton cooling of 
582: the corona by the diskbb emission can affect the temperature and density structure 
583: of the corona. The amount of diskbb radiation cannot be ruled out as an important 
584: parameter in determining when the variability will be observed.
585: 
586: \begin{figure}
587: \plotone{f8.ps}
588: \caption{\label{fig:simmulplot}
589: PSD of GRO~J1655$-$40 during an observation in the LS is shown with the solid 
590: curve obtained by fitting the PSD with two broad Lorentzians and a QPO. The 
591: dotted and the dashed curves are fits to the PSDs when Poisson noise is added 
592: to the original light curve. The PSD under the dotted curve corresponds to a PLR 
593: of 0.6 and the PSD under the dashed curve corresponds to a PLR of 0.3. 
594: }
595: \end{figure}
596: 
597: The power-law emission dominates the LS when the variability is observed but 
598: its presence alone is not enough to create broad-band variability. These 
599: findings suggest several possibilities for the presence of variability in 
600: the LS. The origin of the power-law component in the LS and the other
601: states might be completely different. The passage from the IS to the LS (as 
602: observed for \gx\ and the 2000 outburst of \xf) might be due to a change of the 
603: form of the corona from active regions above the accretion disk to an inner 
604: accretion flow \citep{Zdziarski02_2}. The power-law component in the HS 
605: may be a result of Comptonization of disk photons by non-thermal 
606: electrons \citep{Coppi00}, or bulk motion Comptonization \citep{Laurent01},
607: whereas in the LS, it may be due to thermal Comptonization. However, if the form 
608: or the composition of the corona is changing during the transition, one 
609: would expect a sharp change not only in power-law flux, but also in \ind, especially 
610: if the change is from bulk motion Comptonization to thermal Comptonization. There 
611: are two ways to increase the power-law flux: increasing the number of seed photons; 
612: and/or increasing the area of the corona that intercepts the seed photons. It is 
613: unlikely that the number of seed photons are increasing, the corona is believed to be
614:  optically thin, and an increase in the soft flux should be observed (unless most 
615: photons have energies below the PCA range). Therefore, our findings are consistent 
616: with the idea that, during the transitions, the size of the corona increases to a 
617: threshold value for variability, which would be consistent with the idea that the 
618: second independent parameter determining the spectral states is the size of the 
619: Comptonizing region \citep{Homan01}. The strongest support for this explanation 
620: comes from the behavior of XTE~J1859$+$226. Variability appeared for this 
621: source whenever the power-law flux is higher than a threshold value. No other 
622: spectral variable shows a significant change between observations with and 
623: without variability \citep{Kalemci_tez}. The only caveat of this scenario is 
624: the resemblance of the variability properties of  XTE~J1859$+$226 to the IS 
625: observations of \gx\ and the 2000 outburst of \xf, rather than to the LS 
626: properties.
627: 
628: An increase in the power-law flux does not always correspond to the appearance
629: of variability. The 2000 outburst of \xf\ behaved differently; the rms 
630: amplitude of variability decreased as the the power-law flux increased
631: sharply. But in a couple of days after the first transition, another 
632: transition happened and strong variability that is usually associated with the
633:  LS appeared. The 1999 and the 2001 outbursts of \fu\ did not show a sharp 
634: change in any of the spectral properties at the transition 
635: (see Figs.~\ref{fig:evolmix1}~and~\ref{fig:evolmix2}), but both showed 
636: a change in the slope of the evolution of the power-law flux an observation 
637: earlier (see Fig.~\ref{fig:evolplf}). This suggests the possibility that for some 
638: cases the appearance of variability (in the case of \xf, variability associated 
639: with the LS) is delayed. The transition for these cases  may be happening slower 
640: than that of other cases, and during the restructuring of the accretion 
641: geometry, the coherence could be lost for all timescales and the timing 
642: signatures could be suppressed for a few days. 
643: 
644: \subsection{Evolution of spectral and temporal parameters}
645: 
646: Two important observations about the transitions to the LS can be made by 
647: analyzing the evolution of spectral parameters before the transition. First,
648:  the PLR increases for the majority of the outbursts. The corona (again 
649: assumed to be the source of the power-law flux) must dominate for the transition to
650:  happen, the PLR is greater than 0.65 for all sources after the transition to 
651: the LS. The PLR was as low as 0.45 for \maso\ when the variability was 
652: observed, but it never made the transition to the canonical LS, and returned 
653: back to the HS when the PLR dropped \citep{Kalemci_tez}. Second, the 
654: spectral indices for some of the sources show a step function-like decrease 
655: 3-5 days  before the transition (see Fig.~\ref{fig:evolmix1}b). This may be 
656: interpreted as the time of spectral transition, which would indicate that the 
657: transition in timing properties lags the transition in spectral properties. Then, 
658: the change in  \ind\ may be a sign of changing the form of the corona as 
659: discussed in the previous section, and the lag between the change in the \ind\ 
660: and the appearance of variability is the time for the corona to reach a 
661: threshold volume. However, some sources do not show any drop in \ind, 
662: but still show a sharp change in temporal variability  (Fig.~\ref{fig:evolmix1}a), 
663: making the transitions much easier to identify. 
664: 
665:   
666: After the transition, the evolution of the inner disk temperature, the diskbb
667:  flux, the power-law index and especially the characteristic frequencies of 
668: the variability are consistent with the idea of the inner disk retreat. In the 
669: basic formulation of \cite{Makishima86}, the optically thick disk has a 
670: temperature profile that decreases with increasing radius. Therefore, if the 
671: inner disk evaporates as part of the transition, it is expected that both 
672: \tin\ and diskbb flux decrease. Most sources show a decrease in diskbb flux, 
673: and at least four sources show a decrease in \tin. The drop of spectral index 
674: may also be a sign of increasing inner disk radius. As the disk is closer to 
675: the black hole, its temperature and flux are higher, causing effective cooling of 
676: the corona, increasing the spectral index \citep{Zdziarski02_2}. Four sources show
677: a decreasing \ind. It is intuitively expected that the characteristic 
678: frequencies decrease as the inner disk radius increases. The dynamical 
679: timescale (the fastest timescale) in the accretion disk is shorter close to 
680: the black hole, and it is expected that higher frequency variability is 
681: created in this region. As the inner disk radius retreats (or the inner disk 
682: evaporates) the dynamical timescale (at the inner edge) increases, and 
683: therefore the characteristic frequencies in the PSD decrease. The exact 
684: relation between the characteristic frequencies and the dynamical timescales 
685: is not clear yet. All sources show a decreasing behavior in terms of 
686: characteristic frequencies, except for one observation. The overall evolution 
687: of both spectral and temporal properties therefore indicates that the inner 
688: edge of the accretion disk retreats as the system progresses in the LS, in 
689: accordance with the model of \cite{Esin97}.  In the exceptional case, the QPO 
690: frequencies in the 2000 outburst of \xf\ first increased and then decreased 
691: after the transition (see Fig.~\ref{fig:evolcf}). This behavior is consistent 
692: with the prediction of the ``accretion ejection instability'' QPO model 
693: \citep{Tagger99} if the inner accretion disk is close to the marginally stable 
694: orbit during this observation. This possibility is discussed in detail in 
695: \cite{Rodriguez02b_inprep} and \cite{Kalemci_tez}. 
696: 
697: The interpretation of the behavior of the rms amplitudes is more complex. We 
698: showed the relation between the PLR and the rms amplitudes earlier (see 
699: Fig.~\ref{fig:simmulplot}). The ``Poissonic'' nature of the disk component 
700: causes a decrease in the rms amplitudes at lower energies due to the diskbb 
701: energy spectrum peaking at those energies. This might explain the increase and
702:  leveling of the rms amplitudes in \xs\ and the 2000 outburst of \xf. Although
703:  the rms amplitudes seem to be decreasing after the transition for the 1998 
704: outburst of \xf, it might be a result of not constraining the high frequency 
705: part of the PSD. The first observation requires two Lorentzians to fit, and 
706: results in a high PSD amplitude, whereas for the remaining observations, the 
707: second Lorentzian, although statistically not required, cannot be excluded 
708: \citep{Kalemci_tez}. The fits are better for \reso\ and it also 
709: shows a decreasing behavior. The interpretation of the rms amplitudes of
710:  \reso\ is complicated due to the high absorption column density. The Galactic
711:  ridge emission may also be affecting the rms amplitudes at low flux levels by
712:  supplying additional Poissonic emission. The rms amplitude of \gx\ increases
713:  after the transition although there is no diskbb flux in the 3-25 keV band. 
714: Overall, there is no definite trend of evolution for the rms amplitudes after
715: the transition.
716: 
717: \section{Summary}
718: 
719: An important goal of black hole binary research is to understand the 
720: accretion structure and nature of variability of these systems. This work 
721: addresses these goals by analyzing the X-ray temporal and spectral properties 
722: of a relatively large subset of GBHs during outburst decay. We characterize
723: the evolution of spectral and temporal properties before and after the 
724: transitions, and also work on the changes right at the transition to
725: understand the appearance of broad-band variability.
726: 
727: For this study, a total of seven \rxte\ sources in ten outbursts are analyzed. 
728: Thanks to our group's monitoring program, the coverage close to the transition 
729: is significantly improved, and for the first time allows us to determine the 
730: physical changes that drive the transitions in detail. The first problem we 
731: work on is the evolution of spectral parameters before the state transition and
732:  appearance of broad-band variability at the state transition. We show that 
733: the changes in variability properties (a sudden increase or decrease in the 
734: rms amplitude of variability) are sharper than changes in spectral properties, 
735: and it is easier to identify a transition with the temporal properties. A 
736: change in the spectral index \ind\ is shown to be the pre-cursor of the 
737: transition showing a decrease 3-5 days before for most of our sources. We also 
738: show that the PLR increases close to the transition for all sources. The hard
739:  power-law component must dominate the spectrum for the transition to happen. 
740: Very frequent monitoring observations (\wsim once a day) allow us to 
741: determine the power-law flux (and consequently the PLR) as the parameter 
742: showing a sharp change during the transitions which may indicate
743: a threshold volume for the corona for the appearance of variability.
744: 
745: We also investigate the evolution after the state transition. For most of the
746: cases, the \tin\ and the diskbb flux decrease after the transition, and they 
747: are unobservable with PCA within fifteen days of the transition for all 
748: outbursts. The characteristic frequencies of all except one observation of the 
749: 2000 outburst of \xf\ decrease after the state transition. These spectral and 
750: temporal changes are consistent with the idea that the inner accretion disk 
751: retreats after the state transition. The rms amplitude of variability does not 
752: show any global trend with time after the transition, and its behavior is 
753: consistent with the emission from the soft component having no or very little 
754: temporal signature.
755: 
756: %% /*******************************************************************
757: %% ** Acknowledgments                                               **
758: %% *******************************************************************/
759: 
760: \acknowledgments 
761: E.K. acknowledges useful discussions with David Smith, Ali Alpar and \"Unal 
762: Ertan. The majority of this work was done at the Center for Astrophysics and
763: Space Sciences (CASS) at UCSD as part of E.K.'s dissertation. E.K. acknowledges 
764: the support of the Astrophysics Forum at Sabanc\i\ University where a portion 
765: of this work was prepared, and also acknowledges partial support of  
766: T\"UB\.ITAK. The authors would like to thank all scientists contributed to 
767: the T\"{u}bingen Timing Tools. J.A.T. acknowledges partial support from NASA grant 
768: NAG5-13055. K.P. was supported by grant Sta 173/25-1 and Sta 173/25-3 of the 
769: Deutsche Forschungsgemeinschaft. P.K. acknowledges partial support from NASA grant 
770: NAG5-7405. This work was also supported by NASA contract NAS5-30720.
771:   
772: %% /*******************************************************************
773: %% ** Bibliography                                                   **
774: %% *******************************************************************/
775: %%\bibliographystyle{jwapjbib}
776: %%\bibliography{/home/emrahk/refs}
777: 
778: \begin{thebibliography}{}
779: 
780: \bibitem[\protect\astroncite{{Belloni}, {Psaltis} \& {van der
781:   Klis}}{2002}]{Belloni02}
782: {Belloni}, T., {Psaltis}, D., \& {van der Klis}, M.,  2002, ApJ, 572, 392
783: 
784: \bibitem[\protect\astroncite{{Bradt}, {Rothschild} \& {Swank}}{1993}]{Bradt93}
785: {Bradt}, H.~V., {Rothschild}, R.~E., \& {Swank}, J.~H.,  1993, A\&AS, 97, 355
786: 
787: \bibitem[\protect\astroncite{{Coppi}}{2000}]{Coppi00}
788: {Coppi}, P.~S.,  2000,
789: \newblock in High Energy Processes in Accreting Black Holes, ASP Conf. Ser.
790:   161: High Energy Processes in Accreting Black Hole, eds. {Poutanen}, J. and
791:   {Svensson}, R., (1999)
792: 
793: \bibitem[\protect\astroncite{Corbel et~al.}{2001}]{Corbel01}
794: Corbel, S., et~al., 2001, ApJ, 554, 43
795: 
796: \bibitem[\protect\astroncite{{Ebisawa} et~al.}{1994}]{Ebisawa94}
797: {Ebisawa}, K., et~al., 1994, PASJ, 46, 375
798: 
799: \bibitem[\protect\astroncite{{Esin}, {McClintock} \& {Narayan}}{1997}]{Esin97}
800: {Esin}, A.~A., {McClintock}, J.~E., \& {Narayan}, R.,  1997, ApJ, 489, 865
801: 
802: \bibitem[\protect\astroncite{{Fender} et~al.}{2001}]{Fender01}
803: {Fender}, R.~P., {Hjellming}, R.~M., {Tilanus}, R.~P.~J., {Pooley}, G.~G.,
804:   {Deane}, J.~R., {Ogley}, R.~N., \& {Spencer}, R.~E.,  2001, MNRAS, 322, L23
805: 
806: \bibitem[\protect\astroncite{{Homan} et~al.}{2001}]{Homan01}
807: {Homan}, J., {Wijnands}, R., {van der Klis}, M., {Belloni}, T., {van Paradijs},
808:   J., {Klein-Wolt}, M., {Fender}, R., \& {M{\'e}ndez}, M.,  2001, ApJS, 132,
809:   377
810: 
811: \bibitem[\protect\astroncite{{Kalemci}}{2002}]{Kalemci_tez}
812: {Kalemci}, E.,  2002, \emph{Ph.D. Thesis}, University of California, San Diego
813: 
814: \bibitem[\protect\astroncite{{Kalemci} et~al.}{2003}]{Kalemci02}
815: {Kalemci}, E., {Tomsick}, J.~A., {Rothschild}, R.~E., {Pottschmidt}, K.,
816:   {Corbel}, S., {Wijnands}, R., {Miller}, J.~M., \& {Kaaret}, P.,  2003, ApJ,
817:   586, 419
818: 
819: \bibitem[\protect\astroncite{{Kalemci} et~al.}{2001}]{Kalemci01}
820: {Kalemci}, E., {Tomsick}, J.~A., {Rothschild}, R.~E., {Pottschmidt}, K., \&
821:   {Kaaret}, P.,  2001, ApJ, 563, 239
822: 
823: \bibitem[\protect\astroncite{{Laurent} \& {Titarchuk}}{2001}]{Laurent01}
824: {Laurent}, P., \& {Titarchuk}, L.,  2001, ApJ, 562, L67
825: 
826: \bibitem[\protect\astroncite{{Makishima} et~al.}{1986}]{Makishima86}
827: {Makishima}, K., {Maejima}, Y., {Mitsuda}, K., {Bradt}, H.~V., {Remillard},
828:   R.~A., {Tuohy}, I.~R., {Hoshi}, R., \& {Nakagawa}, M.,  1986, ApJ, 308, 635
829: 
830: \bibitem[\protect\astroncite{{McClintock} \& {Remillard}}{2003}]{McClintock03}
831: {McClintock}, J.~E., \& {Remillard}, R.~A.,  2003,
832: \newblock in {X-ray Binaries, astro-ph/0306213}
833: 
834: \bibitem[\protect\astroncite{{Meyer}, {Liu} \&
835:   {Meyer-Hofmeister}}{2000}]{Meyer00}
836: {Meyer}, F., {Liu}, B.~F., \& {Meyer-Hofmeister}, E.,  2000, A\&A, 361, 175
837: 
838: \bibitem[\protect\astroncite{{Miyamoto} et~al.}{1995}]{Miyamoto95}
839: {Miyamoto}, S., {Kitamoto}, S., {Hayashida}, K., \& {Egoshi}, W.,  1995, ApJ,
840:  442, LL13
841: 
842: \bibitem[\protect\astroncite{{Nowak}}{2002}]{Nowak02b}
843: {Nowak}, M.~A.,  2002,
844: \newblock in Proceedings of the 4th Microquasar Workshop, 2002, eds.
845:   Durouchoux, Fuchs, Rodriguez, to be published by the Center for Space
846:   Physics: Kolkata, 7624
847: 
848: \bibitem[\protect\astroncite{{Nowak} et~al.}{1999}]{Nowak99}
849: {Nowak}, M.~A., {Vaughan}, B.~A., {Wilms}, J.~., {Dove}, J.~B., \& {Begelman},
850:   M.~C.,  1999, ApJ, 510, 874
851: 
852: \bibitem[\protect\astroncite{{Pottschmidt} et~al.}{2003}]{Pottschmidt02}
853: {Pottschmidt}, K., et~al., 2003, A\&A, 407, 1039
854: 
855: \bibitem[\protect\astroncite{{Rodriguez} et~al.}{2002}]{Rodriguez02b_inprep}
856: {Rodriguez}, J., {Corbel}, S., {Kalemci}, E., {Tomsick}, J.~A., \& {Tagger},
857:   M.,  2002, submitted to ApJ
858: 
859: \bibitem[\protect\astroncite{{Sobczak} et~al.}{2000}]{Sobczak00}
860: {Sobczak}, G.~J., {McClintock}, J.~E., {Remillard}, R.~A., {Cui}, W., {Levine},
861:   A.~M., {Morgan}, E.~H., {Orosz}, J.~A., \& {Bailyn}, C.~D.,  2000, ApJ, 544,
862:   993
863: 
864: \bibitem[\protect\astroncite{{Tagger} \& {Pellat}}{1999}]{Tagger99}
865: {Tagger}, M., \& {Pellat}, R.,  1999, A\&A, 349, 1003
866: 
867: \bibitem[\protect\astroncite{{Tanaka} \& {Lewin}}{1995}]{Tanaka95}
868: {Tanaka}, Y., \& {Lewin}, W. H.~G.,  1995,
869: \newblock in {X}-ray {B}inaries, ed. W.~H.~G. {Lewin}, J. {Van Paradjs},
870:   E.~P.~J. {Van den Heuvel},  (Cambridge: Cambridge {U}. {P}ress),  126
871: 
872: \bibitem[\protect\astroncite{{Tomsick}, {Corbel} \&
873:   {Kaaret}}{2001}]{Tomsick01b}
874: {Tomsick}, J.~A., {Corbel}, S., \& {Kaaret}, P.,  2001, ApJ, 563, 229
875: 
876: \bibitem[\protect\astroncite{{Tomsick} \& {Kaaret}}{2000}]{Tomsick00}
877: {Tomsick}, J.~A., \& {Kaaret}, P.,  2000, ApJ, 537, 448
878: 
879: \bibitem[\protect\astroncite{{Tomsick} et~al.}{2003}]{Tomsick03_2}
880: {Tomsick}, J.~A., {Kalemci}, E., {Corbel}, S., \& {Kaaret}, P.,  2003, ApJ, 
881: 592, 1100
882: 
883: \bibitem[\protect\astroncite{{van Straaten} et~al.}{2002}]{vanStraaten01}
884: {van Straaten}, S., {van der Klis}, M., {di Salvo}, T., \& {Belloni}, T.,
885:   2002, ApJ, 568, 912
886: 
887: \bibitem[\protect\astroncite{{Zdziarski} et~al.}{2002}]{Zdziarski02_2}
888: {Zdziarski}, A.~A., {Poutanen}, J., {Paciesas}, W.~S., \& {Wen}, L.,  2002,
889:   ApJ, 578, 357
890: 
891: \end{thebibliography}
892: 
893: \clearpage
894: 
895: %% /*******************************************************************
896: %% ** Figures                                                        **
897: %% *******************************************************************/
898: 
899: 
900: %Tables?
901: 
902: \begin{sidewaystable}[t]%%
903: \scriptsize
904: %\centering
905: \caption[Parameters for \xs]%%
906: {\label{table:1650} Paramters for \xs\footnote{\scriptsize This table is available only on-line 
907: as a machine-readable table.}
908: }%%
909: \begin{minipage}{\textwidth}%\centering
910: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
911: Obs. & Date \footnote{MJD, Modified Julian Date, same for all Tables} & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$\footnote{Lowest peak frequency, same for all Tables.} & QPO Freq.\footnote{Resonance frequency of the QPO if present, same for all Tables.} \\ \hline
912: 1 &      52227.5 &  $4.390 \pm 0.343$ & $0.484 \pm 0.002$ & $2.400 \pm 0.264$ & $5.255 \pm 0.052$ & $0.313 \pm 0.030$ & $<$4\footnote{All upper limits are 2$\sigma$ in all
913: Tables.} & - & - \\
914:  2 &      52228.1 &  $4.026 \pm 0.271$ & $0.464 \pm 0.007$ & $2.211 \pm 0.221$ & $5.822 \pm 0.058$ & $0.275 \pm 0.024$ & $<4$ & - & - \\
915:  3 &      52230.4 &  $2.548 \pm 0.140$ & $0.487 \pm 0.008$ & $3.747 \pm 0.337$ & $5.080 \pm 0.050$ & $0.424 \pm 0.036$ & $<$4 & - & - \\
916:  4\footnote{The state transition happened between this observation and the next observation. The values in Figs.~\ref{fig:evolmix1},~\ref{fig:evolmix2},~\ref{fig:evolplf}
917: are normalized using parameters in this observation.Fig.~\ref{fig:evolcf} 
918: observations are normalized with respect to the next observation. Same for all 
919: tables.} &      52231.0 &  $2.415 \pm 0.038$ & $0.459 \pm 0.004$ & $4.524 \pm 0.135$ & $4.731 \pm 0.047$ & $0.488 \pm 0.015$ & $<$4 & - & - \\ \hline
920: 5 &      52232.0 &  $2.379 \pm 0.019$ & $0.489 \pm 0.009$ & $12.76 \pm 0.127$ & $4.171 \pm 0.041$ & $0.753 \pm 0.009$ & $12.42 \pm 0.546$ & $4.162 \pm 0.498$ & - \\
921:  6 &      52233.2 &  $2.326 \pm 0.023$ & $0.443 \pm 0.016$ & $20.49 \pm 0.204$ & $2.221 \pm 0.022$ & $0.902 \pm 0.010$ & $19.01 \pm 1.672$ & $2.634 \pm 0.202$ & $8.734 \pm 0.219 $ \\
922:  7 &      52234.5 &  $2.093 \pm 0.011$ & $0.410 \pm 0.035$ & $30.62 \pm 0.306$ & $0.767 \pm 0.030$ & $0.975 \pm 0.012$ & $22.28 \pm 1.992$ & $1.274 \pm 0.123$ & $5.131 \pm 0.090 $ \\
923: 8 &      52235.1 &  $2.096 \pm 0.009$ & $0.355 \pm 0.021$ & $29.70 \pm 0.297$ & $0.569 \pm 0.022$ & $0.981 \pm 0.012$ & $23.04 \pm 1.287$ & $1.408 \pm 0.089$ & $5.440 \pm 0.087 $ \\
924:  9 &      52236.1 &  $2.001 \pm 0.009$ & $0.374 \pm 0.033$ & $28.70 \pm 0.287$ & $0.517 \pm 0.036$ & $0.982 \pm 0.012$ & $25.16 \pm 1.614$ & $0.881 \pm 0.054$ & $4.655 \pm 0.072 $ \\
925: 10 &      52236.9 &  $1.976 \pm 0.007$ & $0.321 \pm 0.011$ & $30.58 \pm 0.305$ & $0.271 \pm 0.027$ & $0.991 \pm 0.012$ & $24.75 \pm 0.330$ & $0.999 \pm 0.028$ & $3.799 \pm 0.034 $ \\
926: 11 &      52237.9 &  $1.961 \pm 0.008$ & $0.306 \pm 0.006$ & $29.59 \pm 0.295$ & $0.144 \pm 0.014$ & $0.995 \pm 0.012$ & $25.45 \pm 2.449$ & $0.974 \pm 0.073$ & $3.830 \pm 0.077 $ \\ \hline
927: \end{tabular}
928: \end{minipage}
929: \end{sidewaystable}
930: 
931: \begin{sidewaystable}[t]%%
932: \scriptsize
933: %\centering
934: \caption[Parameters for \xf\ in 1998]%%
935: {\label{table:1550x1} Paramters for \xf\ in 1998}%%
936: \begin{minipage}{\textwidth}%\centering
937: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
938: Obs. & Date  & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$ & QPO Freq. \\ \hline
939: 1 &      1298.09 &  $2.439 \pm 0.134$ & $0.415 \pm 0.011$ & $0.867 \pm 0.065$ & $0.547 \pm 0.010$ & $0.613 \pm 0.048$ & $<$10 & - & - \\
940:  2 &      1299.34 &  $2.688 \pm 0.200$ & $0.392 \pm 0.011$ & $0.567 \pm 0.051$ & $0.389 \pm 0.007$ & $0.593 \pm 0.055$ &  $<$10 & - & - \\
941:  3 &      1302.46 &  $2.754 \pm 0.278$ & $0.394 \pm 0.023$ & $0.444 \pm 0.066$ & $0.219 \pm 0.004$ & $0.669 \pm 0.106$ &  $<$10 & - & - \\
942:  4 &      1303.39 &  $2.635 \pm 0.230$ & $0.379 \pm 0.027$ & $0.532 \pm 0.074$ & $0.187 \pm 0.005$ & $0.739 \pm 0.114$ &  $<$10 & - & - \\
943:  5 &      1305.12 &  $2.047 \pm 0.141$ & $0.464 \pm 0.030$ & $1.038 \pm 0.072$ & $0.228 \pm 0.006$ & $0.819 \pm 0.066$ & $<$10 & - & - \\ \hline
944:  6 &      1307.32 &  $1.968 \pm 0.031$ & $0.479 \pm 0.104$ & $3.746 \pm 0.074$ & $0.065 \pm 0.005$ & $0.982 \pm 0.024$ & $22.13 \pm 1.590$ & $1.175 \pm 0.210$ & - \\
945:  7 &      1309.72 &  $1.739 \pm 0.036$ & $0.595 \pm 0.084$ & $2.377 \pm 0.047$ & $0.092 \pm 0.007$ & $0.962 \pm 0.024$ & $17.74 \pm 0.600$ & $0.551 \pm 0.036$ & - \\
946: 8 &      1311.52 &  $1.785 \pm 0.071$ & $0.518 \pm 0.073$ & $1.216 \pm 0.048$ & $0.069 \pm 0.006$ & $0.946 \pm 0.047$ & $14.81 \pm 1.590$ & $0.144 \pm 0.024$ & - \\
947:  9 &      1313.32 &  $1.773 \pm 0.088$ & $0.489 \pm 0.058$ & $0.861 \pm 0.051$ & $0.061 \pm 0.006$ & $0.933 \pm 0.068$ & $<$8 & - & - \\
948: 10 &      1318.77 &  $1.815 \pm 0.132$ & $0.455 \pm 0.062$ & $0.491 \pm 0.039$ & $0.045 \pm 0.004$ & $0.916 \pm 0.089$ & $<$8 & - & - \\ \hline
949: \end{tabular}
950: \end{minipage}
951: \end{sidewaystable}
952: 
953: \begin{sidewaystable}[t]%%
954: \scriptsize
955: %\centering
956: \caption[Parameters for \xf\ in 2000]%%
957: {\label{table:1550x2} Paramters for \xf\ in 2000}%%
958: \begin{minipage}{\textwidth}%\centering
959: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
960: Obs. & Date  & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$ & QPO Freq. \\ \hline
961: 1 &      51667.7 &  $2.368 \pm 0.016$ & $0.798 \pm 0.010$ & $6.221 \pm 0.124$ & $4.539 \pm 0.045$ & $0.578 \pm 0.012$ & $12.50 \pm 0.470$ & $10.01 \pm 1.264$ & $9.030 \pm 0.240 $ \\
962:  2 &      51669.6 &  $2.386 \pm 0.014$ & $0.763 \pm 0.007$ & $5.170 \pm 0.103$ & $4.142 \pm 0.041$ & $0.555 \pm 0.012$ & $12.37 \pm 0.330$ & $9.571 \pm 0.894$ & $8.160 \pm 0.136 $ \\
963:  3 &      51670.6 &  $2.363 \pm 0.014$ & $0.756 \pm 0.010$ & $6.263 \pm 0.125$ & $3.307 \pm 0.033$ & $0.654 \pm 0.014$ & $13.12 \pm 0.280$ & $9.951 \pm 0.601$ & $9.320 \pm 0.088 $ \\
964:  4 &      51670.8 &  $2.362 \pm 0.015$ & $0.761 \pm 0.008$ & $5.467 \pm 0.109$ & $3.669 \pm 0.036$ & $0.598 \pm 0.013$ & $13.36 \pm 0.840$ & $9.615 \pm 0.932$ & $8.610 \pm 0.116 $ \\
965:  5 &      51671.4 &  $2.366 \pm 0.017$ & $0.748 \pm 0.008$ & $5.022 \pm 0.100$ & $3.550 \pm 0.035$ & $0.585 \pm 0.012$ & $13.06 \pm 0.690$ & $9.699 \pm 0.653$ & - \\
966:  6 &      51672.4 &  $2.342 \pm 0.015$ & $0.741 \pm 0.008$ & $5.031 \pm 0.100$ & $3.030 \pm 0.030$ & $0.624 \pm 0.013$ & $15.21 \pm 0.790$ & $8.906 \pm 0.768$ & $8.810 \pm 0.141 $ \\ \hline
967:  7 &      51673.0 &  $2.414 \pm 0.015$ & $0.761 \pm 0.014$ & $8.486 \pm 0.169$ & $2.857 \pm 0.028$ & $0.748 \pm 0.017$ & $7.230 \pm 0.260$ & $5.959 \pm 1.037$ & - \\
968: 8 &      51673.4 &  $2.422 \pm 0.014$ & $0.759 \pm 0.012$ & $7.261 \pm 0.145$ & $2.798 \pm 0.027$ & $0.721 \pm 0.016$ & $8.440 \pm 0.290$ & $6.735 \pm 1.076$ & - \\
969:  9 &      51674.7 &  $2.331 \pm 0.014$ & $0.756 \pm 0.014$ & $8.471 \pm 0.169$ & $2.400 \pm 0.024$ & $0.779 \pm 0.017$ & $14.77 \pm 0.290$ & $1.019 \pm 0.257$ & $3.580 \pm 0.015 $ \\
970: 10 &      51675.5 &  $2.284 \pm 0.012$ & $0.696 \pm 0.014$ & $7.164 \pm 0.143$ & $1.245 \pm 0.012$ & $0.851 \pm 0.019$ & $16.38 \pm 0.520$ & $1.794 \pm 0.137$ & $7.690 \pm 0.089 $ \\
971: 11 &      51676.4 &  $2.275 \pm 0.014$ & $0.654 \pm 0.017$ & $6.346 \pm 0.126$ & $0.942 \pm 0.009$ & $0.870 \pm 0.020$ & $18.77 \pm 0.960$ & $2.425 \pm 0.429$ & $6.930 \pm 0.020 $ \\
972: 12 &      51678.5 &  $2.116 \pm 0.011$ & $0.613 \pm 0.033$ & $6.233 \pm 0.124$ & $0.417 \pm 0.012$ & $0.937 \pm 0.022$ & $23.88 \pm 0.890$ & $1.148 \pm 0.092$ & $4.490 \pm 0.014 $ \\
973: 13 &      51680.4 &  $2.143 \pm 0.009$ & $0.505 \pm 0.022$ & $5.489 \pm 0.164$ & $0.190 \pm 0.007$ & $0.966 \pm 0.035$ & $24.19 \pm 0.950$ & $1.014 \pm 0.083$ & $4.090 \pm 0.015 $ \\
974: 14 &      51682.3 &  $1.883 \pm 0.007$ & $0.387 \pm 0.011$ & $4.520 \pm 0.135$ & $0.150 \pm 0.007$ & $0.967 \pm 0.035$ & $25.70 \pm 1.530$ & $0.558 \pm 0.029$ & $2.340 \pm 0.010 $ \\ \hline
975: \end{tabular}
976: \end{minipage}
977: \end{sidewaystable}
978: 
979: \begin{sidewaystable}[t]%%
980: \scriptsize
981: %\centering
982: \caption[Parameters for \fu\ in 1998]%%
983: {\label{table:1630x1} Paramters for \fu\ in 1998}%%
984: \begin{minipage}{\textwidth}%\centering
985: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
986: Obs. & Date  & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$ & QPO Freq. \\ \hline
987: 1 &     50937.630 &  $2.367 \pm 0.067$ & $0.764 \pm 0.014$ & $19.03 \pm 0.380$ & $10.05 \pm 0.100$ & $0.654 \pm 0.014$ & $3.000 \pm 0.200$ & - & - \\
988:  2 &    50939.070 &  $2.356 \pm 0.052$ & $0.738 \pm 0.012$ & $18.71 \pm 0.374$ & $8.851 \pm 0.088$ & $0.678 \pm 0.015$ & $3.500 \pm 0.200$ & - & - \\
989:  3 &     50942.010 &  $2.422 \pm 0.068$ & $0.758 \pm 0.011$ & $15.49 \pm 0.309$ & $10.24 \pm 0.102$ & $0.602 \pm 0.013$ & $2.500 \pm 0.300$ & - & - \\
990:  4 &    50945.860 &  $2.317 \pm 0.092$ & $0.711 \pm 0.016$ & $12.70 \pm 0.254$ & $6.887 \pm 0.068$ & $0.648 \pm 0.014$ & $4.000 \pm 0.500$ & - & - \\
991:  5 &    50949.740 &  $2.072 \pm 0.058$ & $0.601 \pm 0.012$ & $10.71 \pm 0.214$ & $3.826 \pm 0.038$ & $0.736 \pm 0.016$ & $2.000 \pm 0.800$ & - & - \\ \hline
992:  6 &    50951.670 &  $1.916 \pm 0.027$ & $0.461 \pm 0.017$ & $13.33 \pm 0.266$ & $1.201 \pm 0.024$ & $0.917 \pm 0.022$ & $10.20 \pm 0.600$ & $4.235 \pm 0.337$ & $3.390 \pm 0.008 $ \\
993:  7 &    50952.490 &  $1.858 \pm 0.032$ & $0.466 \pm 0.024$ & $9.092 \pm 0.181$ & $0.980 \pm 0.029$ & $0.902 \pm 0.022$ & $11.30 \pm 0.800$ & $3.540 \pm 0.247$ & $2.613 \pm 0.012 $ \\
994: 8 &    50953.490 &  $1.798 \pm 0.027$ & $0.487 \pm 0.024$ & $10.64 \pm 0.212$ & $0.822 \pm 0.024$ & $0.928 \pm 0.022$ & $13.60 \pm 0.700$ & $3.140 \pm 0.128$ & $1.351 \pm 0.012 $ \\
995:  9 &    50956.960 &  $1.793 \pm 0.037$ & $0.459 \pm 0.026$ & $7.013 \pm 0.140$ & $0.510 \pm 0.025$ & $0.932 \pm 0.023$ & $15.90 \pm 1.100$ & $1.187 \pm 0.350$ & $0.430 \pm 0.006 $ \\
996: 10 &    50958.960 &  $1.675 \pm 0.031$ & $0.486 \pm 0.025$ & $7.442 \pm 0.148$ & $0.565 \pm 0.028$ & $0.929 \pm 0.023$ & $16.10 \pm 1.500$ & $0.922 \pm 0.197$ & $0.365 \pm 0.011 $ \\
997: 11 &    50962.960 &  $1.584 \pm 0.024$ & $0.452 \pm 0.029$ & $8.827 \pm 0.176$ & $0.406 \pm 0.040$ & $0.956 \pm 0.024$ & $17.30 \pm 0.800$ & $0.887 \pm 0.058$ & $0.228 \pm 0.003 $ \\ \hline
998: \end{tabular}
999: \end{minipage}
1000: \end{sidewaystable}
1001: 
1002: \begin{sidewaystable}[t]%%
1003: \scriptsize
1004: %\centering
1005: \caption[Parameters for \fu\ in 1999]%%
1006: {\label{table:1630x2} Paramters for \fu\ in 1999}%%
1007: \begin{minipage}{\textwidth}%\centering
1008: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
1009: Obs. & Date  & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$ & QPO Freq. \\ \hline
1010: 1 &      51388.3 &  $2.060 \pm 0.036$ & $0.511 \pm 0.014$ & $9.891 \pm 0.197$ & $2.414 \pm 0.048$ & $0.803 \pm 0.019$ & $<$2 & - & - \\
1011:  2 &      51389.1 &  $1.987 \pm 0.028$ & $0.519 \pm 0.011$ & $10.09 \pm 0.201$ & $2.274 \pm 0.045$ & $0.816 \pm 0.019$ & $<$2 & - & - \\
1012:  3 &      51390.1 &  $2.036 \pm 0.032$ & $0.529 \pm 0.014$ & $8.340 \pm 0.166$ & $2.140 \pm 0.042$ & $0.795 \pm 0.019$ & $<$2 & - & - \\
1013:  4 &      51391.4 &  $2.014 \pm 0.032$ & $0.500 \pm 0.016$ & $7.680 \pm 0.153$ & $1.557 \pm 0.046$ & $0.831 \pm 0.020$ & $<$2 & - & - \\
1014:  5 &      51393.2 &  $1.834 \pm 0.024$ & $0.464 \pm 0.014$ & $8.301 \pm 0.166$ & $1.220 \pm 0.061$ & $0.871 \pm 0.022$ & $<$2 & - & - \\ \hline
1015:  6 &      51394.8 &  $1.845 \pm 0.024$ & $0.469 \pm 0.027$ & $7.294 \pm 0.145$ & $0.796 \pm 0.055$ & $0.901 \pm 0.023$ & $12.50 \pm 2.850$ & $4.090 \pm 0.089$ & $1.820 \pm 0.004 $ \\
1016:  7 &      51395.5 &  $1.881 \pm 0.031$ & $0.429 \pm 0.032$ & $6.033 \pm 0.120$ & $0.694 \pm 0.062$ & $0.896 \pm 0.024$ & $8.553 \pm 1.000$ & $0.603 \pm 0.517$ & $0.782 \pm 0.083 $ \\
1017: 8 &      51397.2 &  $2.041 \pm 0.076$ & $0.424 \pm 0.058$ & $3.483 \pm 0.104$ & $0.438 \pm 0.043$ & $0.888 \pm 0.035$ & $<$4 & - & - \\
1018:  9 &      51398.2 &  $1.802 \pm 0.093$ & $0.378 \pm 0.041$ & $3.854 \pm 0.154$ & $0.351 \pm 0.035$ & $0.916 \pm 0.046$ & $<$4 & - & - \\
1019: 10 &      51400.1 &  $2.048 \pm 0.061$ & $0.425 \pm 0.053$ & $2.852 \pm 0.142$ & $0.251 \pm 0.025$ & $0.919 \pm 0.057$ & $<$4 & - & - \\ \hline
1020: \end{tabular}
1021: \end{minipage}
1022: \end{sidewaystable}
1023: 
1024: 
1025: \begin{sidewaystable}[t]%%
1026: \scriptsize
1027: %\centering
1028: \caption[Parameters for \fu\ in 2001]%%
1029: {\label{table:1630x3} Paramters for \fu\ in 2001}%%
1030: \begin{minipage}{\textwidth}%\centering
1031: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
1032: Obs. & Date  & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$ & QPO Freq. \\ \hline
1033: 1 &      52048.0 &  $2.219 \pm 0.039$ & $0.784 \pm 0.012$ & $18.01 \pm 0.360$ & $15.18 \pm 0.151$ & $0.542 \pm 0.011$ & $<$2 & - & - \\
1034:  2 &      52049.8 &  $2.236 \pm 0.034$ & $0.702 \pm 0.012$ & $15.27 \pm 0.305$ & $10.19 \pm 0.101$ & $0.599 \pm 0.013$ & $<$2 & - & - \\
1035:  3 &      52051.6 &  $2.260 \pm 0.029$ & $0.674 \pm 0.012$ & $14.32 \pm 0.286$ & $7.206 \pm 0.072$ & $0.665 \pm 0.014$ & $<$2 & - & - \\
1036:  4 &      52053.8 &  $2.095 \pm 0.020$ & $0.592 \pm 0.011$ & $13.25 \pm 0.265$ & $4.642 \pm 0.046$ & $0.740 \pm 0.016$ & $<$2 & - & - \\
1037:  5 &      52055.6 &  $2.230 \pm 0.038$ & $0.542 \pm 0.014$ & $9.742 \pm 0.194$ & $3.458 \pm 0.069$ & $0.738 \pm 0.017$ & $<$2 & - & - \\
1038:  6 &      52058.0 &  $1.970 \pm 0.030$ & $0.460 \pm 0.017$ & $10.29 \pm 0.205$ & $1.760 \pm 0.052$ & $0.853 \pm 0.021$ & $<$2 & - & - \\ \hline
1039:  7 &      52059.4 &  $1.575 \pm 0.010$ & $0.401 \pm 0.016$ & $11.06 \pm 0.221$ & $0.895 \pm 0.062$ & $0.925 \pm 0.024$ & $11.72 \pm 2.730$ & $2.520 \pm 1.070$ & $1.396 \pm 0.042 $ \\
1040: 8 &      52061.6 &  $1.683 \pm 0.030$ & $0.463 \pm 0.023$ & $4.766 \pm 0.142$ & $0.767 \pm 0.053$ & $0.861 \pm 0.033$ & $9.323 \pm 3.150$ & $0.297 \pm 0.069$ & - \\
1041:  9 &      52063.6 &  $1.751 \pm 0.026$ & $0.444 \pm 0.014$ & $3.961 \pm 0.158$ & $0.590 \pm 0.047$ & $0.870 \pm 0.044$ & $<$5 & - & - \\
1042: 10 &      52065.9 &  $1.694 \pm 0.021$ & $0.430 \pm 0.025$ & $3.880 \pm 0.194$ & $0.491 \pm 0.049$ & $0.887 \pm 0.056$ & $<$5 & - & - \\ \hline
1043: \end{tabular}
1044: \end{minipage}
1045: \end{sidewaystable}
1046: 
1047: \begin{sidewaystable}[t]%%
1048: \scriptsize
1049: %\centering
1050: \caption[Parameters for \reso]%%
1051: {\label{table:1748} Paramters for \reso}%%
1052: \begin{minipage}{\textwidth}%\centering
1053: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
1054: Obs. & Date & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$ & QPO Freq. \\ \hline
1055: 1 &      51002.7 &  $2.643 \pm 0.031$ & $0.892 \pm 0.015$ & $17.73 \pm 0.354$ & $9.967 \pm 0.099$ & $0.640 \pm 0.014$ & $<$2 & - & - \\
1056:  2 &      51007.3 &  $1.967 \pm 0.017$ & $0.437 \pm 0.015$ & $19.62 \pm 0.392$ & $2.496 \pm 0.074$ & $0.887 \pm 0.021$ & $4.250 \pm 0.250$ & - & - \\ \hline
1057:  3 &      51012.2 &  $1.874 \pm 0.005$ & $0.396 \pm 0.030$ & $10.07 \pm 0.201$ & $0.253 \pm 0.025$ & $0.975 \pm 0.024$ & $21.14 \pm 3.160$ & $1.104 \pm 0.186$ & - \\
1058:  4 &      51024.4 &  $1.828 \pm 0.013$ & $0.399 \pm 0.019$ & $8.879 \pm 0.177$ & $0.270 \pm 0.027$ & $0.970 \pm 0.024$ & $20.71 \pm 0.370$ & $0.220 \pm 0.014$ & - \\ \hline
1059: \end{tabular}
1060: \end{minipage}
1061: \end{sidewaystable}
1062: 
1063: \begin{sidewaystable}[t]%%
1064: \scriptsize
1065: %\centering
1066: \caption[Parameters for GRO~J1655$-$40]%%
1067: {\label{table:1655} Paramters for GRO~J1655$-$40}%%
1068: \begin{minipage}{\textwidth}%\centering
1069: \begin{tabular}{l|c|ccccc|ccc} \hline \hline
1070: Obs. & Date & \ind & \tin & Power law Fl. & Dbb Fl. & PLR & rms (\%) & $\nu_{1}$ & QPO Freq. \\ \hline
1071: 1 &      50658.4 &  $2.106 \pm 0.029$ & $0.898 \pm 0.004$ & $16.25 \pm 0.325$ & $57.22 \pm 0.572$ & $0.221 \pm 0.004$ & $1.800 \pm 0.100$ & - & - \\
1072:  2 &      50663.7 &  $2.392 \pm 0.092$ & $0.795 \pm 0.003$ & $9.031 \pm 0.180$ & $29.68 \pm 0.296$ & $0.233 \pm 0.004$ & $<$2 & - & - \\ \hline
1073:  3 &      50674.4 &  $1.948 \pm 0.010$ & $0.610 \pm 0.027$ & $27.59 \pm 0.551$ & $1.220 \pm 0.036$ & $0.957 \pm 0.023$ & $22.66 \pm 0.770$ & $1.514 \pm 0.077$ & $6.457 \pm 0.020 $ \\
1074:  4 &      50678.6 &  $1.670 \pm 0.015$ & $0.629 \pm 0.077$ & $9.284 \pm 0.278$ & $0.203 \pm 0.020$ & $0.978 \pm 0.036$ & $24.34 \pm 2.390$ & $0.313 \pm 0.027$ & $0.785 \pm 0.011 $ \\
1075:  5 &      50685.4 &  $1.751 \pm 0.057$ & $0.485 \pm 0.026$ & $0.971 \pm 0.058$ & $0.072 \pm 0.007$ & $0.930 \pm 0.068$ & $<$5 & - & - \\ \hline
1076: \end{tabular}
1077: \end{minipage}
1078: \end{sidewaystable}
1079: 
1080: \end{document}
1081: