1: % LaTeX doc by Jeewan C. Pandey (jeewan@tifr.res.in)
2: % Start : 19 Dec 2006
3: % Last Mod : 26 May 2008 12:05
4: \documentclass[useAMS,referee]{mn2e}
5: \usepackage{amsmath,amssymb,graphicx,deluxetable}
6: \usepackage{subfigure}
7: \input{def}
8: \title[A study of X-ray flares - I. Active late-type dwarfs]{X-ray flares. Active late-type dwarfs}
9: \author[J. C. Pandey and K. P. Singh]{J. C. Pandey$^1$\thanks{E-mail:jeewan@aries.ernet.in (JCP); singh@tifr.res.in (KPS)}\thanks{Present address: Aryabhatta Research Institute of Observational Sciences, Nainital -263 129, India} and K. P. Singh$^{1\star}$\\
10: $^{1}$Tata Institute of Fundamental Research, Mumbai - 400 005, India}
11:
12: \begin{document}
13:
14: \date{} %TODO
15:
16: \pagerange{\pageref{firstpage}--\pageref{lastpage}} \pubyear{2007}
17:
18: \maketitle \label{firstpage}
19:
20: \begin{abstract}
21: We present temporal and spectral characteristics of X-ray flares observed
22: from six late-type G-K active dwarfs (V368 Cep, XI Boo, IM Vir, V471 Tau,
23: CC Eri and EP Eri) using data from observations with the XMM-Newton
24: observatory. All the stars were found to be flaring frequently and altogether
25: a total of seventeen flares were detected above the ``quiescent'' state X-ray
26: emission which varied from 0.5 to 8.3 $\times 10^{29}$ \lum. The largest flare
27: was observed in a low activity dwarf XI Boo with a decay time of 10 ks and ratio of peak flare luminosity to ``quiescent'' state luminosity of 2. We have studied the spectral changes during the flares by using colour-colour diagram and by detailed spectral analysis during the temporal evolution of the flares. The exponential decay of the X-ray light curves, and time evolution of the plasma temperature and emission measure are similar to those observed in compact solar flares. We have derived the semiloop lengths of flares based on the hydrodynamic flare model. The size of the flaring loops is found to be less than the stellar radius. The hydrodynamic flare decay analysis indicates the presence of sustained heating during the decay of most flares.
28: \end{abstract}
29:
30: \begin{keywords}
31: X-ray:stars -- stars:activity -- stars:coronae -- stars:flare -- stars:late-type
32: \end{keywords}
33: \section{Introduction}
34: \label{sec:intro}
35:
36: Flares are events in which a large amount of energy is released in a short
37: interval
38: of time. Such events take place at almost all frequencies of the electromagnetic
39: spectrum and have been observed in the Sun as well as in many types of cool stars
40: (Garcia-Alvarez, et al. 2002). These stellar flares could be radiating
41: several orders of magnitude more energy
42: than a solar flare. Optical flares in UV Cet (dMe) type stars are a
43: common phenomenon.
44: The flares produced by other stellar sources (e.g. RS CVn and BY Dra) are usually
45: detected only in the UV or X-rays. These UV and X-ray flares show extreme
46: luminosities and very hot temperatures ($\gtrsim 10 $MK). Even though, flares
47: in these stars present many analogies with the solar flares, there are also
48: significant differences, such as the amount of energy released. A general model
49: for stellar flares has emerged from numerous solar flare studies (G\"{u}del 2004).
50: The flare reconnection region, located somewhere at large coronal
51: heights, primarily accelerates electrons and ions upto MeV
52: energy (Dennis \& Schwartz 1989). The accelerated electrons participate
53: along the magnetic fields into chromosphere where they heat the cool plasma
54: to coronal flare temperatures, thus evaporating a part of the chromosphere
55: into the corona. Early X-ray observations with the EXOSAT observatory of flare stars
56: revealed examples
57: of two different type of flares: 1) Impulsive flares which are like
58: compact solar flares, and 2) Long decay flares which are like two-ribbon
59: solar flares (Pallavicini, Tagliaferri \& Stella 1990). The compact flares are less
60: energetic ($\sim$ 10$^{30}$ ergs s$^{-1}$), short in duration ($< 1$ h),
61: and confined to a single loop while the long-decay flares are more energetic
62: ($\sim$ 10$^{32}$
63: ergs s$^{-1}$), of long duration ($\ge 1$ h), and release the energy in an
64: entire arcade of loops (Garcia-Alvarez 2000). The most likely flare process
65: relates to an opening up of magnetic fields and subsequent relaxation by closing
66: the open field lines.
67:
68: Analysis of light curves during flares can provide us with insights into
69: the characteristics of the coronal structures and, therefore, of the magnetic
70: field (e.g., Schmitt \& Favata 1999; Favata, Micela \& Reale 2000a; Reale et al. 2004).
71: Even though stellar flares are spatially unresolved, a great deal of
72: information on the coronal heating and on the plasma structure morphology can
73: be inferred from a detailed modeling of stellar flares; for instance, if
74: sufficient data are available for moderately time-resolved spectral
75: analysis, a study of the complete evolution of a flare can allow us to (a) infer
76: whether the flare occurs in closed coronal structures (loops), (b) determine
77: the size of the flaring structures, (c) determine whether continuous heating is present throughout the
78: flare, and (d) put constraints on the location and distribution of the heating (see Reale et al. 2004).
79:
80: Active solar-type (G-K) dwarfs are known to possess magnetic fields on
81: their surfaces, and the field strengths are as large as several kilogauss, i.e.
82: stronger than the strongest field observed on the Sun (e.g., Saar 1990;
83: Johns-Krull \& Valenti 1996).
84: Cyclic behavior has been identified
85: in several late-type active stars (Baliunas et al. 1995, Olah \& Strassmeier 2002),
86: suggesting that dynamos similar to the solar dynamos are also operative in
87: solar-type stars.
88: These solar-type stars are
89: less active than the dMe stars (so called flare stars).
90: However, they show more frequent flaring activity than the
91: late-type evolved stars (e.g. RS CVns; G{\"u}del 2004).
92: In this paper, we present analysis of archival data obtained from
93: XMM-Newton observations of
94: six G-K dwarfs, namely V268 Cep, XI Boo, IM Vir, V471 Tau, EP Eri and CC Eri
95: with the aim to understand the spectral and temporal characteristics of X-ray flares in them.
96: In subsequent papers we will examine the characteristics of X-ray flares
97: in the subgiants and the giants.
98:
99: The paper is organized as follows: In \S \ref{sec:basic} we give the basic
100: parameters of stars in our sample, in \S \ref{sec:obs} we describe the
101: observational data sets that we have analysed and the methods of data reduction,
102: \S \ref{sec:analysis} contains our analysis and results, and in \S \ref{sec:discussion} we present our discussion and
103: conclusions.
104:
105: \section{Basic parameters of sample stars}
106: \label{sec:basic}
107: The basic parameters of stars in our sample are given in Table 1.
108: V368 Cep (= HD 220140; V = 7.54 mag) is a G9V spectral-type chromospherically active star with photometric period of
109: 2.74 d (Chugainov, Lovkaya \& Petrov 1991, Chugainov, Petrov \& Lovkaya 1993).
110: XI Boo (= HD 131156; V = 4.55 mag) is a nearby visual binary, comprising a primary G8 dwarf and a secondary K4 dwarf
111: with an orbital period of 151 yrs (Hoffleit \& Jaschek 1982). In terms of the outer
112: atmospheric emission, UV and X-ray observations show that the primary dominates
113: entirely over the secondary (Hartman et al. 1979, Ayres, Marstad \& Linsky 1981,
114: Schimtt 1997, Laming \& Drake 1999). XI Boo A (primary) is a slow rotator
115: with rotation period of 6.2 d (Gray et al. 1996).
116: A cool main sequence eclipsing binary (G5V+KV), IM Vir (= HD 111487; V = 9.69 mag ) has an orbital period of 1.31d (Malkov et al. 2006).
117: V471 Tau (= BD+16 516; V = 9.71) is a rapidly rotating detached eclipsing binary first
118: observed by Nelson \& Young (1970). The system is composed of a cool main-sequence
119: chromospherically active K2 V star (Guinan \& Sion 1984) and a degenerate
120: hot white dwarf separated by 3.4 stellar radii.
121: CC Eri (= HD 16157, V = 8.8) is a spectroscopic binary with orbital period of
122: 1.561 d. It consists of a K7Ve primary and a dM4 secondary, with mass ratio
123: of $\sim 2$ (Strassmeier et al. 1993). The primary star corotates with orbital motion
124: due to the tidal lock, and is one of the fastest rotating K dwarfs in the solar
125: neighborhood. EP Eri (=HD 17925, V = 6.0 mag) is very
126: nearby (10.3 pc), very young (high Li I abundances),
127: active K2-type dwarf with a rotation period of 6.85 days (Cayrel de Strobel
128: \& Cayrel 1989; Cutispoto 1992; Henry, Fekel \& Hall 1995). The presence of an
129: unresolved companion in this star has been
130: suggested by {\rm Henry et al. (1995)} based on the variable widths of the
131: photospheric absorption lines reported in the literature ($v\sin{i}$ ranges
132: from 3 to 8 km s$^{-1}$; see Fekel 1997).
133:
134: \section{Observations and Data Reduction}
135: \label{sec:obs}
136: The late-type dwarf stars in our sample were observed with the XMM-Newton satellite using varying detector set-ups.
137: The XMM-Newton satellite is composed of three co-aligned X-ray telescopes
138: (Jansen et al. 2001) which observe
139: a source simultaneously, accumulating photons in three CCD-based instruments: the twin
140: MOS 1 and MOS 2 and the PN (Turner et al. 2001; Str\"{u}der et al. 2001),
141: all three detectors constituting the EPIC (European Photon Imaging Camera) camera.
142: The EPIC instrument consists of three
143: CCD cameras with two different types of CCD, two MOS and one PN, providing the imaging
144: and spectroscopy in the energy range from 0.15 to 15 keV with a good angular
145: (PSF = 6 arcsec (FWHM)) and a moderate spectral resolution ($E/\Delta E \approx 20 - 50 $).
146: Exposure time for each star was in the range of 30-60 ks.
147: A log of observations is provided in Table 2.
148:
149: The data were reduced with standard XMM-Newton Science Analysis System (SAS)
150: software, version 7.0 with updated calibration files (Ehle et al. 2004).
151: The preliminary processing of raw EPIC Observation Data Files was done
152: using the {\it epchain} and {\it emchain} tasks which allow calibration both in energy and
153: astrometry of the events registered in each
154: CCD chip and to combine them in a single data file for MOS
155: and PN detectors.
156: The background contribution is particularly relevant at high energies
157: where coronal sources have very little flux and are often undetectable.
158: Therefore, for further analysis we have selected the energy range between
159: 0.3 to 10.0 keV.
160: Event list file was extracted using the SAS task {\it evselect}.
161: The {\it epatplot} task was used for checking the existence of pile-up
162: affecting the inner region of all stars. Only the star CC Eri was found to be
163: affected by pile-up.
164: X-ray light curves and spectra of all target stars were generated
165: from on-source counts obtained from circular regions with
166: a radius $\sim$ 40-50\arcs ~around each source. However, for the star CC Eri,
167: X-ray light curves and spectra were extracted with events taken from an
168: annulus of 32\arcs with inner radius of 12\arcs to avoid pile-up effect in the inner region.
169: The background was taken from several source free regions on the detectors
170: at nearly the same offset as the source and surrounding the source.
171:
172:
173: \section{Analysis and Results}
174: \label{sec:analysis}
175: \subsection{Multi-band X-ray light curves and characterization of flares}
176: The background subtracted X-ray light
177: curves of six stars viz. V356 Cep, IM Vir, V471 Tau, CC Eri and EP Eri,
178: as observed with the MOS and PN detectors are shown in Figures \ref{fig:mospnlc}(a) to
179: \ref{fig:mospnlc}(f), respectively. All light curves are in the energy band 0.3-10.0 keV.
180: The MOS and PN light curves
181: are represented by solid and open circles, respectively. The temporal
182: binning of light curves is 200 s for all sources.
183: The PN light curve of the star IM Vir was affected by proton flares twice
184: during the observations, first for 800 s after 10 ks from the beginning of the
185: observation, and second at the end, after 40 ks of the observations. However,
186: the MOS light curve was affected only at the end of the observation. For the
187: star EP Eri, the high proton flare background region in the light curve
188: is for a duration of 3 ks after 14.7 ks from the beginning of the observation.
189: We have removed data during the proton flare background from the MOS and PN light curves of
190: both the stars IM Vir and EP Eri. The star XI Boo
191: was not observed by the PN detector, and only the light curve obtained with
192: the MOS is presented here.
193:
194: The light curves of all the individual stars show variability on a time
195: scale of ks, most of which resembles flaring activity.
196: The customary definition of a flare is a significant increase in intensity,
197: after which the initial or quiescent level of intensity is reached again.
198: Such patterns are observed in the light curves of the sample stars
199: shown in Figures \ref{fig:mospnlc}(a) to \ref{fig:mospnlc}(f), and
200: where the
201: ``flare regions'' are represented by arrows and marked by Fi, where i=1,2...17 refers to the flare number. The
202: mean quiescent level count rates of the sources are taken from
203: regions that are free from flares and are marked by Q in the Figure \ref{fig:mospnlc}.
204: The start
205: time, end time, flare duration and flux during quiescent and at flare peak for all the
206: 17 flares observed are listed in Table 3. The peak and quiescent state
207: count rates as given in Table 3 are converted into flux by using the WebPIMMS\footnote{http://heasarc.gsfc.nasa.gov/Tools/w3pimms.html},
208: where we assumed the plasma temperature of 1 keV.
209: Flares F6, F7, F9, F14 and F16 were found to be long lasting ($\geq 2.3$ h) flares.
210: Infact, the flare F7 of the XI Boo
211: is longest flare observed with total duration of $\sim 3.1$ h.
212: The peak flux in these long duration flares was found to be more than two times
213: than the quiescent state. Other detected flares
214: were shorter than 1.3 h. The shortest duration flare was detected one
215: in V368 Cep (F3) and other in CC Eri (F12). In these short duration flares,
216: peak X-ray flux was 1.3 - 1.8 times more than that of the quiescent state.
217:
218: To characterize the flares,
219: we have fitted the light curves of the flares with an exponential function
220:
221: \begin{equation}
222: c(t) = A_{0}{\rm e}^{-\frac{t-t_0}{\tau_d}} + q
223: \label{eq:efold}
224: \end{equation}
225:
226: where c(t) is the count rate as a function of time t, t$_0$ is the time of peak
227: count rate, q is the count rate in the quiescent state, $\tau_d$ is the decay
228: time of the flare, and $A_0$ is the count rate at flare peak. The best fitted
229: parameters for the flares are given in Table 4 for both the MOS and the PN data.
230: Five flares, F6 and F7 of XI Boo,
231: F9 of IM Vir, F14 of CC Eri and F16 of EP Eri the e-fold decay time was
232: found to be more than 1 h.
233: Flare F7 of XI Boo was found to be the longest decay
234: flare ($\tau_d = 10$ ks) observed among all the seventeen flares.
235: The flare F16 shows a very different structure when
236: compared to the other flares, i.e. a sharp rise and a highly structured decay
237: as if more flares were present during the decay. Though it was tough to
238: locate the peak of flare F16 because of its complex structure, the decay
239: time was determined to be 4 ks. For
240: the remaining eleven flares the decay time was found to be in the range of
241: 8 - 47 minutes. The decay phase of the flare F17 was not observed. The rise times
242: of all flares were found to be less than 1 h. The rise and decay time of the
243: shortest flare F3 of V368 Cep was found to be 0.5 ks.
244: In XI Boo, after the flare F6
245: an active level U was identified, where average flux was 1.8 times more than
246: that of the Q state (see Figure \ref{fig:mospnlc}(b)), that was higher than
247: the peak of the flare F5 indicating a very high activity level.
248: Two similar highly active regions U1 and U2 were identified between the flares F13 and
249: F14. The average fluxes during the
250: regions U1 and U2 were 1.3 and 1.4 times more than that of the quiescent
251: state. Presence of
252: strong substructures in these active regions indicates superposition of a large
253: number of flaring regions.
254:
255: To investigate the behavior of the flares detected in different energy bands,
256: the light curves of the stars obtained with the MOS and PN
257: are divided into three energy bands namely soft (0.3 - 0.8 keV), medium (0.8 -
258: 1.6 keV), and hard (1.6 - 10 keV). The boundaries of the selected energy bands are
259: chosen as the line free regions of the low resolution PN spectra.
260: The hardness ratio HR1 and HR2 are defined
261: by the ratio of medium to soft band, and hard to medium band count rate,
262: respectively. The soft, medium, hard band intensity curves, and the hardness
263: ratio curves HR1 and HR2 as a function of time
264: are shown in the sub-panels running from top to bottom in left(MOS) and
265: right(PN) of Figure \ref{fig:lchr} (a) for the star V368 Cep.
266: Similar plots of intensity and hardness ratios observed from the
267: stars XI Boo, IM Vir, V471 Tau, CC Eri and EP Eri are shown in the
268: Figures \ref{fig:lchr}(b) to \ref{fig:lchr}(f), respectively.
269: The light curves in individual band also show significant variability in all sources.
270: We have determined the e-folding decay time of observed flares
271: in the three bands by fitting equation (\ref{eq:efold}).
272: The decay times, rise times, count rates at flare peaks,
273: and quiescent state count rates at soft, medium and hard band
274: for PN data are
275: given in the Table 5. For most of the flares, e-folding decay time in the soft
276: band was found to be more than the decay time in the medium and the hard band.
277: However, for the flare F3 of V368 Cep and F7 of XI Boo, it appears that
278: decay time in the medium band was more than in the soft band.
279: For the flares F6, F7 and F9 the decay time in the soft band was
280: found to be more than 3.0 ks than that of the hard band.
281: However, for other flare events the decay time in the
282: soft band was $\lesssim 2$ ks than that in the hard band.
283: Two smaller flare like structures were seen just after the peak
284: of flare F13 of CC Eri, in the hard band but were not observed in the soft and medium band
285: light curves (see Figure \ref{fig:lchr}(e)). The rise time for all flares in all band
286: was found to be well within $1\sigma$ level (Table 5).
287:
288: The flare peak to the quiescent state flux ratio in the hard band was found in between
289: 2-16, which is more than that of the similar ratios in the medium and the soft band (see also Table 5) for all
290: flares. In terms of the peak flux, the flares F6 and F7 of XI Boo, and F13 and F14 of CC Eri
291: were stronger in the medium band than in the soft and the hard band.
292: The variation in the hardness ratio are indicative of changes in coronal temperature.
293: An X-ray flare is well defined by a rise in the temperature and subsequent decay.
294: For the flares F6, F7, F9, F12, F13 and F14, both HR1 and HR2 varied in the similar fashion
295: to their light curves (see Figure \ref{fig:lchr}). This implies that
296: an increase in the temperature at flare peak and subsequent cooling. The structures
297: observed during the decay phase of the flares F2 of V368 Cep and F13 of CC Eri in the hard
298: band were also
299: observed in the HR2 curve. This could probably be due to emergence of flares
300: during the decay phase of these flares, which were comparatively stronger in the hard band.
301: During the flares F2, F3 and F4 of V368 Cep, and F16 of EP Eri the HR1 did not
302: show any variation but HR2 varied according to the flare intensity. However, HR1 and HR2 did
303: not show any significant variation during the flares F1, F5, F8, F10, F11, F15 and F17.
304: The PN data have more count rates than the MOS and have similar characteristics
305: to that of MOS, therefore, for further analysis in this paper we use only
306: the PN data.
307:
308:
309: \subsection{Colour-colour diagrams}
310: \label{sec:ccd}
311: Plots of hardness ratio in the form of colour-colour (CC) diagram can reveal
312: information about spectral variations and serve as a guide for a more
313: detailed spectral analysis. For this purpose, the light curves and the
314: hardness ratio curves of the flares were divided into different time segments
315: covering their rising and decaying phases as
316: shown in Figure \ref{fig:mospnlc}, where Ri(i=1,2..) represents the rising phase, Di(i=1,2,..) represents the
317: decay phase and Q represents the quiescent state.
318: We determined the HR1 and HR2 from the PN data for each of these
319: segments and plotted the values in Figure \ref{fig:hr}.
320: To understand the observed behavior of HR1 and HR2 in terms of simple spectral models,
321: we generated the soft(0.3-0.8 keV), medium(0.8-1.6 keV) and
322: hard(1.6-10.0 keV) band count rate using the {\it ``fakeit''}
323: provision in the XSPEC,
324: using the most recent Canned Response Matrices downloaded from
325: http://xmm.vilspa.esa.es/external/xmm\_SW\_cal/calib/epic\_files.html, and
326: simple two temperature plasma models. The two temperature APEC models
327: were mixed such that the emission measure were in the
328: ratio $EM_2/EM_1$ of 0.4, 0.6, 0.8, 1.0 and 1.2 (based on spectral
329: results in \S 4.3 and \S 4.4). The abundances (0.25)
330: and hydrogen column density ($10^{20}$ cm$^{-2}$) were kept fixed for each
331: set of
332: the $EM_2$/$EM_1$ (see \S \ref{sec:quiescent} and \S \ref{sec:flare}). Further by keeping $kT_1$ fixed at different values
333: ranging from 0.2 to 1.2 keV in steps of 0.2 and varying $kT_2$ from 0.4 to 2.4 keV such that
334: $kT_2 > kT_1$, we
335: generated count rates in the soft, medium and hard bands. These predicted count rates
336: were used to generate the hardness ratios.
337: The families of curve thus generated
338: are over-plotted on the data points in Figure \ref{fig:hr} to understand
339: the CC diagram. The kT$_2$ in Figure \ref{fig:hr} increases from
340: bottom to top of each curve.
341:
342: Figure \ref{fig:hr} (a) shows the plot between HR1 and HR2 for a set of
343: $EM_2$/$EM_1$ = 1.0 for the star V368 Cep.
344: In the case of the flare F1, 'D1' intersects
345: the generated CC curves for which $kT_1$ is 0.6 keV, however,
346: 'D4' intersects the curve generated for $kT_1 ~=$
347: 0.2 keV. A similar pattern is seen for the other flares F2 and F4 of the stars
348: V368 Cep, where the top of the flare is located near the CC curve
349: for a high temperature component whereas the colours near the
350: end of the flare intersect the CC curve corresponding to a
351: lower temperature. The flare F2 shows a very high
352: temperature during its rise phase 'R1' and the decay phase 'D5', as both are
353: located between the CC curves of 1.0 and 1.2 keV. The decay phase D6
354: of the flare F2 is located near the 0.8 keV CC curve.
355: The peak seen during the decay phase D6 of flare F2 (Fig \ref{fig:hr}(a)) can be
356: explained by a rise in the temperature of coronal plasma. It appears that
357: the quiescent state 'Q' of the star V368 Cep has a cool temperature between
358: 0.2 to 0.4 keV.
359:
360: The peak of the largest flare 'F9' observed in the star IM Vir appears to have a temperature
361: between 1.0 - 1.2 keV. As the flare decayed the temperature was also observed to decrease.
362: A similar trend is seen during the flares observed in the stars XI Boo (Figure \ref{fig:hr}(b)), V471 Tau
363: (Figure \ref{fig:hr}(d)), CC Eri (Figure \ref{fig:hr}(e)), and EP Eri (Figure \ref{fig:hr}(f)). The observed CC curves for the star IM Vir, V471 Tau and CC Eri
364: were well matched with the generated family of CC curves for a set of $EM_1/EM_2$
365: = 1.0. However, the observed CC diagram of EP Eri were well matched with
366: generated CC curves for a set of $EM_1/EM_2 = 0.4$.
367: It appears that high temperatures are needed during
368: the flare peak, and the plasma temperatures decrease as flare decays.
369: To, further confirm and quantify the spectral changes during the
370: flare, we have also performed a detailed spectral analysis of the flares and
371: the quiescent states of the stars.
372: The results of these analyses are presented below.
373:
374: \subsection{Quiescent state X-ray spectra}
375: \label{sec:quiescent}
376: X-ray spectra for each star during the quiescent state were
377: analysed using XSPEC version 12.3 (Arnaud 1996).
378: Data were cleaned of proton flares by removing the affected time
379: periods for the stars IM Vir and EP Eri.
380: Spectral analysis of EPIC data was performed in the energy band
381: between 0.3-10.0 keV for the star V368 Cep, XI Boo, V471 Tau and
382: CC Eri, while for the stars IM Vir and EP Eri spectral analysis
383: was performed between 0.3-5.1 keV.
384: Individual spectra were binned so as to have a minimum of 20 counts per energy bin.
385: The EPIC spectra of the stars were fitted with a single (1T) and two (2T)
386: temperature
387: collisional plasma model known as APEC (Astrophysical Plasma Emission Code, Smith et al. 2001),
388: with variable elemental abundances.
389: Abundances for all the elements in the APEC were varied together.
390: The interstellar hydrogen
391: column density ($N_H$) was left free to vary.
392: For all the stars, no 1T or 2T
393: plasma models with solar photospheric (Anders \& Grevesse 1989)
394: abundances could fit the data, as unacceptably large values of
395: $\chi^{2}_\nu$ were obtained.
396: Acceptable APEC 2T
397: fits were achieved only when the abundances were allowed to depart
398: from the solar values.
399: The best-fit 2T plasma models with sub-solar abundances along with the
400: significance of
401: the residuals in terms of $\Delta \chi^2$ are shown in Figures
402: \ref{fig:qspec}(a) to (f) for all the stars.
403: Table 6 summarizes the best-fit values obtained for the various
404: parameters along with the minimum $\chi_\nu^2 = \chi^2/\nu$
405: (where $\nu$ is degrees of freedom) and the
406: 90\% confidence error bars estimated from the minimum $\chi^2+2.71$.
407: The derived values $N_H$ from the spectral analysis were found to be lower than that of
408: the total galactic HI column density (Dicky \& Lockman 1990) towards the
409: direction of that star. The cool and hot temperatures at the quiescent state
410: of these stars were found in the range of 0.2 - 0.5 and 0.6 - 1.0 keV, respectively.
411: The corresponding EM2/EM1 during the quiescent state of each star was found in the range of
412: 0.8 - 1.2, except for the star EP Eri where it was found to be $\sim 0.4$.
413:
414: \subsection{Spectral evolution of X-ray flares}
415: \label{sec:flare}
416: In order to trace the spectral changes during
417: the flares, we have analysed the spectra of the different
418: time intervals shown in Figure \ref{fig:mospnlc}.
419: To study the flare emission only, we have performed 1-T spectral fits of the
420: data, with the quiescent emission taken into account by including its
421: best-fit 2-T model as a frozen background contribution. This is equivalent
422: to consider the flare emission subtracted of the quiescent level, allows us
423: to derive one ``effective'' temperature and one emission measure of the
424: flaring plasma.
425: The abundances were kept fix to that of the quiescent emission.
426: The best fit spectral parameters for each flare segments are listed in
427: Table 7. Figures \ref{fig:v368ZTEM} - \ref{fig:epZTEM} show the
428: temporal evolution of the temperature and
429: corresponding emission measure of the flares detected in the targeted stars,
430: where '0' time corresponds to the flare peak time.
431:
432: \subsubsection{V368 Cep}
433: \label{sec:V368}
434: Spectral data of flares F1, F2 and F4 were binned into four segments. However,
435: for the flare F3 data could be collected over only
436: one segment. Both the temperature and the corresponding
437: emission measure were found to decrease from the flare peak to the quiescent
438: state (see also Table 7). As shown in Figure \ref{fig:v368ZTEM}(a), the highest temperature
439: for the flare F1 was 14 MK and decreased along the decay path to 6.8 MK. A similar
440: trend was also found in the emission measure, where it increased by a
441: factor of seven (see Figure \ref{fig:v368ZTEM} (b)).
442: The temperature and emission measure for the flare F2
443: were peaked simultaneously during the decay phase D5.
444: However, for the flare
445: F4 both parameters were peaked during the rise phase R2.
446: The maximum luminosity was found during the flares F2 and
447: peaked at $1.18\times10^{30}$ \lum, which is $\sim 1.7$ times more than
448: that of the quiescent state.
449:
450: \subsubsection{XI Boo}
451: To see the flare evolution, flare F6 and F7 have binned into five and six intervals, respectively.
452: For the flare segments R2, D5, D6, D7 and D8, the 1T APEC fit gave high value of
453: $\chi^2_\nu$ (=1.5 for R2, 1.75 for D5, 1.61 for D6, 1.55 for D7 and 1.34 for D8).
454: The spectral fit to the data for these time segments improves the $\chi^2$
455: by using the 2T plasma model (see Table 7). The temperature and the corresponding
456: emission measure of the cool component for these time segments were found to be constant
457: at $0.6\pm0.1$ keV and $4.0\pm1.0 \times 10^{51}$ cm$^{-3}$, respectively.
458: The spectral parameters for hot component of the best fit 2T plasma model
459: are reported in Table 7.
460: The evolution of the temperature of hot component and the corresponding emission measure are shown
461: in Figures \ref{fig:xiZTEM} (a) and \ref{fig:xiZTEM} (b), respectively.
462: For the flares F6 and F7, both the temperature and corresponding emission measure were
463: reached to the maximum value during the decay phase.
464: As mentioned in Table 6 and 7, during the flares F6 and F7,
465: XI Boo was $\sim 2$ times more X-ray luminous than that of the
466: quiescent state.
467:
468: \subsubsection{IM Vir}
469: The spectral data of the flares F8 and F9 were binned in two and four different time segments, respectively
470: (see Figure \ref{fig:mospnlc}(c)) and spectra of each segment were fitted with 1T APEC model.
471: The plots of temperature and the corresponding emission
472: measure are shown
473: in Figures \ref{fig:imZTEM} (a), and (b), respectively.
474: For the flare F9, the maximum fit temperature,
475: was reached during the rising phase R2, somewhat earlier than the
476: maximum of its emission measure. The emission measure was increased by a factor
477: of four during the flare F9. The luminosity at the peak of flare F9 was
478: found to be $\sim 2$ times
479: more than that of the quiescent state.
480:
481: \subsubsection{V471 Tau}
482: Two flares detected in the star V471 Tau were examined by analysing four
483: different time intervals during the flares. Figures.
484: \ref{fig:v471ZTEM} (a) and (b) show the plot of temperature
485: and the corresponding emission measure during the flares F10
486: and F11. For the flare F10 the temperature reached its
487: maximum value of $17.8$ MK and decreased along the decay
488: phase (see Table 7). The corresponding emission measure was also
489: peaked during the decay phase of the flare F10 (see Figure
490: \ref{fig:v471ZTEM}(b) and Table 7). For the flare F11,
491: both the temperature and the emission measure follow similar
492: trend to the flare light curve.
493: The X-ray luminosity was increased by a factor of $\sim 1.6$ for the
494: flare F10 and $\sim 1.4$ for the flare F11.
495:
496:
497: \subsubsection{CC Eri}
498: The evolution of temperature and the corresponding emission measure for
499: three flares F12, F13 and F14 are shown in Figures \ref{fig:ccZTEM} (a) and (b),
500: respectively.
501: Small variations were found in the temperature and the emission
502: measure during
503: the flares F13 and F14. Both the temperatures and the corresponding emission
504: measures were found to decrease during the decay
505: phase for all flares (see also Table 7). The temperature and the corresponding
506: emission measure for all the flares were peaked during the decay phase. However, it appears that
507: temperature is peaked before the emission measure for the flare F13. The peak
508: luminosity during flares F13 and F14 was found to be $6.0 \times 10^{28}$ \lum ~more
509: than that of the quiescent state.
510:
511:
512: \subsubsection{EP Eri}
513: Spectra of different time segments of the flares are fitted with 1T plasma models.
514: Figures \ref{fig:epZTEM} (a) and (b) show the evolution of
515: temperature and the corresponding emission
516: measure for the flares F15 and F16. For the flare F16, both the temperature and
517: the corresponding emission measure were peaked during the rise phase R2, remained
518: constant upto the decay phase D2 and then decreased along the flare decay
519: without any appreciable changes in the luminosity. The luminosity during both
520: flares was found to be similar to that of the quiescent state (see Table 6 and Table 7).
521:
522: \subsection{Loop modeling of the X-ray flares}
523: \label{sec:model}
524: Flares can not be resolved spatially on a star. However, by an analogy
525: with solar flares and using flare loop models, it is possible to infer
526: the physical size and morphology of the loop structures involved in a stellar
527: flare.
528: The most widely used but approximate methods for analyzing stellar flares are:
529: i) Two-ribbon flare method (Kopp \& Poletto 1984), ii) Quasi-Static
530: Cooling method (van den Oord \& Mewe 1989), iii) Pure radiation
531: cooling method ({\rm Pallavicini et al. 1990}), iv) Rise and decay time method
532: (Hawley et al. 1995), and v) Hydrodynamic method (Reale et al. 1997).
533: The Two-ribbon flare model assumes that the flare decay is entirely
534: driven by heating released by magnetic reconnection of higher and higher
535: loops and neglects completely the effect of plasma cooling. The other
536: four methods are instead based on the cooling of plasma confined in a single
537: flaring loop. In the
538: hypothesis of flares occurring inside closed coronal structures,
539: the decay time of the X-ray emission roughly scales as the
540: plasma cooling time. In turn, the cooling time scales with the
541: length of the structure which confines the plasma: the longer
542: the decay, the larger is the structure (e.g. Haisch et al.1983). A loop
543: thermodynamic decay time has been derived (van den Oord \&
544: Mewe 1989; Serio et al. 1991) as:
545:
546: \begin{equation}
547: \tau_{th} = \frac{120 L_9}{\sqrt{T_7}}
548: \label{eq:l1}
549: \end{equation}
550:
551: where $L_9$ and $T_7$ are the loop half-length and the maximum
552: temperature (T$_{max}$) of the flaring plasma, in units of $10^9$ cm and $10^7$ K,
553: respectively. The timescale above is derived under the
554: hypothesis of impulsive heat released at the beginning of a flare.
555:
556: The hydrodynamic model includes both plasma cooling and
557: the effect of heating during flare decay.
558: Reale et al. (1997) presented a method to infer the geometrical size
559: and other relevant physical parameters of the flaring loops, based on the
560: decay time and on evolution of temperature and the emission measure (EM)
561: during the flare decay. The empirical formula for the estimation of the
562: unresolved flaring loop length has been derived as:
563:
564: \begin{equation}
565: L_{9} = \frac{\tau_{d}\sqrt{T_{7}}}{120 f(\zeta)} ~~~~~~~~ f(\zeta) \geq 1
566: \label{eq:l2}
567: \end{equation}
568:
569: \noindent
570: where $\tau_{d}$ is the decay time
571: of the light curve, and $f(\zeta)$ is a non-dimensional correction factor larger
572: than one. $\zeta$ is the slope of the decay path in the density-temperature
573: diagram (Sylwester et al. 1993) and is maximum ($\sim 2$) if heating is
574: negligible and minimum ($\sim 0.5$) if heating dominates the decay. The
575: correction factor $f(\zeta)$ and loop maximum temperature are calibrated
576: for the XMM-Newton EPIC spectral response and are given as (Reale 2007):
577:
578: \begin{equation}
579: f(\zeta) = \frac{0.51}{\zeta - 0.35} + 1.36 ~~~~~~~~~~~\rm{ (for ~0.35 < \zeta \leq 1.6)}
580: \label{eq:fxi}
581: \end{equation}
582:
583: \begin{equation}
584: T_{max} = 0.13 T_{obs}^{1.16}
585: \label{eq:tmax}
586: \end{equation}
587:
588: Other physical properties of the flaring plasma can be inferred from the
589: analysis of flare data. The analysis of X-ray spectra provides
590: values of the temperature and emission measure.
591: From the emission measure (EM) and the plasma density ($n_e$), the
592: volume of the flaring loop is estimated as (Reale 2002):
593:
594: \begin{equation}
595: V = \frac{EM}{n^2_e}
596: \label{eq:density}
597: \end{equation}
598:
599: \noindent
600: where V is the loop volume. It has been
601: shown that in the equilibrium condition the loop scaling laws hold, linking maximum
602: pressure, temperature, loop length and heating rate per unit volume as
603: (Rosner, Tucker \& Vaiana 1978, Kuin \& Martens 1982):
604:
605: \begin{equation}
606: T_{max} = 1.4 \times 10^3 (pL)^{1/3}
607: \label{eq:pressure}
608: \end{equation}
609:
610: \begin{equation}
611: E_H \approx 10^{-6} T_{max}^{3.5} L^{-2} ~~~ergs ~s^{-1} cm^{-3}
612: \label{eq:EH}
613: \end{equation}
614:
615: \noindent
616: The minimum magnetic field necessary to confine the flaring plasma can be
617: simply estimated as
618:
619: \begin{equation}
620: B = \sqrt{8 \pi p}
621: \label{eq:B}
622: \end{equation}
623:
624: Because of limited statistics, we have modeled only ten flares.
625: The results of the model parameters are listed in Table 8. In this
626: table,
627: column 1 is name of the star and the corresponding flare, column 2 is
628: slope of density-temperature diagram ($\zeta$), column 3 contains
629: loop maximum temperature based on spectral fit and equation (5),
630: semiloop length is determined using the equation (3) and given in column
631: 4, pressure in the loop at the flare peak is estimated using the
632: loop scaling law (see equation (7); column 5), column 6 contains maximum
633: electron density at the flare loop (we have assumed a totally ionized
634: hydrogen plasma i.e. $p = 2n_ekT_{max}$; since pressure derived from equation (7) is a maximum value, therefore, a upper limit on density is estimated), volume of the flaring plasma
635: is estimated from maximum emission measure and from the estimated
636: maximum electron density, and therefore, is treated as
637: lower limit (equation (6); column 7), column 8
638: is heating rate per unit volume and estimated using equation (8),
639: the minimum magnetic filed necessary to confine the flaring plasma
640: at loop apex $B$ (see equation \ref{eq:B}) is given in column 9,
641: the loop aspect ratio $\beta$ (=$r/L$, $r$ is radius of the loop
642: cross section, assuming circular) estimated from the
643: loop scaling law $V = 2\pi \beta^2 L^3$ and number of loops
644: filling the flare volume estimated assuming that $\beta = 0.1$
645: for a single loop are given in column 9 and 10, respectively.
646: The model parameters for the different flares from each star are
647: presented and discussed below.
648:
649: \subsubsection{Flares F1, F2 and F4 of V368 Cep}
650: The average temperatures of the loops, usually
651: lower than the real loop maximum temperatures, are found from the
652: spectral analysis of the data as given in \S \ref{sec:flare}.
653: According to equation (\ref{eq:tmax}), the loop maximum temperatures for the flare
654: F1, F2 and F4 are found to be $25.5\pm2.5$, $62\pm8$ and $40.3\pm1.8$ MK, respectively. Figure \ref{fig:nt} (a)
655: shows the density-temperature (n-T) diagram,
656: where EM$^{1/2}$ has been used as a proxy of density. In Figure \ref{fig:nt}
657: (a) the flares F1 ,F2 and F4 are represented by solid circles , solid triangles and solid squares, respectively.
658: The solid lines represent the best linear fit to the corresponding data,
659: providing the slope $\zeta$.
660: These values of
661: $\zeta$ indicate the presence of a sustained
662: heating during the decay of the flares F1 and F2. However, sustained heating is negligible during the decay of the flare F4.
663: As discussed in \S \ref{sec:model}, the e-folding time of the light curve and the slope
664: of n-T path can be used to estimate the loop half-length.
665: According to equation (3) the semiloop lengths for the flares F1, F2 and F4 are
666: determined to be $6.0\pm0.6 \times 10^9 ~(=0.1R_\star)$, $19.8\pm8.6
667: ~(= 0.4R_\star)$ and $11.7\pm1.3 \times 10^9 ~(= 0.2R_\star)$ cm,
668: respectively. These loop lengths are much smaller than the pressure
669: scale height\footnote{defined as $h_p = 2kT/\mu g \sim 5000 \times T_{max}/(g/g_\odot)$,
670: where T is plasma temperature in the loop, $\mu$ is the molecular weight of the
671: plasmas and g is the surface gravity of the star.} $h_p > 1.2\times10^{11}$ cm.
672: Maximum pressure in the loop at the flare peak is estimated to be
673: $>10^3$ dyne cm$^{-2}$ for the flares F1, F2 and F4,
674: respectively (see equation (7)).
675: Assuming that the hydrogen plasma is totally ionized ($p = 2n_ekT_{max}$),
676: the maximum plasma density in the loop at the flare peak is estimated
677: in the order of $10^{11}$ cm$^{-3}$
678: for the flares F1, F2 and F4, respectively.
679: We computed a volume of $\sim 1.78\times 10^{30}$ cm$^{-3}$ for the flare F1
680: using the observed peak EM of $3.7\times10^{52}$ cm$^{-3}$.
681: Similarly, loop volume is determined to be $7.2\times10^{29}$ cm$^3$ for flare F2
682: and $7.4 \times10^{29}$ cm$^3$ for the flare F4.
683: A hint for the heat pulse intensity comes from the flare maximum temperature. By applying
684: the loop scaling laws and loop maximum temperature (see equation (\ref{eq:EH})) the heat
685: pulse intensity for the flares F1, F2 and F4 would be the order of 2.36, 4.75
686: and 3.06 ergs cm$^{-3}$ s$^{-1}$, respectively.
687: From the pressure of the flare plasma the magnetic field required
688: to confine the plasma should be more than the 160 Gauss for all the flares.
689:
690: \subsubsection{Flares F6 and F7 of XI Boo}
691: The evolution of flare F6 and F7 in n - T plane is shown in
692: Figure \ref{fig:nt} (b), together with a least square fit to the decay phase.
693: The resulting best-fit slopes ($\zeta$) for the decaying phase of flare F6 and F7
694: indicate
695: that the flare F6 is driven by the time scale of the heating process, whereas
696: sustained heating is negligible during the decay of the flare F7. The intrinsic flare
697: peak temperatures for flare F6 and F7 are, applying equation (5) to the observed maximum temperature
698: , $18.9\pm0.8$ and $25.3\pm2.7$ MK, respectively.
699: For the flare F6 the semiloop length is estimated to
700: be less than the pressure
701: scale height, $h_p = 10^{11}$ cm. Maximum pressure and density in the loop at
702: the peak of the flare F6 are estimated to be $130\pm 6$ dyne cm$^{-2}$ and $2.5\pm1.0
703: \times 10^{10}$ cm$^{-3}$, respectively. Using equation \ref{eq:density},
704: volume of the flaring plasma is estimated to be $1.1\times10^{31}$ cm$^3$.
705: A minimum magnetic field of $\sim 58$ Gauss is required to confine
706: the flaring plasma at the peak of the flare.
707: Although, the value of $\zeta$
708: for the flare F7 is outside the domain of the validity of the
709: method (see Real 2007), the estimated loop parameters obtained
710: from the hydrodynamic loop modeling are listed in Table 8.
711: The estimated semiloop length for the flare F7
712: is smaller than the pressure scale height (=$1.3\times10^{11}$ cm).
713:
714: \subsubsection{Flare F9 of IM Vir}
715: Using equation (\ref{eq:tmax}) the maximum loop temperature for the
716: flare F9 is found to be $46.6\pm3.7$ MK.
717: Figure \ref{fig:nt} (c) shows the n-T diagram for the flare F9.
718: From the best linear fit, we found the slope of the n-T diagram
719: implies that the heating during the flare decay
720: was as strong as in the flares F1, F2 and F6.
721: The resulting semiloop length
722: is found to be
723: ten times smaller than the pressure
724: scale height ($h_p = 2.3\times10^{11}$ cm),
725: which implies a loop in hydrostatic
726: equilibrium with a plasma pressure of $2020\pm1100$ dyne cm$^{-2}$ and
727: maximum heating rate of $\approx 2.07 $ ergs cm$^{-3}$ s$^{-1}$.
728: The maximum plasma density and loop volume for the flare are estimated
729: to be $1.6\pm0.8 \times10^{11}$ cm$^{-3}$ and $4.55\times10^{30}$ cm$^{3}$, respectively.
730: The minimum confining magnetic field is estimated to be $\sim 226$ Gauss.
731:
732: \subsubsection{Flare F10 of V471 Tau}
733: The maximum temperature in the loop at the flare peak was found to be $33.5\pm3.3$ MK for the flare F10.
734: (see Figure \ref{fig:nt}(d) for the n-T diagram of this flare).
735: The slope of the decay path of n-T diagram
736: indicates the presence of sustained heating.
737: Using equation (3) the semiloop length of the flare F10 is calculated to be
738: $\sim 0.4R_\star$ and is $\sim 5$ times smaller than the pressure scale height $h_p = 1.2\times10^{11}$ cm.
739: Using these parameters and the scaling laws for the loops, the maximum pressure, the maximum density
740: and the minimum magnetic field confining the flaring plasma in the
741: loop at the flare peak, the volume of flaring plasma and the heating rate per unit volume for
742: F10 are estimated to be $p= 630\pm200$ dyne cm$^{-2}$, $n_e=6.8\pm1.0\times10^{10}$ cm$^{-3}$ and $B=126$ Gauss,
743: $8.9\times10^{30}$ cm$^{3}$ and 0.46 ergs s$^{-1}$ cm$^{-3}$.
744:
745: \subsubsection{Flares F13 and F14 of CC Eri}
746: Using equation (\ref{eq:tmax}) the maximum
747: temperature in the loop responsible for the flares F13 and F14 are found to be $25.7\pm2.9$ and $27.0\pm2.1$ MK,
748: respectively.
749: Figure \ref{fig:nt} (e) shows the n-T diagram for these flares. The slopes in
750: n-T diagram for the flares F13 and F14
751: indicate the presence of strong sustained heating, and showing that
752: the observed decay is driven by the time-evolution of the heating process.
753: The value of $\zeta$ at the low extreme of the error bar
754: for the flare F13 is compatible with the lower
755: asymptotic value for which equation (4) can be applied.
756: Therefore, an upper limit for semi-length of the loop can be derived
757: by using the value of $\zeta$ at the high extreme of the error bar.
758: The resulting semi-length for the flares F13 and F14 are $<1.4 \times 10^{10} (=0.3/0.5 R_\star) $ and
759: $2.0\pm1.0 \times 10^{10} (=0.4/0.7 R_\star)$ cm, respectively.
760: These values of the loop length are much smaller than the pressure scale
761: height $h_p = 1.1\times 10^{11}$ for flare F13 and $h_p = 7.3\times10^{10}$
762: cm for flare F14.
763: A magnetic field more than 95 Gauss is required to confine the flaring plasma
764: at the top of both flares. A detailed analysis
765: of these flares is also given in Crespo-Chacon et al. (2007). The estimated
766: loop parameters for both flares are found well within $1\sigma$ level to that of
767: Crespo-Chacon et al. (2007).
768:
769: \subsubsection{Flare F16 of EP Eri}
770: The trend in n-T diagram suggests the
771: presence of significant heat during the decay of the flare F16
772: (See Figure \ref{fig:nt} (f) and Table 8). The temperature at
773: the peak of the flare is determined to be $20.2\pm2.2$ MK.
774: The derived semiloop length $L = 2.4\pm0.5 \times 10^{10}$ cm ($=0.4R_\star$) is
775: less than the pressure scale height $h_p = 9 \times 10^{10}$ cm.
776: Using the same approach as discussed above, the maximum pressure
777: and the maximum density are derived at $p = 130\pm50$ dyne cm$^{-2}$ and $n_e =
778: 2.3\pm0.5 \times 10^{10}$ cm$^{-3}$. Using the maximum emission
779: measure, the loop volume is estimated to be
780: $5.38\times10^{30}$ cm$^{3}$. The confining magnetic filed is estimated as
781: $B > 57$ Gauss.
782:
783: \section{Discussion and Conclusions}
784: \label{sec:discussion}
785: We have carried out an analysis of XMM-Newton observations
786: of six G-K dwarfs. Light curves revealed flaring in all the stars
787: on time scale ranging from seconds to kiloseconds, with
788: various peak strengths. A total of 17 flares were
789: identified in these G-K dwarfs. The decay time of these flares ranges
790: from 0.5 to 10 ks. A similar range of flare decay times
791: in the pre-main-sequence stars of Pleiades cluster was found by Stelzer, Neuh{\"a}user \& Hambaryan (2000).
792: Among the 17 flares presented here, only four flares (F6 and F7 of XI Boo, F9 of IM Vir and
793: F14 of CC Eri) appear to be long decay flare(\td ~$\geq$ 1 hr). Infact, the flare F7 of
794: XI Boo was one of the longest duration flares observed with \td ~
795: $\sim 10$ ks. Such long decay flares have so far been reported
796: in CF Tuc (\td ~= 79.2 ks; Kr\"{u}ster \& Schmitt 1996), EV Lac (\td ~= 37.8 ks; Schmitt
797: 1994), Algol (\td ~= 30.2 ks; Ottmann \& Schmitt 1996), AD Leo
798: (\td ~= 7.9 ks; {\rm Favata, Micela \& Reale 2000a}) and some pre-main sequence stars ({\rm Stelzer, Neuh{\"a}user \& Hambaryan 2000, Favata et al. 2005}).
799: This flare classification is purely based on the decay time.
800: It has been shown that even much longer flares can occur
801: in a single loop (Favata et al. 2005), while relatively short flares with heating dominated
802: decay are probably arcade flares. For the flares F6, F9 and F14, the decay
803: path is driven by the sustained heating, therefore, these flares can not
804: be classified as long decay flares. However, the sustained heating during the decay of
805: the flare F7 is negligible. Thus, these flares can be classified as arcade.
806: The flare decay time for the remaining 12 flares are similar to that of the solar compact
807: flares. The morphological differences in these different types of flares
808: indicate the different processes of energy released. In compact flares energy
809: is probably released only
810: during an impulsive phase, whereas in the two-ribbon flares a prolonged energy release
811: is apparently required to explain their long decay time (Pallavicini, Serio \& Vaiana 1977; Pallavicini et al. 1988;
812: Priest 1981; Poletto, Pallavicini \& Kopp 1988).
813: The rise time of these flares has been found to be less than 1 ks,
814: which is similar to the rise time of impulsive flares observed in the
815: M dwarfs ({\rm Pallavicini et al. 1990}).
816: However, for some flares (F6, F7 and F9) the rise time was found to be more than 1 ks.
817: We found that the decay times of the flares observed here are more in the soft
818: band than in the medium band or the hard band.
819: Some flares, e.g., F2, actually showed a hard peak during the decay.
820: But even then the \td ~in soft band being more than in the hard band.
821: This could probably be due to the softening of the spectrum during
822: the decay due to the plasma cooling i.e. emission gradually exists
823: form high energy band and enters more deeply in the soft energy band
824: or the higher energy electrons streaming down from
825: the coronal heights to the chromosphere that can heat the plasma
826: to higher temperature, lose their energy faster than the lower
827: energy electrons.
828: Most of the flares observed in our sample were found to have more
829: peak flux in the soft band than in the medium and hard band.
830:
831: The total energy released during the X-ray flares observed in the sample of
832: G-K dwarfs is observed to be in the range of $2.3\times10^{32}$ to $6.1\times10^{34}$ ergs.
833: This shows that these flares are ten times more
834: energetic than the flares in the M-type dwarfs ({\rm Pallavicini
835: et al. 1990}), and $10^2$ to $10^4$ times more energetic than the solar
836: flares (Moore et al. 1980, Wu et al. 1986). However, they are not as energetic
837: as a large flare observed in Algol (E$_{tot} = 10^{35}$ ergs;
838: White et al. 1986), but are as energetic as the flares observed in
839: G9 dwarf ZS 76 (E$_{tot}$ = $1.4 \times 10^{34}$ ergs; Pillitteri et al. 2005).
840: The large flare, F7, of XI Boo was the largest in terms of both the
841: decay time and the ratio of the peak luminosity to the quiescent luminosity (a
842: factor of 2.2). However, given the low quiescent emission (\lx ~=
843: $5.1\times10^{28}$ \lum) from the XI Boo, this flare was not particularly
844: prominent in terms of either peak X-ray luminosity (\lx ~= $1.1\times10^{29}$
845: \lum) or the total
846: X-ray energy (E$_{tot}$ = $6.4\times10^{33}$ ergs). Other large flare F9 of IM Vir
847: shows a large peak X-ray luminosity ($1.6\times10^{30}$ \lum)
848: as well as the total energy released (E$_{tot}$ = $6.12\times10^{34}$ ergs).
849:
850: Inspite of the fact that the
851: peak flare luminosity may vary by a factor of 2 (see, e.g., IM Vir in Table 7)
852: , there is a good correlation between the peak flare luminosities (L$_{Xf}$) and the quiescent state
853: stellar X-ray luminosities (L$_{Xq}$). The form of the correlation is
854: log L$_{Xf}$ = 1.04 log L$_{Xq}$ + 1.3, with the correlation coefficient, r, of
855: 0.9 (see Figure \ref{fig:lxlq}). A similar correlation was found by {\rm Pallavicini et al. (1990)}
856: for the M dwarfs. This indicates that there is a direct connection between
857: the flaring and the quiescent X-ray emission, which is probably due to their
858: common emission mechanism for X-rays.
859:
860: Flares F2-F3-F4, F6-F7 and F10-F11 show the similar structure i.e. before
861: ending the first flare the next flare starts.
862: Similar loop systems have been
863: observed to the flare on the Sun (e.g. so-called Bastille-day
864: flare; Aschwanden \& Alexander 2001), and in a stellar analogue
865: dMe star Proxima Centauri (Reale et al. 2004).
866: In these two events, a double ignition in nearby loops was observed or
867: suggested, and the delay between the ignitions appears to scale with the
868: loop sizes. Similar type of double exponential
869: decay was also found in the young stars
870: ZS 76 (Pillitteri et al. 2005) and Cygnus OB2 (Albacete Colombo et al. 2007).
871: This implies that
872: in these dwarfs a steady active region undergoes a strong magnetic
873: reconnection event resulting in an
874: intense flare, and is followed by an arcade of reconnected
875: loops that slowly decay.
876:
877: We have performed time resolved spectroscopy of these flares.
878: The coronal spectrum during
879: the flare can be represented with a 1T model, with quiescent state
880: taken into account as a frozen background contribution.
881: During the flares F1, F9 and F11 the emission measure
882: was increased by a factor of five or more. Such large
883: variation in emission measures were also seen during the
884: flares detected in AB Dor (Maggio et al. 2000). Both
885: the temperature and emission measure show the well-defined trends
886: i.e. the changes in the temperature and the emission measure are correlated
887: with the variations observed in the light curves during the flares.
888: Similar trends were also seen during the flares detected in the pre-main sequence stars
889: of the Orion nebula cluster (Favata et al. 2005) and star-forming complex
890: L1551 in Taurus (Giardino et al. 2006).
891: Reale, Peres \& Orlando (2001)
892: found that the height of emission measure distributions are variable during the different
893: phases of the solar flares, while its width and the peak temperature of the distribution undergo much smaller changes. The peak flare temperature
894: is found in the range of 20 - 60 MK for all the flares.
895: These values are intermediate between those of flares observed on active
896: stars like Algol ($T_{max} \sim $ 100-150 MK, Favata et al. 2000b) and
897: AB Dor ($T_{max} \sim$ 140-170 MK, Maggio et al. 2000) and those
898: found in flares on the dMe star AD Leo (range 20-50 MK,
899: Reale \& Micela 1998; Favata et al. 2000a), and on solar-type
900: Pleiades (20-40 MK, Briggs \& Pye 2003), and young stellar
901: objects ($T_{max} \sim$ 80-270 MK; Favata et al. 2001,2005).
902: For most of flares both emission measure and temperature
903: peaked simultaneously. However,
904: for few flares (F9, F11, F13), it appears that the temperature
905: was evolved before the emission measure.
906: Similar delay is often observed both in solar flare (Sylwester et al. 1993)
907: and in flares from the stars Algol(van Den Oord \& Mewe 1989),
908: EV Lac (Favata et al. 2000a),
909: AB Dor (Maggio et al. 2000), and YY Gem (Stelzer et al. 2002).
910: This is
911: probably due to an impulsive flare event , in which loop does not
912: reach equilibrium conditions, the density begins to decay later than
913: the temperature.
914:
915: We have modeled the flares using the
916: hydrodynamic model (see \S \ref{sec:model};
917: Reale et al. 1997) based on the decay phase of the flare.
918: The derived semiloop lengths for all flares are found to be in the range
919: of $0.6 - 8.0 \times 10^{10}$ cm.
920: Alternatively, Reale (2007)
921: derive the semiloop length from the rise phase and peak phase of the flare
922: as:
923: \begin{equation}
924: L_9^\prime \approx 3 \psi^2 T_{max}^{1/2} t_M
925: \label{eq:loop_rise}
926: \end{equation}
927:
928: \noindent
929: where $L_9^\prime$ is semiloop length in the unit of $10^9$ cm,
930: $\psi = T_{max}/T_M$, $T_{max}$ is maximum temperature
931: in the unit of $10^7$ K, $T_M$ is temperature at density maximum and
932: $t_M$ is time in the unit of $10^3$ s at which density maximum occurs.
933: The time of the maximum emission measure is a good proxy
934: for $T_M$. In case where both temperature and emission measure were
935: simultaneously peaked during the rise phase, only rise time was used
936: in the estimation of the semiloop length.
937: The semiloop lengths estimated
938: from this approach for the flares F4 ($L_9^\prime = 13.6\pm3.2$ cm), F6($L_9^\prime = 16.8\pm1.8$ cm),
939: F13($L_9^\prime = 6.6\pm1.4$ cm), F14($L_9^\prime = 11.4\pm3.1$ cm) and
940: F16($L_9^\prime = 20.8\pm5.8$ cm) are found to be
941: consistent with that of estimated from the decay phase analysis (see also Table 8).
942: The estimated loop length, $L_9^\prime$, for the flares
943: F2 ($9\pm2$ cm), F7($36\pm8$ cm) and F10 ($7\pm2$) are found to be less than that of determined from decay phase.
944: However, for the flare F9 of IM Vir the estimated loop length ($\L_9^\prime = 41\pm13$) is two times
945: more than that of estimated from decay phase. The inconsistency in the estimation of
946: the loop length from the two approaches is probably due to the involvement of
947: other different coronal loops during the decay of the flare or the heat pulse
948: triggering the flare is not a top-hat function.
949: This new analysis allows us
950: to derive the loop lengths of the flares for which either time resolved spectral
951: information is not available or decay phase is not observed.
952: These flares are F11 of V471 Tau, and F15 and F17 of EP Eri.
953: The estimated semiloop lengths for these flares F11, F15 and F17 are $6.2 \times 10^{10}$,
954: $2.1\times10^{10}$ cm and $3.8\times 10^9$ cm, respectively.
955: Therefore, the magnetic structures confining the plasma for all the observed flares
956: in G-K dwarfs are smaller than the star themselves, and
957: are not as large as observed in a single giant HR 9024 ($L=R_\star/2$; Testa et al. 2007)
958: and some pre-main-sequence analogues ($L >> R_\star$; Favata et al. 2005, Giardino et al.
959: 2006).
960:
961: For all the flares, the estimated maximum electron density under assumption of a totally ionised
962: hydrogen plasma is found in the order of $10^{10-11}$ cm$^{-3}$. This is compatible with
963: the values expected for the plasma in coronal condition (Landini et al. 1986).
964: To satisfy the energy balance relation for the flaring as a whole, the maximum
965: X-ray luminosity must be lower than the total energy rate ($H = E_H.V$; see \S
966: \ref{sec:model}) at the flare peak.
967: The rest of the input energy is used for thermal conduction, kinetic energy and
968: radiation at lower frequencies. For the flares F2 and F4 of V368 Cep
969: the maximum X-ray luminosity observed is about 34\% of $H$. Similarly, for the flares F7 of
970: XI Boo and F10 of V471 Tau the maximum X-ray luminosity is about 30\% of $H$.
971: However, for other flares (see Table 7) only 16 - 22 \% of $H$ is observed as a peak
972: luminosity.
973: These values are in agreement with those reported
974: for the solar flares, where the soft X-ray radiation only accounts for upto 20\% of total energy
975: (Wu et al. 1986). In comparison, the fraction of X-ray
976: radiation to the total energy has been found to be 15\% and 35\% for the M dwarfs Proxima Centauri and EV Lac, respectively (Reale et al.
977: 2004; Favata et al. 2000c).
978: Applying loop scaling law $V = 2\pi \beta^2 L^3$, and if the detected flares are produced
979: by a single loop, their aspect ratio ($\beta$) were estimated in the range of 0.1 to 0.3 for the
980: flares F2, F4, F7, F9, F13, F14 and F16 (see Table 8). Similar cross section was also observed for the
981: solar coronal loops, for which typical values of $\beta$ are in the range of 0.1 -0.3. If we assume
982: $\beta =0.1$, these flares occurred in the arcades, which are composed of 2 to 11 loops (see also Table 8).
983: However, for the flares F6 and F10 the value of $\beta$ (assuming single loop) is
984: estimated to be 0.51 and 0.37, respectively. In comparison to the
985: solar coronal loops, the cross section for these flare loops are very large.
986: It appears that the flares F6 and F10 occurred in the arcades that contain 26 and 14 loops, respectively.
987:
988: Therefore, we conclude that the observed flares in G-K
989: dwarfs are similar to the solar arcade flares, which are as strong as M dwarfs and are much smaller than the
990: flare observed in dMe star, giants and pre-main-sequence analogous.
991:
992: \section*{Acknowledgments}
993: We thank the reviewer of the paper for very useful comments and suggestions.
994: This research has made use
995: of data obtained from HEASARC, provided by the NASA
996: Goddard Space Flight Center.
997:
998:
999: \clearpage
1000: \input{ref}
1001: \newpage
1002: \input{tab1}
1003: \input{tab2}
1004: \input{tab3}
1005: \input{tab4}
1006: \input{tab5}
1007: \input{tab_Q}
1008: \input{tab6}
1009: \input{tab7}
1010: \input{fig}
1011: \end{document}
1012: