1: %\documentstyle[12pt,aasms4,flushrt]{article}

2: \documentstyle[emulateapj,apjfonts,flushrt]{article}

3: %

4: %       X-ray Detection of the Brown Dwarf \lp\ with \chandra

5: %       Rutledge, Basri, Martin, and Bildsten

6: %       Submitted April 28, 2000

7: %       Revised  May 17, 2000

8: %

9: %

10: \makeatletter

11: % ASTRO.STY is a modification of ASTRON.STY by Sake J. Hogeveen that

12: % makes it conform to Ap.J. Citation conventions.  Modified by

13: % Christopher B. Moore, February, 1993.  Modified by R. Rutledge Sept 1996

14: % ASTRON.BST is an adaptation of APALIKE.BST by Oren Patashnik and Suzan King.

15: % This style produces citations in the `author-year' format, which is widely

16: % used among astronomical journals.

17: % It supports four forms of citation: the \cite command produces:  (Author, year)

18: % in the text; the \cite* command only:  (year) .

19: % Other two forms are \citenp, which produces Author, year in the text

20: % (no parenthesis); the \citenp* command only: year.  This was added

21: % by R. Rutledge Sept 1996. 

22: 

23: % The long and short citation trickery is adapted from NAMED.BST by

24: % Peter F. Patel-Schneider.

25: 

26: \newlength{\bibhang}

27: \setlength{\bibhang}{1.4em}

28: 

29: \let\@internalcite\cite

30: \def\cite{\@ifstar{\citeyear}{\citefull}}

31: \def\cite{\let\@citeleft(\let\@citeright)%

32:     \@ifstar{\citeyear}{\citefull}}

33: \def\citenp{\let\@citeleft\relax\let\@citeright\relax

34:     \@ifstar{\citeyear}{\citefull}}

35: \def\citefull{\def\astroncite##1##2{##1~##2}\@internalcite}

36: \def\citeyear{\def\astroncite##1##2{##2}\@internalcite}

37: 

38: \def\@citex[#1]#2{\if@filesw\immediate\write\@auxout{\string\citation{#2}}\fi

39:   \def\@citea{}\@cite{\@for\@citeb:=#2\do

40:     {\@citea\def\@citea{; }\@ifundefined

41:        {b@\@citeb}{{\bf ?}\@warning

42:        {Citation `\@citeb' on page \thepage \space undefined}}%

43: {\csname b@\@citeb\endcsname}}}{#1}}

44: 

45: 

46: %\def\@cite#1#2{#1\if@tempswa , #2\fi}

47: \def\@cite#1#2{\@citeleft#1\if@tempswa , #2\fi\@citeright}

48: \def\@biblabel#1{}

49: 

50: \makeatother

51: 

52: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

53: \newcommand{\PSbox}[3]{\mbox{\rule{0in}{#3}\special{psfile=#1}\hspace{#2}}}

54: \newcommand{\FigNum}[1]{\unitlength 1pt \begin{picture}(55,10)(-400,35) 

55:                         \put(0,0){Figure #1}

56:                         \end{picture}}

57: \newcommand{\perval}[2]{{#1\mbox{$^{#2}$}}}

58: \newcommand{\msun}{$M_\odot$} 

59: \newcommand{\persec}{\perval{sec}{-1}\/}

60: \newcommand{\percm}{\mbox{$\cm^{-2}$}}

61: \newcommand{\peryear}{{yr$^{-1}$}}

62: \newcommand{\ppm}{\mbox{$\pm$}}

63: \newcommand{\cgsflux}{\erg~\percm~\persec}

64: \newcommand{\cgslum}{\erg~\persec}

65: \newcommand{\kmpersec}{\km~\persec}

66: \newcommand{\approxlt}{\mbox{$\lesssim$}}

67: \newcommand{\approxgt}{\mbox{$\gtrsim$}}

68: \def\etal{{et~al.}}

69: \def\eg{{e.g.}}

70: \def\ie{{i.e.}}

71: \def\ebv{{$E(B-V)$}}

72: \newcommand{\nh}{\mbox{$N_{\rm H}$}}

73: \newcommand{\nhtt}{\mbox{$N_{\rm H, 22}$}}

74: \def\chisqr{\mbox{$\chi^2$}}

75: \newcommand{\ud}[2]{\mbox{$^{+ #1}_{- #2}$}}

76: \newcommand{\supp}[1]{\mbox{~~$^{#1}$}}

77: \newcommand{\ee}[1]{\mbox{$10^{#1}$}}

78: \newcommand{\tee}[1]{\mbox{$\times 10^{#1}$}}

79: \newcommand{\tento}[1]{\mbox{$10^{#1}$}}

80: \newcommand{\keV}{\mbox{$\rm\,keV$}}

81: \newcommand{\MeV}{\mbox{$\rm\,MeV$}}

82: \newcommand{\cm}{\mbox{$\rm\,cm$}}

83: \newcommand{\km}{\mbox{$\rm\,km$}}

84: \newcommand{\second}{\mbox{$\rm\,s$}}

85: \newcommand{\ksec}{\mbox{$\rm\,ksec$}}

86: \newcommand{\yr}{\mbox{$\rm\,yr$}}

87: \newcommand{\GramPerCc}{\mbox{$\rm\,g\,cm^{-3}$}}

88: \newcommand{\K}{\mbox{$\rm\,K$}}

89: \newcommand{\erg}{\mbox{$\rm\,erg$}\/}

90: \newcommand{\gauss}{\mbox{$\rm\,G$}}

91: \newcommand{\Qnuc}{Q_{\rm nuc}}

92: \newcommand{\Msun}{\mbox{$\,M_\odot$}}

93: \newcommand{\mJy}{\mbox{$\rm\,\mu Jy$}}

94: \newcommand{\axaf}{{\em Chandra\/}}

95: \newcommand{\chandra}{{\em Chandra\/}}

96: \newcommand{\rosat}{{\em ROSAT\/}}

97: \newcommand{\asca}{{\em ASCA\/}}

98: \newcommand{\rxte}{{\em RXTE\/}}

99: \newcommand{\xmm}{{\em XMM\/}}

100: \newcommand{\beppo}{{\em BeppoSAX\/}}

101: \newcommand{\kpc}{\mbox{$\rm\,kpc$}}

102: \newcommand{\Aem}{\mbox{$A_{\rm em}$}}

103: \newcommand{\Teff}{T_{\rm eff}}

104: \newcommand{\kteff}{$kT_{\rm eff}$}

105: \newcommand{\lp}{{LP~944-20}}

106: \newcommand{\usno}{{USNO-A2}}

107: \newcommand{\hal}{{H$\alpha$}}

108: \newcommand{\sig}{{$\sigma$}}

109: \newcommand\lsun{L$_\odot$}             % solar luminosity

110: \newcommand\lxlbol{$L_{X}$/$L_{\rm bol}$}  

111: \newcommand\lx{$L_{X}$}  

112: 

113: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%the document begins%%%%%%

114: %\clubpenalty=10000

115: %\widowpenalty=10000

116: \received{April 28, 2000}

117: \revised{May 17, 2000}

118: % Manuscript # 

119: %\accepted{}

120: \slugcomment{\it ApJ Letters, Submitted April 28, 2000; Revised May 17, 2000}

121: \begin{document}

122: 

123: \title{\chandra \ Detection of an X-ray Flare from the Brown Dwarf \lp}

124: \author{Robert E. Rutledge}

125: \affil{Division of Physics, Mathematics and Astronomy}

126: \affil{MS 220-47, California Institute of Technology, Pasadena, CA

127: 91107\\ {\em rutledge@srl.caltech.edu} }

128: \smallskip

129: \author{Gibor Basri}

130: \affil{Department of Astronomy, University of California at Berkeley}

131: \affil{Berkeley, CA 94720-3411\\ {\em basri@soleil.berkeley.edu}}

132: 

133: \author{Eduardo L. Mart\'\i n}

134: \affil{Division of Geological and Planetary Sciences, California

135: Institute of Technology}

136: \affil{MS 150-21,

137: Pasadena, CA 91125\\ {\em ege@gps.caltech.edu}}

138: 

139: \centerline{\sc and}

140: \author{Lars Bildsten}

141: \affil{Institute for Theoretical Physics and Department of Physics}

142: \affil{Kohn Hall, University of California, Santa Barbara, CA 93106\\ {\em bildsten@itp.ucsb.edu}}

143: 

144: 

145: \begin{abstract}

146: 

147:    We have detected a bright X-ray flare from the nearby ($d=$5.0 pc)

148: brown dwarf \lp\ with the \chandra/ACIS-S. This is an old (500 Myr),

149: rapidly rotating, lithium-bearing M9 object, with a bolometric

150: luminosity of $\approx 6$\tee{29} \cgslum. It was only detected by

151: {\it Chandra} during an X-ray flare of duration 1-2 hours near the end

152: of a 12.1 hour observation.  The peak X-ray luminosity was

153: 1.2\ud{0.5}{0.3}\tee{26} \cgslum\ in the brightest $\approx 550$

154: seconds, corresponding to $L_X/L_{\rm bol}\approx 2\times 10^{-4}$.  A

155: total of 2\tee{29} ergs was released during the 43,773 sec

156: observation, giving a time-averaged $L_X/L_{\rm bol}\approx 7\times

157: 10^{-6}$.  \lp\ was not detected before the flare, with a 3$\sigma$

158: upper limit on the emission at $L_X/L_{\rm bol}< 2\times 10^{-6}$

159: ($L_X<$1\tee{24} \cgslum).  This is faint for a rapidly rotating

160: late-type star, and establishes a record lower limit to the quiescent

161: flux about an order of magnitude below the flux limit (and a factor of

162: 5 below the \lxlbol\ limit) placed on quiescent X-ray emission from

163: the M8 dwarf VB 10. The inferred flaring duty cycle is comparable to

164: that measured via variable H$\alpha$ emission for other late M-type,

165: fully convective stars.

166: 

167: \end{abstract}

168: 

169: 

170: \keywords{ stars: coronae -- stars: low-mass, brown dwarfs -- 

171:    stars: individual(\lp)

172: }

173: 

174: \section{Introduction}

175: 

176: \lp\ (=BRI 0337$-$3535) is an isolated, non-accreting, brown dwarf

177: identified through its Li abundance and low luminosity

178: \cite{tinney98discovery}.  Its parallactic distance and bolometric

179: luminosity are 5.0\ppm0.1 pc and 6\tee{29}\cgslum

180: \cite{tinney96}. Tinney ~\cite*{tinney98discovery} infers an age of

181: about $\tau=$500 Myr, which implies that it is a fully collapsed

182: object (see \citenp{feigelson99} for a recent review of protostellar

183: evolution).  This requires that any observed coronal activity not be

184: due to accretion, as may power younger brown dwarfs ($\tau<$10 Myr) in

185: analogy with T-Tauri stars.  We report here on a bright 1-2 hour X-ray

186: flare detected with \chandra \ during a 12 hour observation, and a

187: strong upper limit on persistent X-ray emission. We attribute the

188: energy release in the flare to transient magnetic activity on this

189: fully convective, rapidly rotating star.

190: 

191: 

192: Stars with masses $M>0.3 M_\odot$ have an outer convective zone and

193: an interior radiative region that need not be rotating at the same

194: rate. A poloidal magnetic field in the convective layers will be

195: stretched and amplified into strong toroidal fields when it is dragged

196: by convective overshoot (see \citenp{weiss94}) into the radial shear

197: in rotation that resides at the boundary (in and near the so-called

198: "tachocline"; \citenp{spiegel92}).  In these cases, rapid rotation is

199: associated with enhanced coronal activity. The activity level

200: correlates with the Rossby number -- the ratio of the stellar rotation

201: period to the convective turnover time \cite{Noyes84}. For Rossby

202: numbers between 10 and 0.1, coronal activity (as measured by \lxlbol)

203: increases with decreasing Rossby number (that is, with more rapid

204: rotation), ``saturating'' at about \lxlbol $\sim 10^{-3}$ for Rossby

205: numbers of 0.1 to 0.01 \cite{randich98}.

206: 

207:  For less massive stars and young brown dwarfs, the energy is

208: transported throughout the star by convection; no radiative core is

209: present. For this reason, it has been supposed that the activity and

210: its dependence on rotation might change near the spectral type where

211: the radiative layer disappears (about M5.5; see \citenp{giampapa96}

212: for a physical overview).  However, a search for this effect found no

213: evidence for a change in the saturated value of \lxlbol $\sim

214: 10^{-3}$, down to spectral types as late as M7, well into the region

215: of fully convective stars \cite{fleming93}.  This implied no dramatic

216: change in the rotation-activity dependence in fully convective stars.

217: 

218: The first hint that activity might be decreasing in the very late M

219: dwarfs was observed from the rapidly rotating ($v \sin i = 40$ \km\

220: \persec) M9.5 star BRI 0021-0214, for which a strong upper-limit on

221: \hal\ emission indicated a substantially lower persistent coronal

222: activity than expected from a rapidly rotating, fully convective star

223: \cite{basri95}.  On the other hand, Reid \etal\ \cite*{reid99dec}

224: observed a strong \hal\ flare from BRI 0021-0214; this indicated that

225: activity was indeed present, and that the star has outbursts no more

226: than 7\% of the time. Liebert \etal\ ~\cite*{liebert99} reported a

227: bright \hal\ flare from 2MASSW J0149090+295613, an M9.5 V star which

228: is otherwise quiescent. Intensity variability from brown dwarfs has

229: been searched for, with limited success.  Bailer-Jones \& Mundt

230: \cite*{bailerjones99} found no infra-red variability in three Pleiades

231: brown dwarfs, with a limit of $\delta I<$0.05 mag on timescales

232: between 25 min-27 hrs.  A search for ``weather'' in two brown dwarfs

233: (\lp\ and DENIS-P J1228-1547) produced a claim of evidence of

234: variability in one, at the 2.3$\sigma$ level \cite{tinney99}.

235: 

236: % What about Briceno 1998 -- an X-ray object is a BD candidate. 

237: Previous X-ray observations of brown dwarfs and brown dwarf candidates

238: have produced detections of persistent emission from several

239: \cite{neuhauser98,neuhauser99a}, but these are all young objects

240: (\approxlt 10 My), still in the process of proto-stellar collapse, and

241: so are actively accreting; neither are they as cool or faint as older

242: brown dwarfs such as \lp.  The activity observed from these collapsing

243: brown dwarfs is analogous to that in the also-young T-Tauri stars --

244: powered by accretion and collapse -- rather than that in fully

245: collapsed, older M dwarfs.  There are only a few observations of X-ray

246: emission in older late M dwarfs and brown dwarfs. Giampapa

247: \etal~\cite*{giampapa96} reported on a ROSAT detection from VB 8

248: (spectral type M7) with a time-averaged \lxlbol =1.6\tee{-3}. Fleming

249: \etal~\cite*{fleming00} detected the M8 star VB 10 during a flare at

250: \lxlbol =5\tee{-4}. No quiescent X-ray emission was detected, limiting

251: it to \lxlbol $<$\ee{-5}. In the only previous observations of \lp ,

252: an upper limit of \lxlbol $<$7\tee{-5} was found

253: \cite{neuhauser99a}. Hence, as with \hal\ observations, the X-ray

254: behavior of late M stars tends toward flaring activity and an absence

255: of persistent activity at the X-ray detection limits of present

256: instrumentation.  Our work confirms this tendency.

257: 

258: We report here on our \chandra \ observation of an X-ray flare from

259: the brown dwarf \lp.  In \S~\ref{sec:obs}, we describe the

260: observation, analysis and detected flux level. In \S 3, we interpret

261: the light curve of the flare.  We conclude in \S~\ref{sec:con} with a

262: brief summary of our results and a comparison to other work. 

263: 

264: \section{Chandra Observation and Analysis}

265: \label{sec:obs}

266: 

267: 

268: The observation (\chandra\ sequence number 200049) occurred on 15 Dec

269: 1999 00:05:50-13:03:05 UTC, for 43,773 seconds. \lp\ was targeted at

270: the nominal aimpoint for the ACIS-S3 chip (backside illuminated) in a

271: faint imaging mode, with 3.2 sec time resolution.  Analysis of the

272: ACIS-S3 chip countrate during the observation showed no evidence of

273: the background flares that sometimes appear (ACIS background

274: calibration memo

275: \footnote{http://asc.harvard.edu/cal/Links/Acis/acis/Cal\_prods/bkgrnd/11\_18/bg181199.html}),

276: on timescales longer than 50 sec, greater than a factor of $\sim$few

277: (background variability is discussed in more detail in

278: Sec.~\ref{sec:var}). The standard analysis data products from \chandra\

279: found 117 X-ray objects in the field of view.

280: 

281:   We now describe how we have used the observations to measure the 

282: X-ray emission from \lp. 

283: 

284: \subsection{Astrometry and Source Identification}

285: 

286: The X-ray source closest to the ACIS-S detector aimpoint was offset

287: from the detector aimpoint by $\delta$RA=$-2.2$\arcsec and

288: $\delta$dec=$-$8.3\arcsec\ according to the standard product

289: astrometry.  This offset is consistent with known systematic

290: uncertainties in the \chandra\ pre-processing analysis astrometry, of

291: the version which produced the astrometry for this observation.  We

292: performed astrometry using the absolute positions of objects in \usno.

293: We used the \chandra\ Interactive Analysis of Observations (CIAO) V1.0

294: software tool {\em celldetect} to find X-ray point source relative

295: positions using counts detected in PHA channels 10-400 (correspondig

296: roughly to 0.1-4.0 keV). We found 39 X-ray sources within 10\arcmin\

297: of the aimpoint, with relative astrometry accurate to between

298: 0.03-0.18\arcsec.  We extracted all optical source positions (from the

299: USNO-A2 catalog) within 15\arcsec\ of the X-ray source positions. We

300: did the same for 20 background fields for each X-ray source (a total

301: of 780 fields), offset by increments of 15\arcsec\ from each X-ray

302: source position, to find the USNO-A2 source density in the region of

303: $\rho_{\rm USNO-A2}=(5.2\ppm0.3)\tee{-4}$ \perval{arcsec}{-2}.  Of the

304: 39 \chandra\ X-ray sources, 12 had a \usno\ object within 15\arcsec.

305: Of these 12, seven were within 1.0\arcsec\ of the offset

306: $\delta$RA=$-2.2$\arcsec and $\delta$dec=$-$8.3\arcsec.  The

307: probability of n=7 (of m=12) optical point sources being found within

308: $r=1.0$\arcsec\ of a previously selected position is $P(<r)=

309: \frac{m!}{(m-n)!n!}(1 - \exp(-\pi \rho_{\rm USNO-A2}

310: r^2))^n=$2\tee{-17}.  Thus, the clustered \usno\ sources are -- taken

311: together -- likely to be the optical counterparts for their

312: corresponding X-ray sources, useful for astrometry.

313: 

314: 

315: To calculate the systematic shift in RA and Dec, we adopted the

316: relative positional uncertainties for the X-ray sources found by {\em

317: celldetect}, and adopted absolute positional uncertainties of

318: 0.25\arcsec\ for USNO-A2 sources.  

319: 

320: The astrometric correction is ($\delta$RA=$+2.2$\arcsec,

321: $\delta$dec=8.4\arcsec), with an uncertainty of \ppm0.1 \arcsec.  The

322: astrometricly corrected X-ray position of the aimpoint source is then

323: RA=03h39m35.16s, dec=$-$35d25h44.0s \ppm 0.1\arcsec\ (1$\sigma$;

324: J2000, epoch 1999.95).  The positional difference between the aimpoint

325: source and the optical position of \lp\ at this epoch \cite{tinney96}

326: is $\delta$RA=$-0.67\pm0.23$\arcsec\ and

327: $\delta$dec=0.15\ppm0.23\arcsec, which is consistent at the 3$\sigma$

328: level.

329: 

330: The likelihood  of a chance alignment of  a serendipitous X-ray source

331: within $r$=1\arcsec\ of  an arbitrary position is  $P(<r)=0.0002$, for

332: $\rho_{\rm X-ray}=$7.2\tee{-5}  \perval{arcsec}{-2}, found from  the 18

333: sources on the ACIS-S3 chip.  We therefore  identify this X-ray source

334: with the brown dwarf \lp\ with 99.98\%  confidence on the basis of the

335: positional coincidence.

336: 

337: \subsection{The Flare from \lp} 

338: \label{sec:var} 

339: 

340: First, we estimate the background countrate. We obtained background

341: counts from an annulus centered on the source, with inner- and

342: outer-radius of 4 and 140 pixels.  We excluded data within 10 pixels

343: of two X-ray sources localized by {\em celldetect} to be within this

344: annulus.  The total average background countrate was

345: (3.99\ppm0.04)\tee{-6} counts \persec \perval{pixel}{-1}; in the

346: limited PHA channel range of 10-400 (nominally 0.1-4.0 keV), the

347: background countrate was (1.0\ppm0.02)\tee{-6} counts \persec

348: \perval{pixel}{-1}.  For nearly all the analyses we present herein,

349: this level of background is negligible.

350: 

351: We extracted counts from a circle about \lp\ 2.0 pixels in radius

352: (0.98\arcsec-- the 90\% enclosed energy radius at 1.5 keV, on axis).

353: The total number of expected background counts in the source region is

354: 2.27\ppm0.02 counts (0.1-10.0 keV), and 0.57\ppm0.01 counts (0.1-4.0

355: keV).  We produced a (0.1-4.0 keV) light curve of 552 sec time

356: resolution (Fig.~\ref{fig:sub4varicomp}).  Of the 19 detected counts

357: in the source region, 15 were detected during a 2760 sec period

358: beginning at 1999 Dec 15 09:41:25 (UTC), which we arbitrarily define

359: as the ``flare period''.  Of these, 7 were detected in a single 552

360: sec bin.  In 80 such time bins, with an average of 19/80=0.238

361: counts/bin, the probability of randomly finding 7 counts in one of 80

362: bins (assuming a constant countrate) is $\approx 7\tee{-9}$.

363: 

364: The ACIS-S-BI chips suffer from short-term increases in background

365: countrate; these can increase the chip background countrate by up to

366: factors of 100 across the entire chip, and in all PHA channels on

367: timescales of seconds to minutes.  The ratio of the countrates during

368: and outside the flare period in the source region is 18-200 ($3\sigma$

369: range, assuming Poisson statistics), while in the background region it

370: is 0.85-1.07 (3$\sigma$, assuming Gaussian statistics).  The flare

371: cannot be from a variation in the background and must be a change in

372: the X-ray emission of \lp.

373: 

374: \subsection{Spectral Analysis}

375: 

376: There were 15 counts (0.1-10 keV) in the 2.0 pixel (=0.98\arcsec)

377: radius about \lp\ during the 2760.0 seconds of the flare.  We expect

378: 0.14 background counts, which we neglect for this spectral analysis.

379: 

380: We performed a spectral analysis, binning the data into two bins which

381: contained 7 and 8 counts, with a third bin of PHA channels 95-1024

382: which contained no counts.  We fit the resulting spectrum with an

383: assumed Raymond-Smith plasma model, implemented in XSPEC \cite{xspec};

384: the best fit ($\chi^2/\nu$=0.003, for $\nu$=1 degree of freedom) found

385: $kT$=0.26\ud{0.1}{0.07} \ {\rm keV}(90\% confidence), and a

386: time-averaged flux during the 2760 seconds of (1.5\ppm0.4)\tee{-14}

387: \cgsflux (\lx=[4.5\ppm1.2]\tee{25} \cgslum; 0.1-10 keV), which

388: corresponds to a flux conversion factor of 1 count = 2.8\tee{-12}

389: \erg\percm \ (0.1-4.0 keV) for this spectrum.  This reduced chi-square

390: value is small because the high energy bin is essentially unimportant,

391: as there are, in the best fit model, $\sim$0 counts in this bin, and

392: thus the model can be thought of as completely determined, with only

393: two bins and two unknowns, although the high energy bin does provide a

394: constraint on flatter spectra, and thus on $kT$ and flux

395: normalization.

396: 

397: The uncertainty in this conversion factor depends both on the

398: uncertainty in the Raymond-Smith spectral parameters, and on the

399: uncertainty in the intrinsic spectrum.  The best fit power-law

400: spectrum ($\alpha_{photon}$=2.6), however, is acceptable only at 6\%

401: confidence. The best-fit black body spectrum (kT=0.17 keV) gives a

402: flux 10\% lower than the best-fit Raymond Smith spectrum.  We estimate

403: this as the level of spectral uncertainty in the flux conversion

404: factor. 

405: 

406: In Table~\ref{tab:flux}, we list the \lxlbol\ values for the full

407: observational period, pre-flare period, the flare period, and the peak

408: 550 sec bin of the flare period.  The uncertainties in these values

409: are Poisson (counting statistics), plus spectral uncertainty

410: ($\sim$10\%).  The $L_X$ were obtained from the detected countrates,

411: corrected for background and the Enclosed Energy fraction appropriate

412: for our source region ($0.90$), and using the above flux conversion

413: factor.

414: 

415: The total source fluence (including the flare) during this observation

416: is 5.7\tee{-11} ergs \percm, corresponding to a released energy of

417: 1.7\tee{29} ergs, and a time-averaged luminosity of

418: $L_X$=4\tee{24}\cgslum.  The time-averaged \lxlbol\ is then 7\tee{-6}.

419: The peak luminosity during the flare was 1.2\ud{0.5}{0.3}\tee{26}

420: \cgslum (including only the counting statistical uncertainty, not the

421: 10\% systematic uncertainty in the counts-energy conversion factor),

422: corresponding to $L_X/L_{\rm bol}\approx 2\times 10^{-4}$.

423: 

424: Before the flaring period, there is 1 count (0.1-4 keV) in 34535 sec,

425: where 0.45 background counts are expected (probability p=0.36 -- consistent with

426: background).  The 3$\sigma$ upper-limit on the pre-flare flux is

427: $<$4.5\tee{-16} \cgsflux\ ($L_X<$1\tee{24} \cgslum;

428: \lxlbol$<$2\tee{-6}).  This compares to the VB~10 quiescent coronal

429: X-ray limits from Fleming \etal\ \cite*{fleming00} of

430: $L_X<$1.7\tee{25}\cgslum, which is an order of magnitude greater, and

431: \lxlbol$<$1.0\tee{-5}, which is a factor of $\times$ 5 greater than

432: we find here for \lp.

433: 

434: After the flare period, there are 3 counts in 6478 sec (0.1-4 keV)

435: where 0.084 are expected (p=0.0001).  Thus, there are significant

436: counts after the flare period, which indicates that the flare event

437: continues beyond the flaring period we defined.

438: 

439: \section{Constraints on the Form of the Lightcurve}

440: 

441: Due to the low number of counts, it is difficult to measure the

442: lightcurve parameters of this flare in a model-independent way.  We

443: therefore imposed a particular model and measured the resulting

444: parameters.  We imposed a model of an instantaneous rise/exponential

445: decay flare, set above a constant background countrate, to compare

446: with the 0.1-4 keV energy band data (19 counts).  We assumed the

447: background countrate from our background region.  We asserted that the

448: instantaneous rise takes place at the time of the first detected

449: photon in the vicinity of the flare (32969 seconds into the

450: observation).  The resulting lightcurve (background+flare) must

451: produce between 11-27 counts in total (90\% confidence region of the

452: total number of counts observed).  The distribution of counts in time

453: should match that observed over the full observation period, using a

454: Kolmogorov-Smirnov test \cite{press}, that the data and the model be

455: consistent at the 10\% level.  We tested a grid of peak countrate

456: ($I_p$) between \ee{-4} and 1 counts~\persec, in units of \ee{-5}

457: counts~\persec, and exponential decay timescale ($\tau$) between 1

458: and 2\tee{4} in units of 100 sec.  Based on this model and approach,

459: we find a 90\% confidence region for $\tau$=5400\ppm700 s, and

460: $I_p$=0.002-0.006 c \persec.  When we relax the constraint on start

461: time, permitting the flare to begin at any time between 30\tee{3} and

462: 40\tee{3} sec after observation start in grid units of 100 sec, the

463: 90\% confidence interval of $\tau$ is 500-7100 s, and of $I_p$ is

464: 0.0018-0.054 counts~\persec.

465: 

466: This brown dwarf is known to be rapidly rotating, with a measured

467: $V\sin i\approx 28 \ {\rm km \ s^{-1}}$ \cite{tinney98spec}.  For the

468: reported bolometric luminosity of $L_{\rm Bol}\approx

469: (5.5\ppm0.4)\times 10^{29}$\cgslum\ \cite{tinney98discovery} and

470: effective temperature of 2500\ppm100 K \cite{basri00}, the inferred

471: radius is $R\approx (4.5\pm0.4)\times 10^{9} \ {\rm cm}$, which is

472: 35\ppm10\% smaller than the $R=0.1R_\odot$ expected for an object of

473: this mass and age \cite{burrows97}. The expected radius could be

474: derived if the temperature were lowered to 2300K and the bolometric

475: luminosity increased to $8\times 10^{29}$\cgslum. Such changes are

476: within the observational uncertainties.  Presuming a radius of $R=0.1

477: R_\odot$, the rotation period is $4.4 \ {\rm hr} \sin i$.  One might

478: interpret the X-ray lightcurve as a localized flare that rotates out

479: of view for a period and then re-appears. The implied rotation rate of

480: $\approx$ 1 hour does not require an especially unlikely value of $\sin

481: i$.  However, such an interpretation is by no means required by the

482: data to explain the variations, as these are entirely consistent with

483: the uncertainty in the flare time evolution and Poisson counting

484: statistics.

485: 

486: \section{Discussion and Conclusions}

487: \label{sec:con}

488: 

489:  We have detected an X-ray source which we identify with the brown

490: dwarf \lp, with 99.98\% confidence, based on positional

491: coincidence. The detection was during a flare, with a peak X-ray

492: luminosity of $L_X=$1\ud{0.5}{0.3}\tee{26} \cgslum, or $L_X/L_{\rm

493: bol}\approx 2\times 10^{-4}$.  This flare's peak luminosity is a

494: factor of ten below that of an X-ray flare detected from the

495: comparably luminous M8 star VB~10 ($L_{\rm Bol}\approx 1.7 \times

496: 10^{30} \ {\rm erg \ s^{-1}}$; \citenp{fleming00}). In VB~10, the peak

497: X-ray luminosity exceeded  $10^{-3} L_{\rm Bol}$. 

498: 

499: We have analysed the X-ray light curve with a simple sharp rise plus

500: exponential decay model; this indicates a decay time $5400\pm700 \

501: {\rm s}$ when we specify the flare start-time, and in the range of

502: 500-7100 s when we do not specify the flare start time. This implies

503: \lp\ is flaring $\sim$10\% of the time, comparable to the amount of

504: \hal\ flaring among the coolest M dwarfs \cite{gizis00}, although the

505: uncertainty on our value is large.  The light-curve is consistent with

506: rotational modulation, but the counting statistics are too poor to

507: conclude that it is required, and discovery of such modulation must

508: wait for higher signal-to-noise data. A previous attempt to detect

509: this source in X-rays \cite{neuhauser99a} placed an upper-limit on the

510: time-averaged luminosity ($L_X< 4\times 10^{25} \ {\rm erg \ s^{-1}}$)

511: during a 66 ksec \rosat/PSPC observation and a 220 ksec \rosat/HRI

512: observation. These limits are above the time-averaged luminosity we

513: detect here ($L_X\approx 4\times 10^{24} \ {\rm erg \ s^{-1}}$), and

514: are therefore consistent with our results.  X-ray variability

515: was not discussed by these previous works; however, a cursory review

516: of 220~ksec of ROSAT/HRI data revealed no variability on timescales of

517: 500 sec; the detection sensitivity (5 counts in a 500 sec bin, energy

518: conversion factor of 3.2\tee{-11} \cgsflux\ \perval{count}{-1}), is

519: about a factor of 10 above the present peak-flare detection.

520: 

521: The time-averaged luminosity ($L_X$=4\tee{24}\cgslum) is a factor of

522: $10^4$ below the previous detections of brown dwarfs and brown dwarf

523: candidates; the time-averaged ratio \lxlbol(=7\tee{-6}) is factor of

524: $10^2$ below this ratio \cite{neuhauser98,neuhauser99a}.  This might

525: be due to different energy production mechanisms; the previously

526: detected brown dwarfs are young compared with \lp\ ($\sim$1 My vs. 500

527: My) and are still forming, while \lp\ is isolated and undergoing slow

528: gravitational contraction.  The persistent X-ray emission of the very

529: young brown-dwarfs in open clusters is more analogous to the pre-main

530: sequence T-Tauri stars -- where convection is stronger, the

531: atmospheres are less neutral, and accretion may play a role -- while

532: the flaring X-ray emission of \lp\ is more analogous to the flaring

533: emission of late-type main-sequence stars, such as VB~8 and VB~10.

534: 

535: We estimate (order of magnitude) an eddy turnover time of $\sim$1 year

536: for \lp, yielding a Rossby number of $R_0 \sim$5\tee{-4}.  As

537: discussed earlier, this is below the Rossby number limit at which the

538: X-rays are observed to be "saturated" (at \lxlbol$\sim$\ee{-3}). That

539: saturation level is much greater than our quiescent X-ray flux upper

540: limit.  Either the "supersaturation" suggested by Randich has strongly

541: set in, or the connection between Rossby number and observed activity

542: is no longer relevant.  It is possible that the rapid rotation

543: suppresses persistent coronal activity, either by forcing the magnetic

544: field into a more organized form or by suppressing the turbulent

545: dynamo.  It is also possible that the neutrality of a cool photosphere

546: quenches coupling between atmospheric motions and the magnetic field,

547: which forces the field into dissipative configurations (as pointed out

548: by \citenp{fleming00}).  In any case, the detected flare requires that

549: a magnetic field be present on LP 944-20, and that at least

550: occasionally it is forced into a dissipative configuration high in the

551: atmosphere.

552: 

553:    Our results confirm the impression from previous studies of \hal\

554: that stellar activity is dying at the bottom of the main sequence, at

555: least in the form that it has in more massive late-type stars.  We

556: have pushed the limits on quiescent coronal emission levels to new

557: lows for fully convective objects. We have also helped confirm that

558: such objects do apparently still have magnetic fields and an ability

559: to flare about 10\% of the time.

560: 

561: \acknowledgements

562: 

563: The authors are grateful to the \chandra\ Observatory team for

564: producing this exquisite observatory. We are grateful to Andrew

565: Cumming for his critical reading of an early version of this paper,

566: and to the referee Tom Fleming, for comments which dramatically

567: improved the paper's readability.  We thank Daniel Holz for

568: encouraging us to consider the rotational modulation of the

569: lightcurve. This research was supported by NASA Grant No. GO0-1009X

570: and the National Science Foundation under Grant No.  PHY94-07194.

571: L. B. is a Cottrell Scholar of the Research Corporation.

572: 

573: 

574: %\bibliographystyle{astro} \bibliography{lp,complete}

575: 

576: \begin{thebibliography}{}

577: 

578: \bibitem[\protect\astroncite{Arnaud}{1996}]{xspec}

579: Arnaud, K.~A., 1996,

580: \newblock in G. Jacoby \& J. Barnes (eds.), {\em Astronomical Data Analysis

581:   Software and Systems V.}, Vol. 101, p.~17, ASP Conf. Series

582: 

583: \bibitem[\protect\astroncite{{Bailer-Jones} \& {Mundt}}{1999}]{bailerjones99}

584: {Bailer-Jones}, C. A.~L. \& {Mundt}, R., 1999,

585: \newblock {\em \aap} { 348}, 800

586: 

587: \bibitem[\protect\astroncite{{Basri} \& {Marcy}}{1995}]{basri95}

588: {Basri}, G. \& {Marcy}, G.~W., 1995,

589: \newblock {\em \aj} { 109}, 762

590: 

591: \bibitem[\protect\astroncite{{Basri} {\rm et~al.\/}}{2000}]{basri00}

592: {Basri}, G., {Mohanty}, S., {Allard}, F., {Hauschildt}, P.~H., {Delfosse}, X.,

593:   {Mart\'\i n}, E.~L., {Forveille}, T., \& {Bertrand}, G., 2000,

594: \newblock {\em \apj},

595: \newblock in press, astroph/0003033

596: 

597: \bibitem[\protect\astroncite{{Burrows} {\rm et~al.\/}}{1997}]{burrows97}

598: {Burrows}, A., {Marley}, M., {Hubbard}, W.~B., {Lunine}, J.~I., {Guillot}, T.,

599:   {Saumon}, D., {Freedman}, R., {Sudarsky}, D., \& {Sharp}, C., 1997,

600: \newblock {\em \apj} { 491}, 856

601: 

602: \bibitem[\protect\astroncite{{Feigelson} \& {Montmerle}}{1999}]{feigelson99}

603: {Feigelson}, E.~D. \& {Montmerle}, T., 1999,

604: \newblock {\em \araa} { 37}, 363

605: 

606: \bibitem[\protect\astroncite{{Fleming} {\rm et~al.\/}}{2000}]{fleming00}

607: {Fleming}, T.~A., {Giampapa}, M.~S., \& {Schmitt}, J. . H. M.~M., 2000,

608: \newblock {\em \apj} { 533}, 372

609: 

610: \bibitem[\protect\astroncite{{Fleming} {\rm et~al.\/}}{1993}]{fleming93}

611: {Fleming}, T.~A., {Giampapa}, M.~S., {Schmitt}, J. H. M.~M., \& {Bookbinder},

612:   J.~A., 1993,

613: \newblock {\em \apj} { 410}, 387

614: 

615: \bibitem[\protect\astroncite{{Giampapa} {\rm et~al.\/}}{1996}]{giampapa96}

616: {Giampapa}, M.~S., {Rosner}, R., {Kashyap}, V., {Fleming}, T.~A., {Schmitt}, J.

617:   H. M.~M., \& {Bookbinder}, J.~A., 1996,

618: \newblock {\em \apj} { 463}, 707

619: 

620: \bibitem[\protect\astroncite{Gizis {\rm et~al.\/}}{2000}]{gizis00}

621: Gizis, J.~E., Monet, D.~G., Reid, N.~I., Kirkpatrick, J.~D., Liebert, J., \&

622:   Williams, R.~J., 2000,

623: \newblock {\em \aj},

624: \newblock in press, astro-ph/0004361

625: 

626: \bibitem[\protect\astroncite{{Liebert} {\rm et~al.\/}}{1999}]{liebert99}

627: {Liebert}, J., {Kirkpatrick}, J.~D., {Reid}, I.~N., \& {Fisher}, M.~D., 1999,

628: \newblock {\em \apj} { 519}, 345

629: 

630: \bibitem[\protect\astroncite{{Neuh\"auser} {\rm et~al.\/}}{1999}]{neuhauser99a}

631: {Neuh\"auser}, R., {Brice\~no}, C., {Comer\'on}, F., {Hearty}, T., {Mart\'\i

632:   n}, E.~L., {Schmitt}, J. H. M.~M., {Stelzer}, B., {Supper}, R., {Voges}, W.,

633:   \& {Zinnecker}, H., 1999,

634: \newblock {\em \aap} { 343}, 883

635: 

636: \bibitem[\protect\astroncite{{Neuh\"auser} \& {Comer\'on}}{1998}]{neuhauser98}

637: {Neuh\"auser}, R. \& {Comer\'on}, F., 1998,

638: \newblock {\em Science} { 282}, 83

639: 

640: \bibitem[\protect\astroncite{{Noyes} {\rm et~al.\/}}{1984}]{Noyes84}

641: {Noyes}, R.~W., {Hartmann}, L.~W., {Baliunas}, S.~L., {Duncan}, D.~K., \&

642:   {Vaughan}, A.~H., 1984,

643: \newblock {\em \apj} { 279}, 763

644: 

645: \bibitem[\protect\astroncite{Press {\rm et~al.\/}}{1995}]{press}

646: Press, W., Flannery, B., Teukolsky, S., \& Vetterling, W., 1995,

647: \newblock {\em Numerical Recipies in C},

648: \newblock Cambridge University Press

649: 

650: \bibitem[\protect\astroncite{Randich}{1997}]{randich98}

651: Randich, S., 1997,

652: \newblock in R.~A. Donahue \& J.~A. Bookbinder (eds.), {\em Cool Stars, Stellar

653:   Systems and the Sun: Tenth Cambridge Workshop}, No. 154 in Astronomical

654:   Society of the Pacific Conference Series, p. 501, Astronomical Society of the

655:   Pacific

656: 

657: \bibitem[\protect\astroncite{{Reid} {\rm et~al.\/}}{1999}]{reid99dec}

658: {Reid}, I.~N., {Kirkpatrick}, J.~D., {Gizis}, J.~E., \& {Liebert}, J., 1999,

659: \newblock {\em \apjl} { 527}, L105

660: 

661: \bibitem[\protect\astroncite{{Spiegel} \& {Zahn}}{1992}]{spiegel92}

662: {Spiegel}, E.~A. \& {Zahn}, J.~., 1992,

663: \newblock {\em \aap} { 265}, 106

664: 

665: \bibitem[\protect\astroncite{{Tinney}}{1996}]{tinney96}

666: {Tinney}, C.~G., 1996,

667: \newblock {\em \mnras} { 281}, 644

668: 

669: \bibitem[\protect\astroncite{{Tinney}}{1998}]{tinney98discovery}

670: {Tinney}, C.~G., 1998,

671: \newblock {\em \mnras} { 296}, L42

672: 

673: \bibitem[\protect\astroncite{{Tinney} \& {Reid}}{1998}]{tinney98spec}

674: {Tinney}, C.~G. \& {Reid}, I.~N., 1998,

675: \newblock {\em \mnras} { 301}, 1031

676: 

677: \bibitem[\protect\astroncite{{Tinney} \& {Tolley}}{1999}]{tinney99}

678: {Tinney}, C.~G. \& {Tolley}, A.~J., 1999,

679: \newblock {\em \mnras} { 304}, 119

680: 

681: \bibitem[\protect\astroncite{Weiss}{1996}]{weiss94}

682: Weiss, N.~O., 1996,

683: \newblock in M.~R.~E. Proctor \& A.~D. Gilbert (eds.), {\em Lectures on Solar

684:   and Planetary Dynamos}, p.~59, Cambridge University Press, Cambridge

685: 

686: \end{thebibliography}

687: 

688: 

689: % Fig 5

690: \begin{figure}[htb]

691: \caption{ \label{fig:sub4varicomp} Comparison between the 0.1-4.0 keV

692: X-ray light curve of \lp\ (top panel), and the 0.1-10.0 keV background

693: countrate (bottom panel).  Crosses mark the energies (right-hand scale)

694: of the detected counts.  The vertical broken lines mark the

695: (arbitrary) beginning and end of the flare; we refer the period prior

696: to this as the ``pre-flare'' period, and after this the ``post-flare''

697: period.  The variation observed in \lp\ is absent in the background

698: countrate taken from a much larger area on the same chip, but away

699: from the position of \lp. This demonstrates that the variability in

700: \lp\ is not due to variations in the background countrate in the chip.

701: }

702: \end{figure}

703: 

704: 

705: % Figure 5

706: \clearpage

707: \pagestyle{empty}

708: \begin{figure}[htb]

709: \PSbox{fig1.ps hoffset=-80 voffset=-80}{14.7cm}{21.5cm}

710: \FigNum{\ref{fig:sub4varicomp}}

711: \end{figure}

712: 

713: \begin{deluxetable}{lrr}

714: \scriptsize

715: \tablecaption{Derived Counts and \lxlbol\ \label{tab:flux}}

716: \tablewidth{7cm}

717: \tablehead{

718: \colhead{Period} & 

719: \colhead{\#Counts (bkg)}& 

720: \colhead{ $L_X/L_{\rm bol}$}}

721: \startdata

722: Full Obs.& 19 (0.57)	& 7\tee{-6}\\

723: Pre-flare& 1 (0.44)	& $<$2\tee{-6} \\

724: Flare	& 15 (0.14)	& 8\tee{-5} \\

725: Peak Flare& 7 (0.02)	& 2\tee{-4}\\

726: \enddata

727: \tablecomments{Assumed 1 count = 2.8\tee{-12}\erg \percm.   Upper limit

728: is 3$\sigma$. Source distance d=5.0 pc}

729: \end{deluxetable}

730: 

731: 

732: \end{document}

733: 

734: