1: \documentclass[12pt,preprint]{aastex}
2:
3: \shorttitle{A Massive Jet Ejection from SS~433}
4: \shortauthors{Kotani et al.}
5:
6: \begin{document}
7:
8: \title{A Massive Jet Ejection Event from the Microquasar SS~433
9: Accompanying Rapid X-Ray Variability}
10:
11: \author{T. Kotani\altaffilmark{1}, S. A. Trushkin\altaffilmark{2},
12: R. Valiullin\altaffilmark{3},
13: K. Kinugasa\altaffilmark{4},
14: S. Safi-Harb\altaffilmark{5},
15: N. Kawai\altaffilmark{1}, and
16: M. Namiki\altaffilmark{6}}
17: \email{kotani@hp.phys.titech.ac.jp}
18:
19: \altaffiltext{1}{Tokyo Tech, 2-12-1 O-okayama, Tokyo 152-8551, Japan}
20: \altaffiltext{2} {Special Astrophysical Observatory RAS, Nizhnij Arkhyz,
21: Karachaevo-Cherkassia 369167, Russia}
22: \altaffiltext{3} {Astrophysical Institute of Kazakh Academy of
23: Sciences, 480020 Alma Ata, Kazakhstan}
24: \altaffiltext{4} {Gunma Astronomical Observatory, 6860-86 Nakayama,
25: Takayama, Agatsuma, Gunma 377-0702, Japan}
26: \altaffiltext{5} { University of Manitoba, Winnipeg, Manitoba, Canada}
27: \altaffiltext{6} {Osaka University, 1-1 Machikaneyama, Toyonaka, Osaka
28: 560-0043, Japan}
29:
30: \begin{abstract}
31: Microquasars occasionally exhibit massive jet ejections which are
32: distinct from the continuous or quasi-continuous weak jet ejections.
33: Because those massive jet ejections are rare and short events, they have
34: hardly been observed in X-ray so far. In this paper, the first X-ray
35: observation of a massive jet ejection from the microquasar SS 433 with
36: the Rossi X-ray Timing Explorer (RXTE) is reported. SS 433 undergoing a
37: massive ejection event shows a variety of new phenomena including a
38: QPO-like feature near 0.1 Hz, rapid time variability, and shot-like
39: activities. The shot-like activity may be caused by the formation of a
40: small plasma bullet. A massive jet may be consist of thousands of those
41: plasma bullets ejected from the binary system. The size, mass, internal
42: energy, and kinetic energy of the bullets and the massive jet are
43: estimated.
44: \end{abstract}
45: \keywords{X-rays: individual (\objectname{SS 433})}
46:
47: \section{Introduction}
48: Microquasars are stellar X-ray binaries (neutron stars or black holes)
49: from which relativistic jets emanate via an unknown, very efficient
50: mechanism \citep{mirabel99}. Microquasars such as SS 433 and GRS
51: 1915+105 occasionally exhibit massive jet ejections, which are
52: recognized as sporadic flares in their radio light curves
53: \citep{fiedler87,foster96}. Because the massive jet ejections are rare
54: (a few per year), short (within a few days), and aperiodic, pointing
55: X-ray observations of these events have hardly been performed so far.
56: As for SS~433, no X-ray observation has been confirmed to coincide with
57: a radio flare, except for one or two possible coincident observations
58: with Einstein in 1979 \citep{band89}. A monitoring observation over 10
59: days and a long-look observation lasting 13 days were performed with
60: ASCA in 1995 and 2000 \citep{kotani97, namiki01}, but there was no radio
61: flare coinciding the periods. A multi-wavelength observation with RXTE
62: and the Giant Meter Radio Telescope in 2002 also misses radio flares
63: \citep{chakrabarti03}. It should be stressed that the massive jets are
64: distinct from the stable continuous jets of SS~433 and the
65: quasi-continuous or weak jet of GRS~1915+105.
66: % SS 433 is unique even among
67: % microquasars.
68: % The continuous jets of SS 433 have been stable for more than two
69: % decades, while the jets of other sources is intermittent, or in some
70: % cases, have been detected only once at the discovery but have never been
71: % active again \citep{rupen98}. The so-called ``moving emission
72: % lines'' in the optical and X-ray spectra of SS 433, which indicate that
73: % matter is ejected at a quarter of the speed of light, have been detected
74: % in most observation
75: % occasions \citep{goranskii98,kotani97}. On the other
76: % hand, the radio counterpart of SS 433 is not at all steady.
77: The radio activity of SS~433 monitored with the Green Bank
78: Interferometer over years may be characterized as a clustering of flare
79: events separated by periods of quiescent emission \citep{fiedler87}. In
80: those sporadic radio flare events, the radio flux density at 2.3 GHz
81: exceeds 1 Jy, and massive jet blobs, which are recognized as bright
82: extended spots in radio images, are ejected from the core of SS 433 at a
83: quarter of the speed of light \citep{vermeulen93}. The ejection of
84: massive jet blobs from GRS 1915+105 with a radio flux exceeding 100 mJy,
85: by which the source has been recognized as a microquasar in the first
86: place \citep{mirabel94,fender01}, have been hardly observed in X rays
87: \citep{muno01}, in contrast to a number of reports on the X-ray
88: observation of the quasi-continuous or weak jet ejections
89: \citep{mirabel98,klein-wolt02,ueda02}.
90:
91:
92: We report on a successful X-ray
93: observation of a massive jet ejection from SS~433 with the Rossi X-ray
94: Timing Explorer (RXTE)\@. The observation scheme is described in
95: \S~\ref{sec:obs}, the data are analyzed and discussed in
96: \S~\ref{sec:analysis}.
97:
98:
99:
100:
101: \section{Observations}\label{sec:obs}
102: Formerly, a radio flare was the only indicator of a massive jet
103: ejection. Unfortunately, an X-ray observation triggered by a radio
104: flare is too late to catch the moment of the ejection, as experienced in
105: the cases of several previous target-of-opportunity (TOO) X-ray
106: observations. Because the X-ray activity precedes a radio flare, a TOO
107: X-ray observation will not work for a massive jet ejection event. So we
108: have built a strategy to observe a {\it second}\/ massive jet ejection
109: event following the first event. In the active state of SS~433, radio
110: flares are clustered with an interval of~$8-23$ days \citep{fiedler87}.
111: Therefore, a series of monitoring observations triggered by a massive
112: jet ejection may cover the moment of a second ejection in~23 days.
113:
114: We planned a 30-days-long TOO monitoring observation of SS~433 with RXTE
115: to be triggered by a radio flare. The proposal was accepted in the
116: Cycle 6 of the RXTE Guest Observer Program carried out for one year
117: beginning in March 2001. The daily radio activity of the source has
118: been monitored with the RATAN-600 radio telescope \citep{korolkov79} of
119: the Special Astrophysical observatory of the Russian Academy of Sciences
120: (SAO RAS) since September 2001. After two months of static activity
121: with an average flux density of 0.7 Jy at 2.3 GHz, a remarkable flare
122: occurred on 2001 November 2 (MJD = 52215), indicating that the source
123: entered its active state (Fig.~\ref{fig:lc_x_radio}). Flux densities
124: reached 1.3 Jy at 2.3 GHz on MJD = 52216.6 \citep{kotani01, trushkin03}.
125: We started a series of X-ray observations with RXTE on MJD = 52222
126: \citep{kotani03, safi-harb03}. Except for a break at MJD = 52231, SS 433
127: was observed for 3 ks every day. In the X-ray light curve, a temporal
128: variation with time scales of $10-100$ s appeared on MJD = 52225 and the
129: amplitude increased day by day (Fig.~\ref{fig:lc_x}). On MJD = 52232,
130: the amplitude reached a maximum, and the 2-10 keV X-ray flux reached a
131: local maximum of $2.5\times10^{-10}$ erg s$^{-1}$ cm$^{-2}$. The X-ray
132: emission, thought to originate in the hot part of the jets as long as or
133: longer than $10^{12}$ cm, had never shown such a variability in past
134: observations \citep{safi-harb03}. Following the maximum of the flux and
135: the variation amplitude, a second radio flare was detected on MJD =
136: 52235. Due to a missing radio data point at MJD = 52234, the precise
137: onset time and peak flux of the second flare are unfortunately not
138: known, but they are not likely out of a range $52233 < $MJD$ \leq 52235$
139: and $1.5$ Jy $<$ F $<$ 2 Jy. Thus we conclude that the moment of a
140: massive jet ejection was observed in the X-ray band. After the peak,
141: the X-ray flux dropped due to a binary eclipse. The X-ray monitoring
142: observation lasted until MJD = 52238, for 17 days, providing 16 data
143: sets. An observation log is shown in Table~\ref{tbl:log}. Optical
144: spectroscopic observations were performed on MJD = 52220.6, 52221.6, and
145: 52225.6 with the 0.7-m telescope at the observatory Kamenskoe Plato
146: \citep{mironov98}, and on MJD = 52229.39 and 52233.38 with the 0.65-m
147: telescope at the Gunma Astronomical Observatory
148: \citep{hasegawa04,kinugasa02}. Based on the spectroscopic data, the
149: variation of the Doppler shifts of the jets during the campaign are
150: estimated. The Doppler parameter of the receding jet is estimated to
151: increase from $0.07$ on MJD = 52222 to $0.13$ on 52238, and that of the
152: approaching jet decrease from $-0.02$ to $-0.07$.
153:
154: \section{Data Analysis and Discussion}\label{sec:analysis}
155: \subsection{The QPO-like feature}
156: Firstly, we have
157: searched for a periodicity in the data. No coherent pulsation has been
158: detected from the 16 data sets, but a feature which can be interpreted
159: as a QPO has been found at 0.1 Hz in the sum of the 16 power density
160: spectra.
161: The sum of the power density spectra
162: is shown in Fig.~\ref{fig:pds}.
163: %We have tested the significance of the QPO-like feature by the method
164: %described by \citet{vaughan04}.
165: The fraction of the flux accounts for the QPO-like variation is
166: estimated from the ratio of the Gaussian normalization to the area under
167: the power-law continuum
168:
169: This is the first detection of any
170: periodicity or quasi periodicity shorter than~1 day from this source.
171: Interestingly, other microquasars such as GRS~1915+105 also show 0.5-10
172: Hz low-frequency QPOs, which are considered to represent a
173: characteristic time scale in the accretion flow \citep{muno01}.
174: A super-critical accretion flow, which SS~433 is believed
175: to have, had not been observed in the X-ray band because of the bright
176: jets. The 0.1 Hz QPO-like feature may be the first detection of the
177: super-critical accretion flow or disk in the X-ray band. The
178: similarity to the QPO in other microquasars suggests the presence of a
179: common mechanism working in other systems and SS~433, at least when the
180: latter is undergoing a massive jet ejection.
181:
182:
183: \subsection{Spectral fitting}
184: The sixteen data sets have been reducted with the standard reduction
185: method\footnote{RXTE GOF, http://heasarc.gsfc.nasa.gov/docs/xte/xhp\_{}proc\_{}analysis.html}.
186: The spectra are fitted with an empirical model,
187: \begin{equation}
188: e^{-\sigma(E) N_{\rm H}} \times \left [
189: \mbox{bremsstrahlung}(kT)
190: + F_{\rm N} \times {\rm narrow line}(E_{\rm N}, \sigma_{\rm N})
191: + F_{\rm B} \times {\rm borad line} (E_{\rm B}, \sigma_{\rm B})
192: \right ],
193: \end{equation}
194: where $\sigma(E)$ is the absorption cross section, $F_{\rm N}$ and
195: $F_{\rm B}$ are line fluxes, $E_{\rm N}$ and $E_{\rm B}$ are line center
196: energies, and $\sigma_{\rm N}$ and $\sigma_{\rm B}$ are line widths.
197: The hydrogen column density $N_{\rm H}$ and the width of the narrow line
198: $\sigma_{\rm N}$ are fixed to $6\times 10^{21}$ cm$^{-2}$ and 0 keV,
199: respectively. The results are shown in Table~\ref{tbl:best-fit}. The
200: model has been applied to SS~433's spectra obtained with the LAC/Ginga
201: \citep{kawai89, yuan95}, a proportional counter array whose energy
202: resolution and energy band are similar to those of the PCA\@. This
203: model is a simple approximation of the complicated, line-abundant
204: spectrum revealed with finer energy resolutions of SIS/ASCA
205: \citep{kotani96} and HETGS/Chandra \citep{marshall02}. In this model,
206: the Doppler-shifted pairs of Fe{\sc xxv} K$\alpha$, Fe{\sc xxvi}
207: K$\alpha$, and Ni{\sc xxvii} K$\alpha$ lines are blended into the
208: ``narrow'' and ``broad iron lines.'' The parameters which can not be
209: determined from an RXTE spectrum, such as line flux ratios Fe{\sc
210: xxvi}/Fe {\sc xxv} and red/blue, are naturally eliminated from the
211: model. The average spectrum of each data set and its evolution can be
212: reproduced with the model and the spectral parameters in
213: Table~\ref{tbl:best-fit} together with the 2-10 keV fluxes in
214: Fig.~\ref{fig:lc_x_radio}.
215:
216: In the eclipse at MJD = 52234, both the bremsstrahlung temperature and
217: the line fluxes drop, as observed with the LAC/Ginga \citep{kawai89,
218: yuan95}. The equivalent width of the two lines at the flux maximum (MJD
219: = 52232) and the eclipse (MJD = 52234) are 1.76 keV and 1.56 keV,
220: respectively. The equivalent width is not sensitive to eclipse because
221: the base of the jet, which is responsible to both of the Doppler-shifted
222: line emission and the continuum emission, is occulted in eclipse
223: \citep{kawai89, yuan95, gies02}.
224:
225:
226: \subsection{The rapid variability}
227: \paragraph{Data analysis}
228: Then we examined the rapid variation seen on MJD = 52232
229: (Fig.~\ref{fig:lc_x}). The variation, which might appear irregular or
230: chaotic, can be interpreted as a series of ``shots'' or ``spikes''with
231: widths of tens of seconds. Their intervals are random and do not show
232: any periodicity. We have sampled 12 shots as indicated in
233: Fig.~\ref{fig:lc_x}, and folded the light curve to make the average
234: profile of the shots (Fig.~\ref{fig:foldedshot}). The shot rises fast
235: then slightly softens during the decay. The 8.4-21 keV profile is
236: fitted with a burst model,
237: \begin{equation}
238: \mbox{const.} + A \times \left \{
239: \begin{array}{ll}
240: 0 & (t \leq t_0)\\
241: \frac{t-t_0}{-t_0} &(t_0 \leq t \leq 0)\\
242: \exp (-t/\tau_{\rm dec} ) & (0 \leq t)
243: \end{array}
244: \right .,
245: \end{equation}
246: where $A$ is a normalization factor, $t$ is the time from the peak,
247: $t_0$ is the time of the onset of the shot, and $\tau_{\rm dec}$ is the
248: decay time scale. The onset time and decay time scale are fitted to be
249: $-23^{+5}_{-4}$ and $41_{-9}^{+12}$ s, respectively.
250:
251: We have divided the profile into three phases, namely, the ``pre-shot''
252: phase, the ``peak'' phase, and the ``decay'' phase, and made a spectrum
253: from each phase. We have subtracted the pre-shot spectrum from each of
254: the peak and the decay spectra to extract the pure shot component. The
255: pure shot component is shown in Fig.~\ref{fig:spec}, together with the
256: pre-shot spectrum. The 3-20 keV fluxes in the peak and decay phases are
257: $1.9\times10^{-10}$ erg s$^{-1}$ cm$^{-2}$ and $5.7\times10^{-11}$ erg
258: s$^{-1}$ cm$^{-2}$, respectively. The shot component is well fit by
259: either an absorbed power-law model or an absorbed thermal bremsstrahlung
260: model. No emission line is detected. The total spectrum integrated over
261: all data taken on MJD = 52235 is expressed as an attenuated
262: bremsstrahlung model and require the addition of a broad iron line of
263: $7.00 \pm 0.02$ keV\@. The hydrogen column density decreases in the
264: decay phase from $60^{+50}_{-32}\times 10^{22}$ cm$^{-2}$ to
265: $18^{+20}_{-13}\times 10^{22}$ cm$^{-2}$ in both models used to fit the
266: data. The best fit thermal bremsstrahlung temperature and power-law
267: photon index at the peak are $kT_1 = 14^{+86}_{-9}$ keV and
268: $1.6^{+0.3}_{-0.3}$, respectively. The index or the temperature do not
269: change significantly during the decay. The unabsorbed 2-10 keV
270: luminosity at the peak are fitted to be $L_1 = 4.7^{+1.5}_{-1.9} \times
271: 10^{35}$ erg s$^{-1}$ assuming a distance of $D$ = 4.85 kpc
272: \citep{vermeulen93}.
273:
274: \paragraph{Interpretation}
275: This is the first detection of a rapid X-ray variability with a time
276: scale less than 300 s from SS~433 \citep{kotani02, kotani03,
277: safi-harb03}. Although this source had been observed for numerous times
278: with various X-ray observatories, only variability with time scales as
279: long as or longer than a day had been reported. For example, Einstein
280: observed the source to vary by a factor of 2 on time scales of a day
281: \citep{band89}, and daily variations other than the orbital and
282: precessional modulations are seen in Ginga and ASCA data \citep{yuan95,
283: kotani97}. Temporal analysis of ROSAT data shows flickering around 3-10
284: s, but this variability does not appear consistently
285: \citep{safi-harb97}. Since the detection with RXTE in 2001
286: \citep{kotani02, kotani03, safi-harb03}, evidences of rapid X-ray
287: variability have been accumulated. \citet{chakrabarti03} report on
288: X-ray variability with time scales of a few minutes detected in 2002
289: with RXTE, and \citet{revnivtsev04} detected a significant X-ray
290: variability with time scales as short as 100 s with RXTE in 2004.
291:
292: The absence or weakness of a rapid X-ray variability had been explained
293: in terms of the X-ray-emitting jet as long as or longer than $10^{12}$
294: cm. Together with the QPO-like feature in the power density spectrum,
295: this shot-like variability implies the presence of X-ray sources smaller
296: than $10^{12}$ cm in the system. Considering that these shots coincide a
297: massive jet ejection event, we further assume that they are related to
298: the ejection in the following discussion.
299:
300: Since the spectral fit is consistent with a decrease of the absorption
301: hydrogen column density during the evolution of the shots, we attribute
302: the rise of the shots to the decrease of attenuating matter, or in other
303: words, the emergence of an X-ray-emitting small plasma bullet from
304: behind attenuating matter. Each shot literally corresponds to a shot of
305: a small plasma bullet from the nozzle. This interpretation is different
306: from that of the X-ray variability seen in GRS~1915+105, which is
307: explained in terms of the rapid removal and replenishment of matter
308: forming the inner part of an accretion disk \citep{belloni97}. Since
309: both thermal and non-thermal spectral models are consistent with the
310: observed spectrum, it is difficult to determine the emission mechanism.
311: But in either case, physical quantities of the emitting bullets would be
312: derived as follows.
313:
314:
315: \paragraph{Thin-thermal emission}
316: Given a spherical, thin-thermal, freely expanding plasma bullet with a
317: temperature $T(t)$, a radius $R(t) = v_{\rm exp}t$, an expanding
318: velocity $v_{\rm exp} =$ const., and a total number of electrons $N_{\rm
319: e} =$ const., the cooling would be governed with the equation
320: \begin{eqnarray}
321: \frac 3 2 (1+X) N_{\rm e} k_{\rm B} \frac {dT} {dt}
322: &= &- \frac {\Lambda(T) X N_{\rm e}^2} {\frac 4 3 \pi R^3}
323: - 3 (\gamma -1) \frac 3 2 (1+X) N_{\rm e} k_{\rm B} \frac {T v_{\rm exp}} {R},
324: \label{eq:cooling}
325: \end{eqnarray}
326: where $X = N_{\rm i}/N_{\rm e}$ is the ratio of the total number of ions
327: to that of electrons assumed to be 0.93117, $k_{\rm B}$ is the Boltzmann
328: constant, and $\gamma$ is the adiabatic index, assumed to be 5/3. The
329: first term on the right-hand side corresponds to radiative cooling, and
330: the coefficient $\Lambda(T)$ is defined so that $\Lambda(T) N_{\rm e}/(4
331: \pi R^3/3) N_{\rm i}/(4 \pi R^3/3)$ equals to the emitted power per unit
332: volume. The last term corresponds to expansion: For expanding plasma
333: with a volume $V(t)$,
334: \begin{equation}
335: V^{\gamma-1}dT = -T(\gamma-1)V^{\gamma-2}dV \label{eq:adiabatic}
336: \end{equation}
337: is derived from the relation $TV^{\gamma-1} =$ constant. Substituting $V
338: = 4 \pi R^3/3$ to Eq.~(\ref{eq:adiabatic}), the cooling rate by expansion,
339: $ dT/dt = -3 (\gamma-1) T v_{\rm exp}/R$
340: is obtained, which is equivalent to the last term of
341: Eq.~(\ref{eq:cooling}).
342: The time parameter $t$ is defined so that $t = 0$ at $R =
343: 0$, although the radius can never be zero. If the coefficient
344: $\Lambda(T)$ is proportional to $\sqrt T$ and written in the form
345: $\Lambda(T) = \Lambda_T \sqrt T$, Eq.~(\ref{eq:cooling}) has an analytic
346: solution
347: \begin{eqnarray}
348: \sqrt T & = &
349: \left (\sqrt {T_1} - \frac {\Lambda_T N_{\rm e}X }{4 \pi (\frac 7 2 - \frac 3 2 \gamma) (1+X) k_{\rm B} v_{\rm exp}R_1^2} \right )
350: \left ( \frac R {R_1} \right )^{- \frac 3 2 \gamma + \frac 3 2} \nonumber \\
351: &&\mbox{} + \frac {\Lambda_T X N_{\rm e}}{4 \pi (\frac 7 2 - \frac 3 2 \gamma) (1+X) k_{\rm B} v_{\rm exp}R_1^2}
352: \left ( \frac R {R_1} \right )^{-2} \nonumber \\
353: & = & \sqrt {T_1} \left [\left ( 1 - \frac {L_1 \tau_{\rm exp}}{2 E_1} \right ) \left ( \frac R {R_1} \right )^{-1}
354: + \frac {L_1 \tau_{\rm exp}}{2 E_1}
355: \left ( \frac R {R_1} \right )^{-2} \right ],\label{eq:sol}
356: \end{eqnarray}
357: \begin{eqnarray}
358: E_1 &= &\frac 3 2 (1+X) N_{\rm e} k_{\rm B} T_1\\
359: L_1 &= &\frac {3\Lambda_T \sqrt {T_1} X N_{\rm e}^2}{4\pi R_1^3}\\
360: \tau_{\rm exp} &= &R_1/v_{\rm exp}
361: \end{eqnarray}
362: where the subscript ``$_1$'' denotes the value at the peak. $E_1$, $L_1$,
363: and $\tau_{\rm exp}$ correspond to the thermal energy,
364: the
365: luminosity of the bullet at the peak, and the time scale of expansion, respectively. The ratio ${L_1 \tau_{\rm exp}}/ (2E_1)$
366: represents the fraction of the thermal energy in the
367: bullet lost by radiation (cf.\ Kotani et al.\ 1996).
368: Using Eq.~(\ref{eq:sol}), the luminosity of
369: the bullet can be written as
370: \begin{eqnarray}
371: L & = &
372: L_1 \left [\left ( 1 - \frac {L_1 \tau_{\rm exp}}{2 E_1} \right )
373: \left ( \frac R {R_1} \right )^{-4}
374: + \frac {L_1 \tau_{\rm exp}}{2 E_1}
375: \left ( \frac R {R_1} \right )^{-5} \right ].\label{eq:lum}
376: \end{eqnarray}
377:
378: The expansion velocity $v_{\rm exp}$ is estimated from the observed
379: temperature as
380: $v_{\rm exp} = \sqrt{ k_{\rm B}T_1/(\mu m_{\rm H})}
381: = 1.5^{+2.5}_{-0.6}\times10^{8}$ cm s$^{-1}$,
382: where $m_{\rm H}$ is the mass of
383: a hydrogen atom and $\mu$ is the mean molecular weight assumed to be 0.587922.
384: In 40 s, the luminosity of the plasma decreases by a factor of $1/e$. Substituting the factor into Eq.~(\ref{eq:lum}), we obtain
385: \begin{eqnarray}
386: \frac {L_2}{L_1} = \frac 1 e & = &
387: \left ( 1 - \frac {L_1 \tau_{\rm exp}}{2 E_1} \right )
388: \left ( \frac {R_2} {R_1} \right )^{-4}
389: + \frac {L_1 \tau_{\rm exp}}{2 E_1}
390: \left ( \frac {R_2} {R_1} \right )^{-5}, \label{eq:l2l1}
391: \end{eqnarray}
392: where the subscript ``$_2$'' denotes the value at the decay phase.
393: Because the ratio of the radiation loss to the internal energy ${L_1
394: \tau_{\rm exp}}/ (2E_1)$ is between 0 and 1, the expansion in 40 s is
395: constrained as $\exp[1/5] < R_2/R_1 < \exp[1/4]$ from the above
396: equation. Substituting $R_2 = R_1 + v_{\rm exp} \times 40$ s, we obtain
397: the radius as $R_1 = 2.2^{+3.6}_{-0.9} \times 10^{10}$ cm. From this
398: radius and an observed quantity $L_1$, all other parameters are derived;
399: $1.06 \times 10^2 < \tau_{\rm exp} < 1.35 \times 10^2$ s, the number of
400: electrons $N_{\rm e} = 1.0^{+2.1}_{-0.5} \times 10^{45}$, the number
401: density of electrons $n_{\rm e1} = 4.1^{+9.2}_{-3.7} \times 10^{13}$
402: cm$^{-3}$, and the thermal energy $E_1 = 0.7^{+13.7}_{-0.6}\times
403: 10^{38}$ erg. Assuming that the bullet is moving at 0.26 c, the kinetic
404: energy of the bullet is estimated to be $0.6^{+1.2}_{-0.3}\times
405: 10^{41}$ erg. The assumption is consistent with the observed rise time
406: of 20 s, which is naturally explained by the time in which a bullet
407: appears from a nozzle.
408: Eq.~(\ref{eq:sol}) is written with these estimates as
409: \begin{eqnarray}
410: \sqrt T
411: & = & \sqrt {14 \; \mbox{[keV]}} \left [0.3 \left ( \frac R {1.6\times 10^{10} \; \mbox{[cm]}} \right )^{-1}
412: + 0.7 \left ( \frac R {1.6\times 10^{10} \; \mbox{[cm]}} \right )^{-2} \right ],\label{eq:solnum}
413: \end{eqnarray}
414: where errors are omitted.
415:
416: Since a shot and the unmodulated component coexist, the bullets and the
417: continuously emanating jet may coexist. In that case, the small plasma
418: bullets can be interpreted as bright knots in the continuous jet. The
419: knots are created when the mass outflow rate or the density of the
420: continuous jet is temporarily increased. A temporal increase of
421: temperature is not plausible, because it would result in a change of the
422: spectrum, which has not been observed. The properties of the knots
423: would not be much different from those of the small plasma bullets
424: discussed above, and the estimates above are valid if the knots or
425: bullets coexist with the continuous jet.
426:
427:
428: \paragraph{Synchrotron emission}
429: Because the spectral shape does not much change in the decay and because
430: no iron line is detected in the shot component, a non-thermal emission
431: from expanding bullets also can account for the shot component. As for
432: the steady non-variable component, it is definitely a thin-thermal
433: emission with Doppler-shifted iron lines. Therefore, it is natural to
434: interpret the shot component as a thermal emission, and an
435: interpretation of non-thermal emission is rather eccentric. In the
436: following paragraph, we show physical parameters of a plasma bullet assuming that
437: the bullets emit X ray via synchrotron radiation.
438:
439: A power-law distribution of synchrotron electrons,
440: \begin{eqnarray}
441: f(\gamma)d\gamma & \equiv &
442: \left\{
443: \begin{array}{ll}
444: \frac{-p+1}{\gamma_{\rm max}^{-p+1}-1} n_{\rm e,syn} \gamma^{-p} d\gamma
445: \approx (p-1) n_{\rm e} \gamma^{-p} d\gamma
446: & (1 < \gamma \le \gamma_{\rm max})\\
447: 0 & (\gamma_{\rm max} < \gamma)
448: \end{array}
449: \right.
450: \end{eqnarray}
451: is assumed, where $\gamma$ is Lorentz factor of electrons, $n_{\rm
452: e,syn}$ is the synchrotron-electron number density, and $p =
453: 2.2^{+1.4}_{-1.0}$ is the electron energy index derived from the photon
454: index $\Gamma = (p+1)/2$. The maximum Lorentz factor should be at least
455: $\gamma_{\rm max} > 1.4 \times10^{5}$ to account for the X-ray emission
456: up to 10 keV\@. Optically thin synchrotron flux from such a sphere is
457: expressed as
458: \begin{eqnarray}
459: F_\nu &= & \frac{\chi(p)}{4\pi}
460: n_{\rm e,syn} \frac{e^3}{mc^2} B^{p/2+1/2}
461: \left ( \frac{4\pi mc\nu} {3e} \right )^{-p/2+1/2} \frac{4\pi R^3/3}{D^2}\label{eq:intenapprothin}\\
462: \chi(p) & \equiv &\frac{3^{1/2}2^{p/2-1/2}(p-1)}{(p+1)}
463: \frac {\Gamma \left(\frac p 4 + \frac {19}{12}\right)
464: \Gamma \left(\frac p 4 - \frac {1}{12}\right)
465: \Gamma \left(\frac p 4 + \frac 5 4 \right)}
466: {\Gamma \left(\frac p 4 + \frac 7 4 \right)}
467: \end{eqnarray}
468: where $\nu$ is frequency, and $B$ is the magnetic field strength in the
469: plasma, (e.g., \cite{hjellming95}). By substituting the observed flux
470: $F\nu$(1 keV) = $1.0^{+1.0}_{-0.5}\times 10^{-2}$ photons cm$^{-2}$
471: s$^{-1}$ keV$^{-1}$ into Eq.~(\ref{eq:intenapprothin}), the magnetic
472: field strength $B$ and the total number of electrons in a bullet $N_{\rm
473: e}$are constrained as
474: \begin{eqnarray}
475: \log_{10}\left[N_{\rm e}B^{p/2+1/2}\;{\rm [G]}\right]
476: % & = & \log_{10}\left[\frac{1}{\chi(p)} \frac{3c}{e^2}
477: % \left ( \frac {4\pi mc}{3e} \right )^{p/2+1/2}
478: % F_\nu \nu^{p/2-2} D^2
479: % [\rm G]\right] \nonumber\\
480: & = & 44.6^{+3.0}_{-2.9},
481: \end{eqnarray}
482: where the rather large uncertainty is due to the uncertainty of the
483: electron energy index $p$.
484:
485: As
486: the bullet expands, each high-energy electron loses energy as $E =
487: (R/R_1)^{-1}$ and the magnetic field and the luminosity decreases as $B
488: = B_1 \times (R/R_1)^{-2}$ and $L = L_1 (R/R_1)^{-2p}$, respectively.
489: Radiative cooling and heating are neglected. Thus a decrease of flux by
490: a factor of $1/e$ corresponds to an adiabatic expansion by
491: $1.26^{+0.25}_{-0.10}$ of the radiating bullet. Further assuming that
492: the bullet is proceeding at 0.26 $c$ with an expanding half angle of
493: $2.1^\circ$, which are the same value as the velocity and the half
494: opening angle of the continuous jet \citep{namiki03}, the expansion
495: velocity is estimated as $v_{\rm exp} = 2.9 \times 10^{8}$ cm s$^{-1}$.
496: From this expansion velocity and the expansion factor obtained above,
497: the radius of the plasma is determined as $R = 4.5^{+1.4}_{-1.4} \times
498: 10^{10}$ cm, which is consistent with the rise time of 20 s.
499:
500: The synchrotron-electron number density and the strength of the magnetic
501: field can be estimated if their energies are assumed to be in
502: equipartition, i.e.,
503: \begin{eqnarray}
504: \frac {B^2}{8\pi} & = & \int_1^{\gamma_{\rm max}} \frac{p-1}{\gamma_{\rm max}^{-p+1}-1}
505: n_{\rm e} mc^2 \gamma^{-p+1} d\gamma \\
506: &&\approx \left\{
507: \begin{array}{ll}
508: \frac{p-1}{-p+2}\gamma_{\rm max}^{-p+2} n_{\rm e}mc^2 & \;\;\;\; (1 < p < 2)\\
509: \frac{p-1}{p-2} n_{\rm e}mc^2 & \;\;\;\; (2 < p).
510: \end{array}
511: \right. \label{eq:enee}
512: \end{eqnarray}
513: Substituting an electron energy index of $p=2.2$, we obtain $B = 1.8
514: \times 10^{2}$ G and $n_{\rm e} = 2.4 \times 10^{8}$ cm$^{-3}$. The
515: total number of electrons and internal energy in a bullet are estimated
516: as $N_{\rm e} = 9 \times 10^{40}$ per shot and $E_{\rm syn} = 5.3 \times
517: 10^{35}$ erg, respectively. These number of electrons and internal
518: energy derived here are smaller by orders of magnitude than those of the
519: thin-thermal case.
520:
521:
522: \paragraph{Comptonized emission}
523: If the emission mechanism is inverse-Compton scattering of optical
524: photons, the seed-photon density would decrease and the emission would
525: decay as the plasma bullet gets away from the central engine. Based on
526: this Comptonized-emission model, the parameters of the plasma bullet,
527: such as the electron number density, the total number of electrons, and
528: the total energy of electrons, are estimated. They are found to be not
529: much different from those in the case of synchrotron emission, although
530: the uncertainties of parameters are larger in the case of
531: inverse-Comptonization.
532:
533: \subsection{The massive jet}
534: Based on the bullets model, we suggest an explanation of the massive jet
535: ejection: During the massive jet ejection event, small discrete plasma
536: bullets, or knots in the continuously emanating flow, are successively
537: ejected at random intervals of $\sim 150$ s. The radius of bullets is
538: estimated as $R_1 = 2.2^{+3.6}_{-0.9} \times 10^{10}$ cm. The X-ray
539: emission from the small plasma bullets, either thermal or non-thermal,
540: decays in 40 s as it expands. Assuming that the most active state lasts
541: 3 days, the total number of small bullets ejected in a single massive
542: jet event is estimated to be 1700. The total mass and total kinetic
543: energy of all the 1700 bullets are $3.3^{+6.7}_{-1.8} \times 10^{24}$ g
544: and $1.0^{+2.1}_{-0.5} \times 10^{44}$ erg, respectively. The average
545: mass ejection rate and average kinetic luminosity over 3 days are
546: $0.7_{-0.4}^{+1.6} \times 10^{16}$ g s$^{-1}$ and $3.9^{+8.0}_{-2.1}
547: \times 10^{38}$ erg s$^{-1}$, respectively. In the case of synchrotron
548: emission from baryonic plasma, only the lower limits of mass and kinetic
549: energy are derived; the total mass and total kinetic energy of the 1700
550: bullets would be at least $3.0 \times 10^{20}$ g and $9.1 \times
551: 10^{39}$ erg, respectively, and the average mass ejection rate and
552: average kinetic luminosity would be at least $1.1 \times 10^{15}$ g and
553: $3.5 \times 10^{34}$ erg s$^{-1}$, respectively.
554:
555:
556:
557: The estimated average kinetic luminosity of $\sim 10^{38}$ erg s$^{-1}$
558: is considerably lower than estimates based on the quiescent or normal
559: state. For example, \citet{kotani97} has calculated the kinetic
560: luminosity as $1 \times 10^{40}$ erg s$^{-1}$ based on ASCA data,
561: \citet{marshall02} as $3.2 \times 10^{38}$ erg s$^{-1}$ based on
562: HETGS/Chandra data, and \citet{brinkmann05} as $5 \times 10^{39}$ erg
563: s$^{-1}$ based on EPIC/XMM-Newton data. It is puzzling that the mass
564: outflow rate and kinetic luminosity in the massive jet ejection are not
565: so ``massive'' compared to those of the steady continuous jet flow seen
566: in most occasions. There are several possibilities to account for the
567: inconsistency in terms of the bullets model: 1) The mass outflow rate
568: and kinetic luminosity of a massive jet are not larger than those of
569: quiescent steady jet, but the efficiency to accelerate electrons
570: contributing to synchrotron radio emission is far larger. 2) The massive
571: jet is not an assembly of the small plasma bullets, but mainly supplied
572: with the steady flow which coexists with the bullets. 3) In spite of
573: the monitoring observation with a sampling rate of 3 ks a day, we have
574: missed the moment of the true massive jet ejection, which lasts only,
575: say, hours, and a massive jet of $10^{44} - 10^{45}$ erg is ejected at a
576: maximum outflow rate of $10^{40} - 10^{41}$ erg s$^{-1}$. We do not yet
577: have an evidence supporting one of them, but suggest that the second
578: case is unlikely, in which the unvariable X-ray component is expected to
579: rise as the radio flux densities rise.
580:
581: Another question is whether X-ray variability is really related to radio
582: flaring, which is associated with blob rebrightening events out of the
583: system core. At 1.6 GHz, radio flares peak at 35 AU from the core
584: \citep{paragi99}. If radio flares were caused by an environmental
585: condition, it would not be detectable in X-ray band at the ejection, and
586: the coincidence of the rapid X-ray variability and the massive jet event
587: would be accidental. However, \citet{paragi99} suggest that the
588: rebrightening is due to the attenuation by out-flowing gas around the
589: core. If the cause of radio flares is not an environmental condition
590: but a core activity, the activity which changes the radio flux by a
591: factor of 2 might be detectable in X-ray band. That should be tested in
592: future multi-wavelength observations.
593:
594:
595: X-ray shots are firstly seen in the data from MJD = 52225, a week before
596: its maximum activity and the onset of a second radio flare. Provided
597: that shots precede a radio flare, we can predict a massive
598: jet-ejection event based on X-ray monitoring data. On detection of
599: X-ray shots, notice of a massive jet ejection to occur in a week can be
600: sent to ground and space observatories. An observation campaign
601: covering a massive jet ejection is possible. There is still a
602: possibility that the duration of a massive jet ejection is shorter
603: than~1 day and the moment has been missed even in our observations. It
604: can be confirmed in future observations. And this technique may be
605: applicable for prediction of massive jet ejections from other
606: microquasars. In spite of numerous observations performed so far, it is
607: not yet known whether massive jet ejections from other microquasars such
608: as GRS~1915+105 are also preceded by a precursor or not. As a specially
609: coordinated observation is required to detect the shot-like variability
610: from SS~433, a carefully coordinated observations plan is desirable to
611: observe GRS~1915+105 in a massive jet ejection event with an X-ray
612: mission. The findings reported here imply that new and important
613: physics of a microquasar is revealed by observing massive jet ejections.
614: The observation of these events is essential to explore the nature of
615: microquasars. Therefore the technique to observe massive jet ejections
616: is one of the most important results from this study. Future
617: observations of massive jet ejections from microquasars are encouraged.
618:
619: %\section{Summary}\label{sec:summary}
620:
621:
622: \acknowledgments
623: T.~K. is supported by a 21st Century COE Program at Tokyo Tech
624: ``Nanometer-Scale Quantum Physics'' by the Ministry of Education,
625: Culture, Sports, Science and Technology. S.~A.~T. is very grateful to
626: the Russian Base Researches Foundation for support of the SS~433
627: monitoring with the RATAN-600 radio telescope. K.~K. is supported by a
628: grant-in aid 16740121 from the Ministry of Education, Culture, Sports,
629: Science and Technology of Japan. We thank the anonymous referee for
630: useful comments and suggestions to improve this paper.
631:
632:
633:
634: \begin{thebibliography}{99}
635: \bibitem[Band(1989)] {band89} Band, D. 1989, \apj, 336, 937
636:
637: \bibitem[Belloni et al.(1997)] {belloni97} Belloni, T., Mendez, M., King, A. R., van der
638: Klis, M., \& van Paradijs, J. 1997,
639: % An Unstable Central Disk in the Superluminal Black Hole X-Ray Binary GRS 1915+105.
640: \apjl, 479, L145
641:
642: \bibitem[Brinkmann et al.(2005)]{brinkmann05} Brinkmann, W.,
643: Kotani, T., \& Kawai, N.\ 2005, \aap, 431, 575
644:
645: \bibitem[Chakrabarti et al.(2003)]{chakrabarti03} Chakrabarti, S.~K.,
646: Pal, S., Nandi, A., Anandarao, B.~G., \& Mondal, S.\ 2003, \apjl, 595, L45
647:
648:
649: \bibitem[Fender(2001)] {fender01} Fender, R. P. 2001,
650: in Proc.\ ESO Workshop: Black holes in binaries and galactic
651: nuclei, ed.\ L. Kaper, E. P. J. van den Heuvel, \& Woudt, P. A.
652: (Berlin: Spring-Verlag), 193
653:
654: \bibitem[Fiedler et al.(1987)] {fiedler87} Fiedler, R. L., et al.\ 1987,
655: %Observations of SS 433 at 2695 and 8085 MHz, 1979-1985.
656: \aj, 94, 1244
657:
658: \bibitem[Foster et al.(1996)] {foster96} Foster, R. S., et al.\ 1996,
659: % Radio and X-ray variability of the galactic superluminal source GRS 1915+105.
660: \apj, 467, L81
661:
662: \bibitem[Gies et al.(2002)]{gies02} Gies, D.~R., McSwain,
663: M.~V., Riddle, R.~L., Wang, Z., Wiita, P.~J., \& Wingert, D.~W.\ 2002,
664: \apj, 566, 1069
665:
666:
667: \bibitem[Goranskii et al.(1998)] {goranskii98} Goranskii, V. P., Esipov,
668: V. F., \& Cherepashchuk, A. M. 1998,
669: % Optical variability of SS 433 in 1978-1996.
670: {Astron.\ Reports}, 42, 209
671:
672: \bibitem[Hasegawa et al.(2004)]{hasegawa04} Hasegawa, T., Malasan,
673: H.~L., Kawakita, H., Obayashi, H., Kurabayashi, T., Nakai, T., Hyakkai, M.,
674: \& Arimoto, N.\ 2004, \pasj, 56, 295
675:
676: \bibitem[Hjellming \& Han(1995)]{hjellming95} Hjellming, R.~M., \& Han,
677: X. 1995, in X-Ray Binaries, ed.\ W.~H.~G. Lewin, J.~van Paradijs, \&
678: E.~P.~J. van den Heuvel (Cambridge: Cambridge University Press), 308
679:
680:
681: \bibitem[Kawai(1989)]{kawai89} Kawai, N. 1989, in Proc.\ 23rd ESLAB
682: Symp.\ on Two-topic in X-Ray Astronomy, (European Space Agency,
683: Noordwijk), 453
684:
685: \bibitem[Kinugasa et al.(2002)]{kinugasa02} Kinugasa, K., et al.\
686: 2002, \apjl, 577, L97
687:
688:
689: \bibitem[Klein-Wolt et al.(2002)]{klein-wolt02} Klein-Wolt, M.,
690: Fender, R.~P., Pooley, G.~G., Belloni, T., Migliari, S., Morgan, E.~H., \&
691: van der Klis, M.\ 2002, \mnras, 331, 745
692:
693: \bibitem[Korolkov \& Pariiskii(1979)] {korolkov79} Korolkov, D.~V., \& Pariiskii, I.~N. 1979,
694: % The Soviet RATAN-600 radio telescope.
695: \skytel, 57, 324
696:
697: \bibitem[Kotani et al.(1996)]{kotani96} Kotani, T., Kawai, N., Matsuoka, M.,
698: \& Brinkmann, W. 1996, \pasj, 48, 619
699:
700: \bibitem[Kotani(1997)] {kotani97} Kotani, T. 1997,
701: % {\it X-RAY OBSERVATIONS OF A GALACTIC JET SYSTEM SS 433 WITH ASCA\@.}
702: Doctoral Thesis, University of Tokyo
703:
704: \bibitem[Kotani \& Trushkin(2001)] {kotani01} Kotani, T., \& Trushkin, S. 2001,
705: % SS 433.
706: \iaucirc, 7747
707:
708: \bibitem[Kotani et al.(2002)]{kotani02} Kotani, T., Band, D.,
709: Denissyuk, E.~K., Kawai, N., Kinugasa, K., Namiki, M., Safi-Harb, S., \&
710: Trushkin, S.\ 2002, ASP Conf.~Ser.~279: Exotic Stars as Challenges to
711: Evolution, 279, 19
712:
713:
714: \bibitem[Kotani et al.(2003)] {kotani03} Kotani, T., Trushkin, S., Denissyuk, E.~K.,
715: Kawakita, N., Kinugasa, K., Safi-Harb, S., \& Band, D. 2003,
716: in New Views on Microquasars, ed.\ P. Durouchoux, Y. Fuchs, \&
717: J. Rodriguez (Kolkata: Centre for Space Physics), 265
718:
719: \bibitem[Marshall et al.(2002)]{marshall02} Marshall, H.~L.,
720: Canizares, C.~R., \& Schulz, N.~S.\ 2002, \apj, 564, 941
721:
722:
723: \bibitem[Mirabel et al.(1998)]{mirabel98} Mirabel, I.~F., Dhawan,
724: V., Chaty, S., Rodriguez, L.~F., Marti, J., Robinson, C.~R., Swank, J., \&
725: Geballe, T.\ 1998, \aap, 330, L9
726:
727: \bibitem[Mirabel \& Rodr\'{\i}guez(1994)] {mirabel94} Mirabel, I. F., \& Rodr\'{\i}guez, L. F. 1994,
728: % A Superluminal Source in the Galaxy.
729: \nat, 371, 46
730:
731: \bibitem[Mirabel \& Rodr\'{\i}guez(1999)] {mirabel99} Mirabel, I. F., \& Rodr\'{\i}guez, L. F. 1999,
732: % Sources of relativistic jets in the Galaxy.
733: \araa, 37, 409
734:
735: \bibitem[Mironov \& Tereshchenko(1998)] {mironov98} Mironov, A.~V., \& Tereshchenko, V.~M. 1998,
736: % Astronomical Observatories in Kazakhstan.
737: Baltic Astron., 7, 351
738:
739: \bibitem[Muno et al.(2001)]{muno01} Muno, M.~P., Remillard,
740: R.~A., Morgan, E.~H., Waltman, E.~B., Dhawan, V., Hjellming, R.~M., \&
741: Pooley, G.\ 2001, \apj, 556, 515
742:
743: \bibitem[Namiki et al.(2001)]{namiki01} Namiki, M., Kawai, N.,
744: \& Kotani, T.\ 2001, ASP Conf.~Ser.~251: New Century of X-ray Astronomy,
745: 251, 390
746:
747:
748: \bibitem[Namiki et al.(2003)] {namiki03} Namiki, M., Kawai, N., Kotani, T., \& Makishima, K. 2003,
749: % Evidence for Jet Collimation in SS 433 with the Chandra HETGS.
750: \pasj, 55, 281
751:
752: \bibitem[Paragi et al.(1999)]{paragi99} Paragi, Z., Vermeulen,
753: R.~C., Fejes, I., Schilizzi, R.~T., Spencer, R.~E., \& Stirling, A.~M.\
754: 1999, \aap, 348, 910
755:
756:
757:
758: \bibitem[Revnivtsev et al.(2004)]{revnivtsev04} Revnivtsev, M., et
759: al.\ 2004, \aap, 424, L5
760:
761:
762: \bibitem[Rupen et al.(1998)] {rupen98} Rupen, M. P., Hjellming, R. M., \& Mioduszewski, A. J. 1998,
763: % XTE J1748$-$288.
764: \iaucirc, 6938
765:
766: \bibitem[Safi-Harb(1997)] {safi-harb97} Safi-Harb, S. 1997,
767: Ph.D. Thesis, University of Wisconsin.
768:
769: \bibitem[Safi-Harb \& Kotani(2003)] {safi-harb03} Safi-Harb, S., \& Kotani, T. 2003,
770: in New Views on Microquasars, ed.\ P. Durouchoux, Y. Fuchs, \&
771: J. Rodriguez (Kolkata: Centre for Space Physics), 279
772:
773: \bibitem[Trushkin et al.(2003)] {trushkin03} Trushkin, S. A., Bursov, N. N., \& Nizhelskij, N. A. 2003,
774: % The multifrequency monitoring of microquasars. SS 433.
775: Bull.\ Spec.\ Astrophys.\ Obs., 56, 57
776:
777: \bibitem[Ueda et al.(2002)] {ueda02} Ueda, Y., et al.\ 2002,
778: % Study of the Largest Multiwavelength Campaign of the Microquasar GRS 1915+105.
779: \apj, 571, 918
780:
781: %\bibitem[Vaughan(2005)]{vaughan05} Vaughan, S. 2005, \aa, in press,
782: %(Astro-ph/0412697)
783:
784:
785: \bibitem[Vermeulen et al.(1993)] {vermeulen93} Vermeulen, R. C., Schilizzi, R. T., Spencer, R. E.,
786: Romney, J. D., \& Fejes, I. 1993,
787: % A series of VLBI images of SS 433 during the outbursts in
788: % May/June 1987.
789: \aap, 270, 177
790:
791: \bibitem[Watson et al.(1983)]{watson83} Watson, M. G., Willingale, R.,
792: Grindlay, J. E., \& Seward, F. D. 1983,
793: \apj, 273, 688
794:
795: \bibitem[Yuan et al.(1995)]{yuan95} Yuan, W., Kawai, N., Brinkmann, W.,
796: \& Matsuoka, M. 1995, \aap, 297, 451
797:
798:
799: \end{thebibliography}
800:
801: \clearpage
802:
803:
804: \begin{table}
805: \centering
806: \caption{Observation log}\label{tbl:log}
807: \begin{tabular}{cccl}
808: \hline
809: Start &End &Exposure Time &PCU$^{\rm a}$\\
810: (MJD) & (MJD) &(ks) &\\
811: \hline
812: 2001/11/09 07:10 (52222.299) &2001/11/09 08:11 (52222.341) &3.6&0234 \\
813: 2001/11/10 05:19 (52223.222) &2001/11/10 06:21 (52223.265) &3.7&0234 \\
814: 2001/11/11 06:47 (52224.283) &2001/11/11 07:47 (52224.325) &3.6&0234 \\
815: 2001/11/12 06:35 (52225.275) &2001/11/12 07:35 (52225.316) &3.5&0234 \\
816: 2001/11/13 04:46 (52226.199) &2001/11/13 05:46 (52226.240) &3.5&0234 \\
817: 2001/11/14 06:11 (52227.258) &2001/11/14 07:13 (52227.301) &3.7&0234 \\
818: 2001/11/15 06:00 (52228.250) &2001/11/15 07:01 (52228.293) &3.7&0234 \\
819: 2001/11/16 07:26 (52229.310) &2001/11/16 08:32 (52229.356) &4.0&023 \\
820: 2001/11/17 07:14 (52230.302) &2001/11/17 08:20 (52230.349) &4.1&023 \\
821: 2001/11/19 03:41 (52232.154) &2001/11/19 04:33 (52232.190) &3.1&0234 \\
822: 2001/11/20 06:40 (52233.278) &2001/11/20 07:46 (52233.324) &4.0&02 \\
823: 2001/11/21 08:05 (52234.337) &2001/11/21 09:16 (52234.387) &4.3&023 \\
824: 2001/11/22 07:54 (52235.330) &2001/11/22 09:05 (52235.379) &4.2&024 \\
825: 2001/11/23 04:32 (52236.189) &2001/11/23 05:29 (52236.229) &3.5&024 \\
826: 2001/11/24 01:11 (52237.050) &2001/11/24 01:54 (52237.080) &2.6&02 \\
827: 2001/11/25 05:45 (52238.240) &2001/11/25 06:47 (52238.283) &3.7&012 \\
828: \hline
829: \end{tabular}
830: $^{\rm a}$ Proportional counter units in operation.
831: \end{table}
832:
833:
834:
835: \clearpage
836:
837: \begin{table}
838: \caption{Spectral parameters}\label{tbl:best-fit}
839: \begin{tabular}{ccccccccc}
840: \hline
841: Start &$kT$ &$E_{\rm N}$&$F_{\rm N}$ &$E_{\rm B}$&$\sigma_{\rm B}$&$F_{\rm B}$ &$\chi^2/$d.o.f.\\
842: (MJD) &(keV)&(keV) &(ph s$^{-1}$ cm$^{-2}$)&(keV) &(keV) &(ph s$^{-1}$ cm$^{-2}$)&\\
843: \hline
844: 52222.299 &35 &6.5 &$5.1\times10^{-4}$ &6.92 &0.83 &$2.2\times10^{-3}$ &30.5/31\\
845: 52223.222 &49 &6.4 &$4.7\times10^{-4}$ &6.98 &0.81 &$2.2\times10^{-3}$ &51.5/31\\
846: 52224.283 &47 &6.5 &$4.7\times10^{-4}$ &6.98 &0.83 &$2.2\times10^{-3}$ &29.7/31\\
847: 52225.275 &56 &6.7 &$5.2\times10^{-4}$ &6.95 &1.00 &$2.9\times10^{-3}$ &57.0/31\\
848: 52226.199 &55 &6.8 &$6.5\times10^{-4}$ &6.90 &1.07 &$3.1\times10^{-3}$ &82.8/31\\
849: 52227.258 &45 &6.9 &$5.5\times10^{-4}$ &6.90 &1.10 &$3.1\times10^{-3}$ &45.0/31\\
850: 52228.250 &37 &6.8 &$4.8\times10^{-4}$ &6.92 &0.99 &$2.8\times10^{-3}$ &39.9/31\\
851: 52229.310 &48 &6.9 &$6.3\times10^{-4}$ &6.94 &1.07 &$2.9\times10^{-3}$ &38.7/31\\
852: 52230.302 &37 &6.9 &$3.4\times10^{-4}$ &6.97 &0.95 &$2.3\times10^{-3}$ &49.6/31\\
853: 52232.154 &38 &7.0 &$5.9\times10^{-4}$ &6.88 &1.21 &$3.3\times10^{-3}$ &72.2/31\\
854: 52233.278 &25 &7.0 &$2.2\times10^{-4}$ &7.06 &0.93 &$2.0\times10^{-3}$ &87.1/30\\
855: 52234.337 &21 &6.9 &$2.2\times10^{-4}$ &7.11 &1.07 &$2.1\times10^{-3}$ &42.1/30\\
856: 52235.330 &48 &7.0 &$6.0\times10^{-4}$ &6.88 &1.12 &$2.5\times10^{-3}$ &38.4/30\\
857: 52236.189 &53 &7.0 &$5.1\times10^{-4}$ &6.92 &1.05 &$2.7\times10^{-3}$ &53.1/30\\
858: 52237.050 &55 &7.0 &$6.6\times10^{-4}$ &6.74 &1.24 &$3.0\times10^{-3}$ &51.9/30\\
859: 52238.240 &42 &7.1 &$6.9\times10^{-4}$ &6.92 &1.27 &$3.4\times10^{-3}$ &42.9/30\\
860: \hline
861: \end{tabular}
862: The least significant digit of each best-fit value has an uncertainty of
863: the same order. For 2-10 keV flux, see Fig.~\ref{fig:lc_x_radio}.
864: \end{table}
865:
866:
867:
868:
869:
870:
871: \clearpage
872:
873: \begin{figure}
874: %\resizebox{1\textwidth}{!}{\includegraphics{kotani.fig1.lc_x_radio.eps}}
875: \plotone{f1.eps} \caption{2.3 GHz radio light curve (open squares with
876: error bars) and
877: 2-10 keV X-ray light curve (open circles with
878: error bars and solid line) of SS 433.
879: The radio flux taken with RATAN-600 shows two flares, or massive
880: jet-blob ejections, indicated by filled triangles. The epochs of
881: eclipse are indicated in dashed lines. The first flare has triggered
882: RXTE monitoring observations. The rise of the second flare is not
883: prominent because of a lack of the monitoring data on MJD = 52234 due to
884: bad weather condition. The X-ray fluxes show a peak just before the
885: second radio flare and then a dip coinciding with an eclipse.
886: }\label{fig:lc_x_radio}
887: \end{figure}
888:
889: \clearpage
890:
891: \begin{figure}
892: %\resizebox{1\textwidth}{!}{\includegraphics{kotani.fig3.pds.eps}}
893: \plotone{f2.eps}
894: \caption{Sum of power density spectra of the X-ray data. The sum of all 16
895: spectra and the best fit model of a power-law model plus a Gaussian (solid
896: line) are plotted. The data points and upper limits are marked as
897: filled circles and T-shaped bars, respectively. A feature which can be
898: interpreted as a QPO centered at $0.1127 \pm 0.0072$ Hz with a Gaussian
899: sigma of $0.011 \pm 0.006$ Hz is seen. }\label{fig:pds}
900: \end{figure}
901:
902: \begin{figure}
903: %\resizebox{1\textwidth}{!}{\includegraphics{kotani.fig2.lc_x.eps}}
904: \resizebox{!}{0.8\textheight}{\plotone{f3.eps}}
905: \caption{Blown-up X-ray light curves taken on MJD =
906: 52222 (upper panel), 52225 (middle), and 52232 (lower). Only the energy
907: band higher than iron and nickel lines (8.38 keV for blue-shifted Ni
908: {\sc xxvii} K$\alpha$) is shown. Typical 1 $\sigma$ errors are plotted
909: as crosses. The amplitude of variations on MJD = 52222, if any, is not
910: larger than the error bar. On MJD = 52225, the variations becomes
911: significant, and the amplitude reaches the maximum on 52232. The
912: variation on MJD = 52232 can be interpreted as a series of ``shots'' or
913: ``spikes.'' By picking up local maximums in the light curve above a
914: threshold of 74 counts s$^{-1}$ after at least 2 successive increasing
915: bins (30 s), twelve shots are sampled as indicated by the filled
916: triangles in the lower panel. The criterion of successive increasing
917: bins is necessary to cut local maximums due to fluctuation. The
918: threshold of 74 counts s$^{-1}$ (13 $\sigma$) is chosen so that most
919: shots are sampled only once. This sampling is not exhaustive and a
920: different set of shots may be selected under another criterion.
921: }\label{fig:lc_x}
922: \end{figure}
923:
924: \begin{figure}
925: %\resizebox{1\textwidth}{!}{\includegraphics{kotani.fig4.foldedshot.eps}}
926: \plotone{f4.eps}
927: \caption{Average
928: profile of the twelve shots sampled from the data on MJD = 52232
929: (Fig.~\ref{fig:lc_x}). In the 8.4-21 keV band, the
930: shot rises in $\sim 10$ s and decays slowly ($\sim 30$ s). In the
931: 2.4-5.7 keV
932: band, the profile is less pronounced: The peak lags behind that in the hard band by
933: $\sim 30$ s, and both the rise and decay time scales are $\sim 30$ s.
934: The peak is indistinct in the 5.7-8.4 keV band,
935: too. That implies that the iron-line intensity is not much
936: contributing the shot.
937: For a spectroscopic study, we have divided the profile into three phases;
938: the ``pre-shot,'' ``peak,'' and ``decay'' phases. }\label{fig:foldedshot}
939: \end{figure}
940:
941: \begin{figure}
942: %\resizebox{1\textwidth}{!}{\includegraphics{kotani.fig5.spec.eps}}
943: \plotone{f5.eps}
944: \caption{Spectra and best-fit models of the pre-shot phase
945: (crosses), the peak phase (open circles), and the decay phase (filled
946: circles). To emphasize the spectral evolution of the shot component,
947: the pre-shot spectrum is
948: subtracted from the peak and decay spectra.
949: The background spectrum made with a tool pcabackest is
950: subtracted from the pre-shot spectrum.}\label{fig:spec}
951: \end{figure}
952:
953: \clearpage
954:
955:
956:
957: \end{document}
958: